<article>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#article09_06_26_028218</id>
	<title>Facebook VP Slams Intel's, AMD's Chip Performance Claims</title>
	<author>timothy</author>
	<datestamp>1246029780000</datestamp>
	<htmltext><a href="http://www.itworld.com/" rel="nofollow">narramissic</a> writes <i>"In an interview on stage at GigaOm's Structure conference in San Francisco on Thursday, Jonathan Heiliger, Facebook's VP of technical operations, told Om Malik that the latest generations of server processors from Intel and AMD <a href="http://www.itworld.com/server-and-data-center/69802/new-chips-dont-deliver-facebook-says">don't deliver the performance gains</a> that 'they're touting in the press.' 'And we're, literally in real time right now, trying to figure out why that is,' Heiliger said. He also had some harsh words for server makers: 'You guys don't get it,' Heiliger said. 'To build servers for companies like Facebook, and Amazon, and other people who are operating fairly homogeneous applications, the servers have to be cheap, and they have to be super power-efficient.' Heiliger added that Google has done a great job designing and building its own servers for this kind of use."</i></htmltext>
<tokenext>narramissic writes " In an interview on stage at GigaOm 's Structure conference in San Francisco on Thursday , Jonathan Heiliger , Facebook 's VP of technical operations , told Om Malik that the latest generations of server processors from Intel and AMD do n't deliver the performance gains that 'they 're touting in the press .
' 'And we 're , literally in real time right now , trying to figure out why that is, ' Heiliger said .
He also had some harsh words for server makers : 'You guys do n't get it, ' Heiliger said .
'To build servers for companies like Facebook , and Amazon , and other people who are operating fairly homogeneous applications , the servers have to be cheap , and they have to be super power-efficient .
' Heiliger added that Google has done a great job designing and building its own servers for this kind of use .
"</tokentext>
<sentencetext>narramissic writes "In an interview on stage at GigaOm's Structure conference in San Francisco on Thursday, Jonathan Heiliger, Facebook's VP of technical operations, told Om Malik that the latest generations of server processors from Intel and AMD don't deliver the performance gains that 'they're touting in the press.
' 'And we're, literally in real time right now, trying to figure out why that is,' Heiliger said.
He also had some harsh words for server makers: 'You guys don't get it,' Heiliger said.
'To build servers for companies like Facebook, and Amazon, and other people who are operating fairly homogeneous applications, the servers have to be cheap, and they have to be super power-efficient.
' Heiliger added that Google has done a great job designing and building its own servers for this kind of use.
"</sentencetext>
</article>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28485891</id>
	<title>Re:WTF?</title>
	<author>AmberBlackCat</author>
	<datestamp>1246043220000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext>And if you accuse them of lying about the performance increase, or accuse them of being unable to produce a 150mpg car, you're right.</htmltext>
<tokenext>And if you accuse them of lying about the performance increase , or accuse them of being unable to produce a 150mpg car , you 're right .</tokentext>
<sentencetext>And if you accuse them of lying about the performance increase, or accuse them of being unable to produce a 150mpg car, you're right.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477977</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477845</id>
	<title>Re:Facebook's application is poorly coded</title>
	<author>f0dder</author>
	<datestamp>1245954660000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext>Unless the bottle neck isn't in the cpu, probably a combination of I/O + expensive database queries.  Mebbe they need to get halal drives for for those AMD !slamic machines.</htmltext>
<tokenext>Unless the bottle neck is n't in the cpu , probably a combination of I/O + expensive database queries .
Mebbe they need to get halal drives for for those AMD ! slamic machines .</tokentext>
<sentencetext>Unless the bottle neck isn't in the cpu, probably a combination of I/O + expensive database queries.
Mebbe they need to get halal drives for for those AMD !slamic machines.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477085</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478901</id>
	<title>Re:WTF?</title>
	<author>rainer\_d</author>
	<datestamp>1246007160000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext><p><div class="quote"><p>I think we read different articles.  He's not saying he didn't plan well enough, he's saying that Intel and AMD promise that Gen Y processor is 35\% faster than Gen X processor, and he's not seeing anywhere near 35\% in real world performance.  </p></div><p>So his fault is that he actually believed a vendor.<br>Does he also belief in the tooth-fairy?</p></div>
	</htmltext>
<tokenext>I think we read different articles .
He 's not saying he did n't plan well enough , he 's saying that Intel and AMD promise that Gen Y processor is 35 \ % faster than Gen X processor , and he 's not seeing anywhere near 35 \ % in real world performance .
So his fault is that he actually believed a vendor.Does he also belief in the tooth-fairy ?</tokentext>
<sentencetext>I think we read different articles.
He's not saying he didn't plan well enough, he's saying that Intel and AMD promise that Gen Y processor is 35\% faster than Gen X processor, and he's not seeing anywhere near 35\% in real world performance.
So his fault is that he actually believed a vendor.Does he also belief in the tooth-fairy?
	</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477361</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478709</id>
	<title>and this is why...</title>
	<author>RMH101</author>
	<datestamp>1246048800000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext>...big companies have infrastructure architects to plan, test and make recommendations on new infrastructure.  You don't just go out and buy a container load of the latest and greatest based on the advertising copy.</htmltext>
<tokenext>...big companies have infrastructure architects to plan , test and make recommendations on new infrastructure .
You do n't just go out and buy a container load of the latest and greatest based on the advertising copy .</tokentext>
<sentencetext>...big companies have infrastructure architects to plan, test and make recommendations on new infrastructure.
You don't just go out and buy a container load of the latest and greatest based on the advertising copy.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477361</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477535</id>
	<title>Re:Hm...</title>
	<author>mabhatter654</author>
	<datestamp>1245951780000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext><p>they may be "niche" but data centers are the top multiple unit customers, buying thousands of the exact same configurations from OEMS<nobr> <wbr></nobr>.... they've surpassed the "enterprise cube farm" desktops buy a good margin in that data centers buy the MOST expensive server CPU they can where Enterprises are buying the cheapest desktops they can... Data center customers and Gamers are the only ones that NEED new CPUS every six months... nobody else really cares as they're out buying no-profit Atom based netbooks now.</p></htmltext>
<tokenext>they may be " niche " but data centers are the top multiple unit customers , buying thousands of the exact same configurations from OEMS .... they 've surpassed the " enterprise cube farm " desktops buy a good margin in that data centers buy the MOST expensive server CPU they can where Enterprises are buying the cheapest desktops they can... Data center customers and Gamers are the only ones that NEED new CPUS every six months... nobody else really cares as they 're out buying no-profit Atom based netbooks now .</tokentext>
<sentencetext>they may be "niche" but data centers are the top multiple unit customers, buying thousands of the exact same configurations from OEMS .... they've surpassed the "enterprise cube farm" desktops buy a good margin in that data centers buy the MOST expensive server CPU they can where Enterprises are buying the cheapest desktops they can... Data center customers and Gamers are the only ones that NEED new CPUS every six months... nobody else really cares as they're out buying no-profit Atom based netbooks now.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476941</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479461</id>
	<title>Re:Hm...</title>
	<author>Anonymous</author>
	<datestamp>1246012800000</datestamp>
	<modclass>None</modclass>
	<modscore>0</modscore>
	<htmltext><p>  Seriously, you do this for a living? Is it a very good living?</p><p>
&nbsp; &nbsp; You may have a limited range of experience, perhaps specialized. In any case, I can tell you that it is not grounded in current technology limits. The 'average server' doesn't use FibreChannel at all. Simple SAS or SATA storage is the standard. A small server (1U) may have a 6-disk array, pumping out perhaps 650 MB/sec or more of sequential throughput, for both reading and writing. In your terms, that is over "5Gb".</p><p>
&nbsp; &nbsp; "HBA" is really typically a FibreChannel term, which again is not even a factor in the common server.</p><p>
&nbsp; &nbsp; A larger server might have larger arrays, or a couple of RAID controllers. This can easily push the aggregate bandwith over 1GB/sec (or 8Gb, in awkward terms). There are some limits at the higher ends, of course.</p><p>
&nbsp; &nbsp; Driving this full bandwidth requires minimal CPU, when properly configured. "Properly configured" does involve a decent RAID controller, with substantial battery-backed RAM, good drivers, a good filesystem choice, and decent configuration of all. And, any common server today is going to have a minimum of 8 cores anyway.</p><p>
&nbsp; &nbsp; Custom processing pipelines can do significant work and drive the bulk of that bandwidth. A database will likely be far suboptimal - partially because of overhead in the work being done, and partially because they appear to just not be caught up to modern hardware. Still, even PostgreSQL can blow past your 1Gb, handily.</p><p>
&nbsp; &nbsp; Random access times also can come into play. These are still I/O.</p><p>
&nbsp; &nbsp; Basically, there are many common cases where even a database will have low CPU usage and minimal benefit from adding memory. What do you think it is bottlenecked on, in such cases?</p></htmltext>
<tokenext>Seriously , you do this for a living ?
Is it a very good living ?
    You may have a limited range of experience , perhaps specialized .
In any case , I can tell you that it is not grounded in current technology limits .
The 'average server ' does n't use FibreChannel at all .
Simple SAS or SATA storage is the standard .
A small server ( 1U ) may have a 6-disk array , pumping out perhaps 650 MB/sec or more of sequential throughput , for both reading and writing .
In your terms , that is over " 5Gb " .
    " HBA " is really typically a FibreChannel term , which again is not even a factor in the common server .
    A larger server might have larger arrays , or a couple of RAID controllers .
This can easily push the aggregate bandwith over 1GB/sec ( or 8Gb , in awkward terms ) .
There are some limits at the higher ends , of course .
    Driving this full bandwidth requires minimal CPU , when properly configured .
" Properly configured " does involve a decent RAID controller , with substantial battery-backed RAM , good drivers , a good filesystem choice , and decent configuration of all .
And , any common server today is going to have a minimum of 8 cores anyway .
    Custom processing pipelines can do significant work and drive the bulk of that bandwidth .
A database will likely be far suboptimal - partially because of overhead in the work being done , and partially because they appear to just not be caught up to modern hardware .
Still , even PostgreSQL can blow past your 1Gb , handily .
    Random access times also can come into play .
These are still I/O .
    Basically , there are many common cases where even a database will have low CPU usage and minimal benefit from adding memory .
What do you think it is bottlenecked on , in such cases ?</tokentext>
<sentencetext>  Seriously, you do this for a living?
Is it a very good living?
    You may have a limited range of experience, perhaps specialized.
In any case, I can tell you that it is not grounded in current technology limits.
The 'average server' doesn't use FibreChannel at all.
Simple SAS or SATA storage is the standard.
A small server (1U) may have a 6-disk array, pumping out perhaps 650 MB/sec or more of sequential throughput, for both reading and writing.
In your terms, that is over "5Gb".
    "HBA" is really typically a FibreChannel term, which again is not even a factor in the common server.
    A larger server might have larger arrays, or a couple of RAID controllers.
This can easily push the aggregate bandwith over 1GB/sec (or 8Gb, in awkward terms).
There are some limits at the higher ends, of course.
    Driving this full bandwidth requires minimal CPU, when properly configured.
"Properly configured" does involve a decent RAID controller, with substantial battery-backed RAM, good drivers, a good filesystem choice, and decent configuration of all.
And, any common server today is going to have a minimum of 8 cores anyway.
    Custom processing pipelines can do significant work and drive the bulk of that bandwidth.
A database will likely be far suboptimal - partially because of overhead in the work being done, and partially because they appear to just not be caught up to modern hardware.
Still, even PostgreSQL can blow past your 1Gb, handily.
    Random access times also can come into play.
These are still I/O.
    Basically, there are many common cases where even a database will have low CPU usage and minimal benefit from adding memory.
What do you think it is bottlenecked on, in such cases?</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477411</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477185</id>
	<title>Re:Hm...</title>
	<author>Anonymous</author>
	<datestamp>1245948900000</datestamp>
	<modclass>Informativ</modclass>
	<modscore>2</modscore>
	<htmltext><p><div class="quote"><p>Hm, lets see... perhaps because Facebook and Amazon are niche markets?</p></div><p>-Maybe-. Even if they are a niche market, they're a big enough one to hold the attention of the big chipmakers.

</p><p>A traditional business model might use large orders, especially advance orders, to offset or defray the cost of setting up a production line or facility, and get most of the profit from smaller sales. Or they may choose only to do production runs for large, inherently profitable orders. Even in a firing-from-the-hip model, large customers cost less per unit in marketing and sales than do smaller ones, very much so when compared to the general public. And of course there's plenty of wiggle room between extremes. So depending on the diversity of the market and the choice of business model, big customers range from important to desirable. Naturally, in a niche market large customers have a greater importance, since smaller sales are fewer.

</p><p>Presumably, AMD and Intel are selling servers to the likes of Amazon and Facebook 'cause they think it's profitable. If it is a niche market, keeping those guys happy is <i>paramount</i> to profitability.

</p><p>(I don't think the server farm market is really a niche, tho'. But I dunno; I don't keep up with such things.)</p><p><div class="quote"><p>And really, why is a VP complaining about this stuff? That he can't either afford custom solutions or spend the money buying more servers?</p></div><p>Well, because we <i>asked</i>. Well, not "we" as such, but someone asked him and he answered. It sounds like he was answering honestly and openly. I've no problem with that.</p></div>
	</htmltext>
<tokenext>Hm , lets see... perhaps because Facebook and Amazon are niche markets ? -Maybe- .
Even if they are a niche market , they 're a big enough one to hold the attention of the big chipmakers .
A traditional business model might use large orders , especially advance orders , to offset or defray the cost of setting up a production line or facility , and get most of the profit from smaller sales .
Or they may choose only to do production runs for large , inherently profitable orders .
Even in a firing-from-the-hip model , large customers cost less per unit in marketing and sales than do smaller ones , very much so when compared to the general public .
And of course there 's plenty of wiggle room between extremes .
So depending on the diversity of the market and the choice of business model , big customers range from important to desirable .
Naturally , in a niche market large customers have a greater importance , since smaller sales are fewer .
Presumably , AMD and Intel are selling servers to the likes of Amazon and Facebook 'cause they think it 's profitable .
If it is a niche market , keeping those guys happy is paramount to profitability .
( I do n't think the server farm market is really a niche , tho' .
But I dunno ; I do n't keep up with such things .
) And really , why is a VP complaining about this stuff ?
That he ca n't either afford custom solutions or spend the money buying more servers ? Well , because we asked .
Well , not " we " as such , but someone asked him and he answered .
It sounds like he was answering honestly and openly .
I 've no problem with that .</tokentext>
<sentencetext>Hm, lets see... perhaps because Facebook and Amazon are niche markets?-Maybe-.
Even if they are a niche market, they're a big enough one to hold the attention of the big chipmakers.
A traditional business model might use large orders, especially advance orders, to offset or defray the cost of setting up a production line or facility, and get most of the profit from smaller sales.
Or they may choose only to do production runs for large, inherently profitable orders.
Even in a firing-from-the-hip model, large customers cost less per unit in marketing and sales than do smaller ones, very much so when compared to the general public.
And of course there's plenty of wiggle room between extremes.
So depending on the diversity of the market and the choice of business model, big customers range from important to desirable.
Naturally, in a niche market large customers have a greater importance, since smaller sales are fewer.
Presumably, AMD and Intel are selling servers to the likes of Amazon and Facebook 'cause they think it's profitable.
If it is a niche market, keeping those guys happy is paramount to profitability.
(I don't think the server farm market is really a niche, tho'.
But I dunno; I don't keep up with such things.
)And really, why is a VP complaining about this stuff?
That he can't either afford custom solutions or spend the money buying more servers?Well, because we asked.
Well, not "we" as such, but someone asked him and he answered.
It sounds like he was answering honestly and openly.
I've no problem with that.
	</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476941</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478095</id>
	<title>Re:Facebook's application is poorly coded</title>
	<author>Anonymous</author>
	<datestamp>1245957000000</datestamp>
	<modclass>None</modclass>
	<modscore>0</modscore>
	<htmltext><p><div class="quote"><p>Crap code on faster computer is still going to be faster than it was on a slower computer. He's not saying anything about how efficient their software is, just that buying new processors didn't get him the performance delta that it was supposed to. More advanced hardware should deliver a performance benefit no matter what is running on it.</p></div><p>Garbage in, garbage out.</p></div>
	</htmltext>
<tokenext>Crap code on faster computer is still going to be faster than it was on a slower computer .
He 's not saying anything about how efficient their software is , just that buying new processors did n't get him the performance delta that it was supposed to .
More advanced hardware should deliver a performance benefit no matter what is running on it.Garbage in , garbage out .</tokentext>
<sentencetext>Crap code on faster computer is still going to be faster than it was on a slower computer.
He's not saying anything about how efficient their software is, just that buying new processors didn't get him the performance delta that it was supposed to.
More advanced hardware should deliver a performance benefit no matter what is running on it.Garbage in, garbage out.
	</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477085</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477361</id>
	<title>Re:WTF?</title>
	<author>Anonymous</author>
	<datestamp>1245950340000</datestamp>
	<modclass>Insightful</modclass>
	<modscore>4</modscore>
	<htmltext><p>I think we read different articles.  He's not saying he didn't plan well enough, he's saying that Intel and AMD promise that Gen Y processor is 35\% faster than Gen X processor, and he's not seeing anywhere near 35\% in real world performance.  The 35\% is a made up number but it doesn't matter what the number is that they claim.  He's probably correct.  Manufacturers pull this stuff all the time.  Look at the recent articles on battery life claims on notebook's.  AMD came out and called BS on the whole thing and basically said if you guys don't stop lying through your teeth, the FTC is going to regulate us.  From the perspective you are taking, that would mean we have to call AMD incompetent for not understanding how batteries work and not properly selecting them.</p><p>Manufacturers ALWAYS overstate claims in computer related products.  CRT actual inches vs viewable inches (thank you lcd's for finally being honest... about inches anyway.. brightness and contrast however....) Computer speaker wattage being 1/2 or 1/4 of what's claimed.  Power supply efficiency or wattage not measuring up to claims... you name it.  He's calling out what he see's to be bogus claims based on a real world experiences in a demanding environment, the type of environment where one is always looking for better performance.  I think we should get some more information before declaring him to be the problem as I'm sure he has a whole team of people that are working on this issue.</p><p>What I'd like to see from him is some numbers.  On this Intel (or AMD) rig, we get so many operations per hour/minute/whatever.  On this new Intel (or AMD) rig which they claim is 20\% faster than the previous rig, we're only seeing this number of operations per hour which amounts to only a 7\% gain, thus we're seeing 13\% less than they are claiming.  Again, numbers made up for examples sake.  I'd also be very interested in what a typical rig of theirs looks like... X Processor, Y Ram, what type of storage system is it connected to, etc.  I think such numbers are vital to understanding the issues at hand.   We all know that vendors will run the benchmarks that makes their stuff look the best, and that is often not reflective of real world performance.   If I was Intel/AMD I'd be chiming in right about now and opening a dialog with Facebook and looking to see what the issues are.  Facebook is a big customer with huge name recognition and you want to be able to use them as an example of your solution delivering the promised performance for your marketing.  I'm going to assume (I know I know) that they are already working with the server vendor to try and see what's going on here.</p></htmltext>
<tokenext>I think we read different articles .
He 's not saying he did n't plan well enough , he 's saying that Intel and AMD promise that Gen Y processor is 35 \ % faster than Gen X processor , and he 's not seeing anywhere near 35 \ % in real world performance .
The 35 \ % is a made up number but it does n't matter what the number is that they claim .
He 's probably correct .
Manufacturers pull this stuff all the time .
Look at the recent articles on battery life claims on notebook 's .
AMD came out and called BS on the whole thing and basically said if you guys do n't stop lying through your teeth , the FTC is going to regulate us .
From the perspective you are taking , that would mean we have to call AMD incompetent for not understanding how batteries work and not properly selecting them.Manufacturers ALWAYS overstate claims in computer related products .
CRT actual inches vs viewable inches ( thank you lcd 's for finally being honest... about inches anyway.. brightness and contrast however.... ) Computer speaker wattage being 1/2 or 1/4 of what 's claimed .
Power supply efficiency or wattage not measuring up to claims... you name it .
He 's calling out what he see 's to be bogus claims based on a real world experiences in a demanding environment , the type of environment where one is always looking for better performance .
I think we should get some more information before declaring him to be the problem as I 'm sure he has a whole team of people that are working on this issue.What I 'd like to see from him is some numbers .
On this Intel ( or AMD ) rig , we get so many operations per hour/minute/whatever .
On this new Intel ( or AMD ) rig which they claim is 20 \ % faster than the previous rig , we 're only seeing this number of operations per hour which amounts to only a 7 \ % gain , thus we 're seeing 13 \ % less than they are claiming .
Again , numbers made up for examples sake .
I 'd also be very interested in what a typical rig of theirs looks like... X Processor , Y Ram , what type of storage system is it connected to , etc .
I think such numbers are vital to understanding the issues at hand .
We all know that vendors will run the benchmarks that makes their stuff look the best , and that is often not reflective of real world performance .
If I was Intel/AMD I 'd be chiming in right about now and opening a dialog with Facebook and looking to see what the issues are .
Facebook is a big customer with huge name recognition and you want to be able to use them as an example of your solution delivering the promised performance for your marketing .
I 'm going to assume ( I know I know ) that they are already working with the server vendor to try and see what 's going on here .</tokentext>
<sentencetext>I think we read different articles.
He's not saying he didn't plan well enough, he's saying that Intel and AMD promise that Gen Y processor is 35\% faster than Gen X processor, and he's not seeing anywhere near 35\% in real world performance.
The 35\% is a made up number but it doesn't matter what the number is that they claim.
He's probably correct.
Manufacturers pull this stuff all the time.
Look at the recent articles on battery life claims on notebook's.
AMD came out and called BS on the whole thing and basically said if you guys don't stop lying through your teeth, the FTC is going to regulate us.
From the perspective you are taking, that would mean we have to call AMD incompetent for not understanding how batteries work and not properly selecting them.Manufacturers ALWAYS overstate claims in computer related products.
CRT actual inches vs viewable inches (thank you lcd's for finally being honest... about inches anyway.. brightness and contrast however....) Computer speaker wattage being 1/2 or 1/4 of what's claimed.
Power supply efficiency or wattage not measuring up to claims... you name it.
He's calling out what he see's to be bogus claims based on a real world experiences in a demanding environment, the type of environment where one is always looking for better performance.
I think we should get some more information before declaring him to be the problem as I'm sure he has a whole team of people that are working on this issue.What I'd like to see from him is some numbers.
On this Intel (or AMD) rig, we get so many operations per hour/minute/whatever.
On this new Intel (or AMD) rig which they claim is 20\% faster than the previous rig, we're only seeing this number of operations per hour which amounts to only a 7\% gain, thus we're seeing 13\% less than they are claiming.
Again, numbers made up for examples sake.
I'd also be very interested in what a typical rig of theirs looks like... X Processor, Y Ram, what type of storage system is it connected to, etc.
I think such numbers are vital to understanding the issues at hand.
We all know that vendors will run the benchmarks that makes their stuff look the best, and that is often not reflective of real world performance.
If I was Intel/AMD I'd be chiming in right about now and opening a dialog with Facebook and looking to see what the issues are.
Facebook is a big customer with huge name recognition and you want to be able to use them as an example of your solution delivering the promised performance for your marketing.
I'm going to assume (I know I know) that they are already working with the server vendor to try and see what's going on here.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477069</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478529</id>
	<title>Re:PHP</title>
	<author>Anonymous</author>
	<datestamp>1246047180000</datestamp>
	<modclass>None</modclass>
	<modscore>0</modscore>
	<htmltext><p>People choose PHP because the gain in productivity is worth the loss in performance. Developers cost money too. A good developer's yearly salary and benefits, say $200K in total, costs as much as about 100 servers. Web companies decided using PHP was worth the reduced human costs and reduced time to market. Now whether AMD and Intel try to support the workload is their own problem, but if they don't, they shouldn't expect to sell too many Nehalems.</p></htmltext>
<tokenext>People choose PHP because the gain in productivity is worth the loss in performance .
Developers cost money too .
A good developer 's yearly salary and benefits , say $ 200K in total , costs as much as about 100 servers .
Web companies decided using PHP was worth the reduced human costs and reduced time to market .
Now whether AMD and Intel try to support the workload is their own problem , but if they do n't , they should n't expect to sell too many Nehalems .</tokentext>
<sentencetext>People choose PHP because the gain in productivity is worth the loss in performance.
Developers cost money too.
A good developer's yearly salary and benefits, say $200K in total, costs as much as about 100 servers.
Web companies decided using PHP was worth the reduced human costs and reduced time to market.
Now whether AMD and Intel try to support the workload is their own problem, but if they don't, they shouldn't expect to sell too many Nehalems.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477011</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477305</id>
	<title>Rub a lamp, Heiliger</title>
	<author>SeaFox</author>
	<datestamp>1245949920000</datestamp>
	<modclass>Insightful</modclass>
	<modscore>4</modscore>
	<htmltext><blockquote><div><p>'You guys don't get it,' Heiliger said. 'To build servers for companies like Facebook, and Amazon, and other people who are operating fairly homogeneous applications, the servers have to be cheap, and they have to be super power-efficient.'</p></div></blockquote><p>NEWSFLASH! Customer are tightwads.</p><p>Performance/Reliability/Price.</p><p>Pick any two, Heiliger.</p></div>
	</htmltext>
<tokenext>'You guys do n't get it, ' Heiliger said .
'To build servers for companies like Facebook , and Amazon , and other people who are operating fairly homogeneous applications , the servers have to be cheap , and they have to be super power-efficient.'NEWSFLASH !
Customer are tightwads.Performance/Reliability/Price.Pick any two , Heiliger .</tokentext>
<sentencetext>'You guys don't get it,' Heiliger said.
'To build servers for companies like Facebook, and Amazon, and other people who are operating fairly homogeneous applications, the servers have to be cheap, and they have to be super power-efficient.'NEWSFLASH!
Customer are tightwads.Performance/Reliability/Price.Pick any two, Heiliger.
	</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477451</id>
	<title>so what about google then?</title>
	<author>Klintus Fang</author>
	<datestamp>1245951060000</datestamp>
	<modclass>Insightful</modclass>
	<modscore>2</modscore>
	<htmltext><p>I'm bemused that he implies the problems with his servers are due to Intel and AMD no delivering with their chips, yet at the same time he admires google for how good a job they do in building out their machines.</p><p>he must be aware that google uses Intel and AMD chips.</p><p>his reasoning just doesn't square.</p></htmltext>
<tokenext>I 'm bemused that he implies the problems with his servers are due to Intel and AMD no delivering with their chips , yet at the same time he admires google for how good a job they do in building out their machines.he must be aware that google uses Intel and AMD chips.his reasoning just does n't square .</tokentext>
<sentencetext>I'm bemused that he implies the problems with his servers are due to Intel and AMD no delivering with their chips, yet at the same time he admires google for how good a job they do in building out their machines.he must be aware that google uses Intel and AMD chips.his reasoning just doesn't square.</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478265</id>
	<title>Re:Something about his arguement doesn't work</title>
	<author>Trepidity</author>
	<datestamp>1245958620000</datestamp>
	<modclass>Insightful</modclass>
	<modscore>2</modscore>
	<htmltext><p><div class="quote"><p>If that's his argument, then it would seem that the real conclusion is that Facebook can't build systems as good as Google's, even though they are using the same processor technology.</p></div><p>Google does have approximately 30x as many employees as Facebook, so it's not implausible that they've got a much greater ability to build in-house custom tech.</p></div>
	</htmltext>
<tokenext>If that 's his argument , then it would seem that the real conclusion is that Facebook ca n't build systems as good as Google 's , even though they are using the same processor technology.Google does have approximately 30x as many employees as Facebook , so it 's not implausible that they 've got a much greater ability to build in-house custom tech .</tokentext>
<sentencetext>If that's his argument, then it would seem that the real conclusion is that Facebook can't build systems as good as Google's, even though they are using the same processor technology.Google does have approximately 30x as many employees as Facebook, so it's not implausible that they've got a much greater ability to build in-house custom tech.
	</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476989</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28480457</id>
	<title>Re:Facebook's application is poorly coded</title>
	<author>ceoyoyo</author>
	<datestamp>1246023540000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext><p>And we all know how much better it is when the crap is flying faster.</p><p>Sure, a faster processor will make your crap code go faster, but there are all sorts of reasons why it won't make it go as fast as the CPU manufacturer's specs.</p><p>Where I'm at right now instead of linking libraries they have a bunch of little utilities that all perform fairly basic tasks on large datasets.  In order to build up a complex algorithm you have to chain together a bunch of these utilities, with LOTS of disk, or at least cache, writes in between.</p><p>A faster CPU makes a difference, but not much unless you throw in a faster disk as well.</p></htmltext>
<tokenext>And we all know how much better it is when the crap is flying faster.Sure , a faster processor will make your crap code go faster , but there are all sorts of reasons why it wo n't make it go as fast as the CPU manufacturer 's specs.Where I 'm at right now instead of linking libraries they have a bunch of little utilities that all perform fairly basic tasks on large datasets .
In order to build up a complex algorithm you have to chain together a bunch of these utilities , with LOTS of disk , or at least cache , writes in between.A faster CPU makes a difference , but not much unless you throw in a faster disk as well .</tokentext>
<sentencetext>And we all know how much better it is when the crap is flying faster.Sure, a faster processor will make your crap code go faster, but there are all sorts of reasons why it won't make it go as fast as the CPU manufacturer's specs.Where I'm at right now instead of linking libraries they have a bunch of little utilities that all perform fairly basic tasks on large datasets.
In order to build up a complex algorithm you have to chain together a bunch of these utilities, with LOTS of disk, or at least cache, writes in between.A faster CPU makes a difference, but not much unless you throw in a faster disk as well.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477085</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477475</id>
	<title>Re:WTF?</title>
	<author>MidnightBrewer</author>
	<datestamp>1245951240000</datestamp>
	<modclass>Interestin</modclass>
	<modscore>2</modscore>
	<htmltext><p>How can you be blamed for finding an acceptable solution when there simply isn't one available? He is a software developer, not a hardware one.  Not everybody can just go out and design their own servers like Google does. He's saying he's been tripped up by the fact that the server manufacturers aren't delivering on their promises; hardly something he should be blamed for. Your attempts to read more into his comment about "not being cheap" and compare it to the false words of a politician seems like a pretty big stretch.</p><p>If you read the entire article, he not only doesn't say that his decisions have led to disasters, but instead says that his infrastructure development decisions have led to very smooth transitions, even when Facebook rolls out big, new features like the customized home page URLs. He is only voicing his disappointment in saying that the servers aren't living up to the hype, and that he is still looking for a better solution.</p><p>I will say that his comment to not be cheap seems to be in direct conflict with the rest of his argument, since his criticism over AMD and Intel revolves around the fact that they need to be cheaper.  Seems a bit counter-intuitive.</p></htmltext>
<tokenext>How can you be blamed for finding an acceptable solution when there simply is n't one available ?
He is a software developer , not a hardware one .
Not everybody can just go out and design their own servers like Google does .
He 's saying he 's been tripped up by the fact that the server manufacturers are n't delivering on their promises ; hardly something he should be blamed for .
Your attempts to read more into his comment about " not being cheap " and compare it to the false words of a politician seems like a pretty big stretch.If you read the entire article , he not only does n't say that his decisions have led to disasters , but instead says that his infrastructure development decisions have led to very smooth transitions , even when Facebook rolls out big , new features like the customized home page URLs .
He is only voicing his disappointment in saying that the servers are n't living up to the hype , and that he is still looking for a better solution.I will say that his comment to not be cheap seems to be in direct conflict with the rest of his argument , since his criticism over AMD and Intel revolves around the fact that they need to be cheaper .
Seems a bit counter-intuitive .</tokentext>
<sentencetext>How can you be blamed for finding an acceptable solution when there simply isn't one available?
He is a software developer, not a hardware one.
Not everybody can just go out and design their own servers like Google does.
He's saying he's been tripped up by the fact that the server manufacturers aren't delivering on their promises; hardly something he should be blamed for.
Your attempts to read more into his comment about "not being cheap" and compare it to the false words of a politician seems like a pretty big stretch.If you read the entire article, he not only doesn't say that his decisions have led to disasters, but instead says that his infrastructure development decisions have led to very smooth transitions, even when Facebook rolls out big, new features like the customized home page URLs.
He is only voicing his disappointment in saying that the servers aren't living up to the hype, and that he is still looking for a better solution.I will say that his comment to not be cheap seems to be in direct conflict with the rest of his argument, since his criticism over AMD and Intel revolves around the fact that they need to be cheaper.
Seems a bit counter-intuitive.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477069</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477965</id>
	<title>Re:Facebook's application is poorly coded</title>
	<author>Anonymous</author>
	<datestamp>1245955920000</datestamp>
	<modclass>None</modclass>
	<modscore>0</modscore>
	<htmltext><p>I program a lot as a hobby. I have made/ am making games that are thousands of lines of code long. I may not have a PHD in programming nor do i claim to be the authority on programming but i know this. The difference between a well written program and a poorly written program is night and day. The actually performance may easily vary by a factor of 2, 10 or 100. The latest and greatest cpu will give say a 20\% performance increase...So if it is poorly written...the real thing to be looking at is fixing up the program.</p><p>So i have to completely agree with the statement</p><p>"But I don't care what sort of hardware you through at crap code you are always going to get crap performance."</p></htmltext>
<tokenext>I program a lot as a hobby .
I have made/ am making games that are thousands of lines of code long .
I may not have a PHD in programming nor do i claim to be the authority on programming but i know this .
The difference between a well written program and a poorly written program is night and day .
The actually performance may easily vary by a factor of 2 , 10 or 100 .
The latest and greatest cpu will give say a 20 \ % performance increase...So if it is poorly written...the real thing to be looking at is fixing up the program.So i have to completely agree with the statement " But I do n't care what sort of hardware you through at crap code you are always going to get crap performance .
"</tokentext>
<sentencetext>I program a lot as a hobby.
I have made/ am making games that are thousands of lines of code long.
I may not have a PHD in programming nor do i claim to be the authority on programming but i know this.
The difference between a well written program and a poorly written program is night and day.
The actually performance may easily vary by a factor of 2, 10 or 100.
The latest and greatest cpu will give say a 20\% performance increase...So if it is poorly written...the real thing to be looking at is fixing up the program.So i have to completely agree with the statement"But I don't care what sort of hardware you through at crap code you are always going to get crap performance.
"</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476965</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477873</id>
	<title>Re:Hm...</title>
	<author>Anonymous</author>
	<datestamp>1245954960000</datestamp>
	<modclass>None</modclass>
	<modscore>0</modscore>
	<htmltext><p>because when you are writing zeros, the zeros are not being copied from one device, or memory to another.</p><p>So writing zeros is more cpu intensive than writing random data that is already in memory.</p><p>If the data is already available in memory, or on a block device the cpu will issue a DMA request, and the memory controller/north bridge will complete the <br>copy task without loading any data into the cpu registers. When the data is done copying, the dma controller will generate a hardware interrupt.</p><p>The same thing can happen over the network, with tcp offloading cards, and rdma: http://www.rdmaconsortium.org/home</p></htmltext>
<tokenext>because when you are writing zeros , the zeros are not being copied from one device , or memory to another.So writing zeros is more cpu intensive than writing random data that is already in memory.If the data is already available in memory , or on a block device the cpu will issue a DMA request , and the memory controller/north bridge will complete the copy task without loading any data into the cpu registers .
When the data is done copying , the dma controller will generate a hardware interrupt.The same thing can happen over the network , with tcp offloading cards , and rdma : http : //www.rdmaconsortium.org/home</tokentext>
<sentencetext>because when you are writing zeros, the zeros are not being copied from one device, or memory to another.So writing zeros is more cpu intensive than writing random data that is already in memory.If the data is already available in memory, or on a block device the cpu will issue a DMA request, and the memory controller/north bridge will complete the copy task without loading any data into the cpu registers.
When the data is done copying, the dma controller will generate a hardware interrupt.The same thing can happen over the network, with tcp offloading cards, and rdma: http://www.rdmaconsortium.org/home</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477411</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28491011</id>
	<title>Why the F--- do I need to subsidize Facebook?</title>
	<author>tjstork</author>
	<datestamp>1246032720000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext><p>What the VP of Facebook is really asking is that Intel and AMD spend a lot of their R&amp;D on designing features they want, and then pass those costs onto consumers everywhere, so Facebook can get cheaper servers.</p><p>F--- that.</p></htmltext>
<tokenext>What the VP of Facebook is really asking is that Intel and AMD spend a lot of their R&amp;D on designing features they want , and then pass those costs onto consumers everywhere , so Facebook can get cheaper servers.F--- that .</tokentext>
<sentencetext>What the VP of Facebook is really asking is that Intel and AMD spend a lot of their R&amp;D on designing features they want, and then pass those costs onto consumers everywhere, so Facebook can get cheaper servers.F--- that.</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479055</id>
	<title>Re:You're Computin' for a Shootin' Mister</title>
	<author>metacell</author>
	<datestamp>1246008540000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext><p><div class="quote"><p>I assure you, despite your misconception that the world revolves around you everyone has those requirements.  From the people who build supercomputers right down to the netbook I am typing on while watching Gurren Lagann.</p></div><p>No, not everyone has those requirements. A scientific workstation needs to have great multi-processor performance, at the expense of cost and power effiency. A computer embedded in a weapon needs to have solid reliabilty, at the expense of cost and sometimes, performance. A gaming desktop needs to have great performance, at the expense of cost and power effiency. A gaming laptop needs to have good performance and power-efficiency, at the expense of cost. And so on.</p></div>
	</htmltext>
<tokenext>I assure you , despite your misconception that the world revolves around you everyone has those requirements .
From the people who build supercomputers right down to the netbook I am typing on while watching Gurren Lagann.No , not everyone has those requirements .
A scientific workstation needs to have great multi-processor performance , at the expense of cost and power effiency .
A computer embedded in a weapon needs to have solid reliabilty , at the expense of cost and sometimes , performance .
A gaming desktop needs to have great performance , at the expense of cost and power effiency .
A gaming laptop needs to have good performance and power-efficiency , at the expense of cost .
And so on .</tokentext>
<sentencetext>I assure you, despite your misconception that the world revolves around you everyone has those requirements.
From the people who build supercomputers right down to the netbook I am typing on while watching Gurren Lagann.No, not everyone has those requirements.
A scientific workstation needs to have great multi-processor performance, at the expense of cost and power effiency.
A computer embedded in a weapon needs to have solid reliabilty, at the expense of cost and sometimes, performance.
A gaming desktop needs to have great performance, at the expense of cost and power effiency.
A gaming laptop needs to have good performance and power-efficiency, at the expense of cost.
And so on.
	</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476905</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479041</id>
	<title>Re:Facebook's application is poorly coded</title>
	<author>gbjbaanb</author>
	<datestamp>1246008360000</datestamp>
	<modclass>Insightful</modclass>
	<modscore>2</modscore>
	<htmltext><p>throwing proportionally faster CPUs at *good* code should make it proportionally faster.</p><p>Crap code.... probably not. For example, I once had to improve the performance of an app. The app spent most of its time context switching from one thread to another, more time was taken up stopping a thread, switching to another, refilling the cache lines, and so on that was spent processing the data! Think what a faster processor would do here - the CPU would process the little bit of data it was given faster thus providing much more CPU time for context switching.</p><p>Similarly with other aspects of modern code - relatively little of it is spent spinning CPU cycles. I'd say more was spent dealing with memory IO (as there is a lot of RAM used nowadays, getting that data to and from the CPU is, relatively speaking, slow as treacle) so it wouldn't matter if you could crunch the data faster if you still had to wait for it to be delivered to you.</p><p>Then we put more stuff on the network, and connect to it via Web services and the like, and the amount of CPU power required gets less and less relevant.</p><p>I'd say the single best thing you can do to get good performance, and therefore energy efficiency, and cheapness of resources is to write efficient code that requires little resources itself. Even if it takes you longer to do the job, tough on you - there's just you as a programmer but millions of users, the extra time spent developing at a lower level (instead of pointy-clicking in the IDE) is time well spent.</p><p>If Facebook's code could be made 10\% more efficient, they'd require 10\% less servers with all the reduced energy bill that entails. But the Facebook chap doesn't care about that - that'd cost him programmer time, and that costs short-term money! Far better for him to whinge that Intel and AMD aren't fixing his shit for him instead.</p></htmltext>
<tokenext>throwing proportionally faster CPUs at * good * code should make it proportionally faster.Crap code.... probably not .
For example , I once had to improve the performance of an app .
The app spent most of its time context switching from one thread to another , more time was taken up stopping a thread , switching to another , refilling the cache lines , and so on that was spent processing the data !
Think what a faster processor would do here - the CPU would process the little bit of data it was given faster thus providing much more CPU time for context switching.Similarly with other aspects of modern code - relatively little of it is spent spinning CPU cycles .
I 'd say more was spent dealing with memory IO ( as there is a lot of RAM used nowadays , getting that data to and from the CPU is , relatively speaking , slow as treacle ) so it would n't matter if you could crunch the data faster if you still had to wait for it to be delivered to you.Then we put more stuff on the network , and connect to it via Web services and the like , and the amount of CPU power required gets less and less relevant.I 'd say the single best thing you can do to get good performance , and therefore energy efficiency , and cheapness of resources is to write efficient code that requires little resources itself .
Even if it takes you longer to do the job , tough on you - there 's just you as a programmer but millions of users , the extra time spent developing at a lower level ( instead of pointy-clicking in the IDE ) is time well spent.If Facebook 's code could be made 10 \ % more efficient , they 'd require 10 \ % less servers with all the reduced energy bill that entails .
But the Facebook chap does n't care about that - that 'd cost him programmer time , and that costs short-term money !
Far better for him to whinge that Intel and AMD are n't fixing his shit for him instead .</tokentext>
<sentencetext>throwing proportionally faster CPUs at *good* code should make it proportionally faster.Crap code.... probably not.
For example, I once had to improve the performance of an app.
The app spent most of its time context switching from one thread to another, more time was taken up stopping a thread, switching to another, refilling the cache lines, and so on that was spent processing the data!
Think what a faster processor would do here - the CPU would process the little bit of data it was given faster thus providing much more CPU time for context switching.Similarly with other aspects of modern code - relatively little of it is spent spinning CPU cycles.
I'd say more was spent dealing with memory IO (as there is a lot of RAM used nowadays, getting that data to and from the CPU is, relatively speaking, slow as treacle) so it wouldn't matter if you could crunch the data faster if you still had to wait for it to be delivered to you.Then we put more stuff on the network, and connect to it via Web services and the like, and the amount of CPU power required gets less and less relevant.I'd say the single best thing you can do to get good performance, and therefore energy efficiency, and cheapness of resources is to write efficient code that requires little resources itself.
Even if it takes you longer to do the job, tough on you - there's just you as a programmer but millions of users, the extra time spent developing at a lower level (instead of pointy-clicking in the IDE) is time well spent.If Facebook's code could be made 10\% more efficient, they'd require 10\% less servers with all the reduced energy bill that entails.
But the Facebook chap doesn't care about that - that'd cost him programmer time, and that costs short-term money!
Far better for him to whinge that Intel and AMD aren't fixing his shit for him instead.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477167</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477999</id>
	<title>Re:Facebook's application is poorly coded</title>
	<author>Anonymous</author>
	<datestamp>1245956160000</datestamp>
	<modclass>None</modclass>
	<modscore>0</modscore>
	<htmltext>Question marks. Have you heard of them?</htmltext>
<tokenext>Question marks .
Have you heard of them ?</tokentext>
<sentencetext>Question marks.
Have you heard of them?</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476965</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28480465</id>
	<title>Re:You're Computin' for a Shootin' Mister</title>
	<author>Anonymous</author>
	<datestamp>1246023600000</datestamp>
	<modclass>None</modclass>
	<modscore>0</modscore>
	<htmltext><p>When you're company isn't pulling in the revenues you were expecting and you have a lot of investors to please, you'll find all sorts of people to blame for why you're not making billions in profit.</p></htmltext>
<tokenext>When you 're company is n't pulling in the revenues you were expecting and you have a lot of investors to please , you 'll find all sorts of people to blame for why you 're not making billions in profit .</tokentext>
<sentencetext>When you're company isn't pulling in the revenues you were expecting and you have a lot of investors to please, you'll find all sorts of people to blame for why you're not making billions in profit.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476905</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477425</id>
	<title>Re:PHP</title>
	<author>MightyMartian</author>
	<datestamp>1245950880000</datestamp>
	<modclass>Insightful</modclass>
	<modscore>2</modscore>
	<htmltext><p>Exactly.  All these interpreted languages, even with some special tricks, will have serious scalability issues.  At some point you have to look at the application and ask some serious questions.</p></htmltext>
<tokenext>Exactly .
All these interpreted languages , even with some special tricks , will have serious scalability issues .
At some point you have to look at the application and ask some serious questions .</tokentext>
<sentencetext>Exactly.
All these interpreted languages, even with some special tricks, will have serious scalability issues.
At some point you have to look at the application and ask some serious questions.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477011</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479831</id>
	<title>Re:Facebook's application is poorly coded</title>
	<author>hattig</author>
	<datestamp>1246017180000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext><p>Well CPUs process data.</p><p>If the data can't be provided quickly enough, then all that will happen is that the faster CPU will process what it gets quicker, and then have a nap waiting for some more.</p><p>I'm sure Facebook has multiple replicated databases in order to handle the load, and I'm sure that the servers have tonnes of RAM in order to cache information in user sessions to reduce load on the database... Yes? I'm sure they're not using SOAP/XML for internal messaging, but using a binary protocol like Google's protocol buffers, but above that doing all they can to reduce network I/O, such as keeping each user session on the same front-end webserver, etc.</p><p>See, Google's system works because Google hires good software engineers who solve problems so that their internal networks can cope.</p></htmltext>
<tokenext>Well CPUs process data.If the data ca n't be provided quickly enough , then all that will happen is that the faster CPU will process what it gets quicker , and then have a nap waiting for some more.I 'm sure Facebook has multiple replicated databases in order to handle the load , and I 'm sure that the servers have tonnes of RAM in order to cache information in user sessions to reduce load on the database... Yes ? I 'm sure they 're not using SOAP/XML for internal messaging , but using a binary protocol like Google 's protocol buffers , but above that doing all they can to reduce network I/O , such as keeping each user session on the same front-end webserver , etc.See , Google 's system works because Google hires good software engineers who solve problems so that their internal networks can cope .</tokentext>
<sentencetext>Well CPUs process data.If the data can't be provided quickly enough, then all that will happen is that the faster CPU will process what it gets quicker, and then have a nap waiting for some more.I'm sure Facebook has multiple replicated databases in order to handle the load, and I'm sure that the servers have tonnes of RAM in order to cache information in user sessions to reduce load on the database... Yes? I'm sure they're not using SOAP/XML for internal messaging, but using a binary protocol like Google's protocol buffers, but above that doing all they can to reduce network I/O, such as keeping each user session on the same front-end webserver, etc.See, Google's system works because Google hires good software engineers who solve problems so that their internal networks can cope.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477085</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28481125</id>
	<title>I wonder..</title>
	<author>Hoonis</author>
	<datestamp>1246026720000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext><p>Are they using the intel compiler?<br>Are they making their binaries more thread-friendly?<br>Are they using cc flags that exploit the new cache/features/instructions?<br>Are they running their OS's with at least basic tuning to contain interrupts and kernel activity to particular CPUs?<br>Are they using new hyperthreading efficiently and consciously or just disabling it?<br>Are they running with taskset wrappers to decrease context switching on an obviously stochastic workload?<br>Are they tweaking their networking configuration internally to optimize for their specific packet loads?</p><p>Because I've found that the new hardware (Nehalems, for example) will give you a heck of a speed boost if you're doing the above.</p><p>It's real easy to have the compatibility features mask performance, really you gotta look at it like a new platform.</p><p>Anyway, those comments smell like making excuses to me, they are not even remotely specific enough to be defensible.</p></htmltext>
<tokenext>Are they using the intel compiler ? Are they making their binaries more thread-friendly ? Are they using cc flags that exploit the new cache/features/instructions ? Are they running their OS 's with at least basic tuning to contain interrupts and kernel activity to particular CPUs ? Are they using new hyperthreading efficiently and consciously or just disabling it ? Are they running with taskset wrappers to decrease context switching on an obviously stochastic workload ? Are they tweaking their networking configuration internally to optimize for their specific packet loads ? Because I 've found that the new hardware ( Nehalems , for example ) will give you a heck of a speed boost if you 're doing the above.It 's real easy to have the compatibility features mask performance , really you got ta look at it like a new platform.Anyway , those comments smell like making excuses to me , they are not even remotely specific enough to be defensible .</tokentext>
<sentencetext>Are they using the intel compiler?Are they making their binaries more thread-friendly?Are they using cc flags that exploit the new cache/features/instructions?Are they running their OS's with at least basic tuning to contain interrupts and kernel activity to particular CPUs?Are they using new hyperthreading efficiently and consciously or just disabling it?Are they running with taskset wrappers to decrease context switching on an obviously stochastic workload?Are they tweaking their networking configuration internally to optimize for their specific packet loads?Because I've found that the new hardware (Nehalems, for example) will give you a heck of a speed boost if you're doing the above.It's real easy to have the compatibility features mask performance, really you gotta look at it like a new platform.Anyway, those comments smell like making excuses to me, they are not even remotely specific enough to be defensible.</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478839</id>
	<title>Re:WTF?</title>
	<author>Anonymous</author>
	<datestamp>1246049880000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext>So my take was that he's a total choad.  He's blaming the chip makers for his software's poor performance and his company's inability to design a scalable architecture?

If he is so damn sure of what he wants, then why isn't he putting it together himself?  He blames the server manufacturers for failing to provide both super fast and ultra low-power machines.  He spec'd the whole thing out of Unobtanium and then complained that *surprise* nobody can deliver it.

Maybe his app is I/O or memory bound, and a 35\% increase in CPU performance isn't going to do shit.  Maybe he should have purchased a couple new CPUs and tested them before committing to a large purchase and then pointing his finger at the marketing reps.

He says servers have to be powerful, power efficient, and cheap.  Then he says his advice is "don't be cheap."   What a choad-farmer.  I'll bet he bitches that his wife isn't Smart, Beautiful, AND Sane.  He doesn't seem to get the concept of "You have three options: pick two."

That's my two cents--keep the change.</htmltext>
<tokenext>So my take was that he 's a total choad .
He 's blaming the chip makers for his software 's poor performance and his company 's inability to design a scalable architecture ?
If he is so damn sure of what he wants , then why is n't he putting it together himself ?
He blames the server manufacturers for failing to provide both super fast and ultra low-power machines .
He spec 'd the whole thing out of Unobtanium and then complained that * surprise * nobody can deliver it .
Maybe his app is I/O or memory bound , and a 35 \ % increase in CPU performance is n't going to do shit .
Maybe he should have purchased a couple new CPUs and tested them before committing to a large purchase and then pointing his finger at the marketing reps . He says servers have to be powerful , power efficient , and cheap .
Then he says his advice is " do n't be cheap .
" What a choad-farmer .
I 'll bet he bitches that his wife is n't Smart , Beautiful , AND Sane .
He does n't seem to get the concept of " You have three options : pick two .
" That 's my two cents--keep the change .</tokentext>
<sentencetext>So my take was that he's a total choad.
He's blaming the chip makers for his software's poor performance and his company's inability to design a scalable architecture?
If he is so damn sure of what he wants, then why isn't he putting it together himself?
He blames the server manufacturers for failing to provide both super fast and ultra low-power machines.
He spec'd the whole thing out of Unobtanium and then complained that *surprise* nobody can deliver it.
Maybe his app is I/O or memory bound, and a 35\% increase in CPU performance isn't going to do shit.
Maybe he should have purchased a couple new CPUs and tested them before committing to a large purchase and then pointing his finger at the marketing reps.

He says servers have to be powerful, power efficient, and cheap.
Then he says his advice is "don't be cheap.
"   What a choad-farmer.
I'll bet he bitches that his wife isn't Smart, Beautiful, AND Sane.
He doesn't seem to get the concept of "You have three options: pick two.
"

That's my two cents--keep the change.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477361</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28512067</id>
	<title>Re:Facebook's application is poorly coded</title>
	<author>Anonymous</author>
	<datestamp>1246276680000</datestamp>
	<modclass>None</modclass>
	<modscore>0</modscore>
	<htmltext><p>Crap code usually does not scale linearly.</p></htmltext>
<tokenext>Crap code usually does not scale linearly .</tokentext>
<sentencetext>Crap code usually does not scale linearly.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477167</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477167</id>
	<title>Re:Facebook's application is poorly coded</title>
	<author>Anonymous</author>
	<datestamp>1245948780000</datestamp>
	<modclass>Insightful</modclass>
	<modscore>3</modscore>
	<htmltext><p>Developers have been known to trade off performance for development ease.  Frequently the result is crap code.  Yes, it performs like crap on both sets of processors.  But if the application is CPU-limited (rather than IO or memory or...), then throwing faster CPUs at it ought to make it proportionally faster, no?  Obviously they thought the previous performance was acceptable, is it unreasonable to think that buying CPUs marketed as 50\% faster should give a 50\% performance increase?  Clearly crap code will still run like crap, but you ought to be able to throw more CPU power at it and get 150\% of crap performance.</p></htmltext>
<tokenext>Developers have been known to trade off performance for development ease .
Frequently the result is crap code .
Yes , it performs like crap on both sets of processors .
But if the application is CPU-limited ( rather than IO or memory or... ) , then throwing faster CPUs at it ought to make it proportionally faster , no ?
Obviously they thought the previous performance was acceptable , is it unreasonable to think that buying CPUs marketed as 50 \ % faster should give a 50 \ % performance increase ?
Clearly crap code will still run like crap , but you ought to be able to throw more CPU power at it and get 150 \ % of crap performance .</tokentext>
<sentencetext>Developers have been known to trade off performance for development ease.
Frequently the result is crap code.
Yes, it performs like crap on both sets of processors.
But if the application is CPU-limited (rather than IO or memory or...), then throwing faster CPUs at it ought to make it proportionally faster, no?
Obviously they thought the previous performance was acceptable, is it unreasonable to think that buying CPUs marketed as 50\% faster should give a 50\% performance increase?
Clearly crap code will still run like crap, but you ought to be able to throw more CPU power at it and get 150\% of crap performance.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476965</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476975</id>
	<title>Take that...</title>
	<author>actionbastard</author>
	<datestamp>1245947520000</datestamp>
	<modclass>Offtopic</modclass>
	<modscore>-1</modscore>
	<htmltext>bitches.</htmltext>
<tokenext>bitches .</tokentext>
<sentencetext>bitches.</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28480675</id>
	<title>Re:You're Computin' for a Shootin' Mister</title>
	<author>rgviza</author>
	<datestamp>1246024860000</datestamp>
	<modclass>Insightful</modclass>
	<modscore>2</modscore>
	<htmltext><p>that guy is an ass.</p><p><div class="quote"><p>the latest generations of server processors from Intel and AMD don't deliver the performance gains that 'they're touting in the press</p></div><p>then</p><p><div class="quote"><p>Google has done a great job designing and building its own servers for this kind of use</p></div><p>I wonder who makes the server processors for Google's servers. Hmmm.....</p></div>
	</htmltext>
<tokenext>that guy is an ass.the latest generations of server processors from Intel and AMD do n't deliver the performance gains that 'they 're touting in the pressthenGoogle has done a great job designing and building its own servers for this kind of useI wonder who makes the server processors for Google 's servers .
Hmmm.... .</tokentext>
<sentencetext>that guy is an ass.the latest generations of server processors from Intel and AMD don't deliver the performance gains that 'they're touting in the pressthenGoogle has done a great job designing and building its own servers for this kind of useI wonder who makes the server processors for Google's servers.
Hmmm.....
	</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476905</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476905</id>
	<title>You're Computin' for a Shootin' Mister</title>
	<author>Anonymous</author>
	<datestamp>1245947040000</datestamp>
	<modclass>Insightful</modclass>
	<modscore>5</modscore>
	<htmltext>

So let me get this straight, the Vice President of a web company is criticizing the hardware guys in two of the world's biggest chip makers?<p><div class="quote"><p>You guys don't get it</p></div><p>Is it possible to take out a massive life insurance policy on Jonathan Heiliger?</p><p><div class="quote"><p>To build servers for companies like Facebook, and Amazon, and other people who are operating fairly homogeneous applications, <b>the servers have to be cheap, and they have to be super power-efficient</b>.</p></div><p>I assure you, despite your misconception that the world revolves around you everyone has those requirements.  From the people who build supercomputers right down to the netbook I am typing on while watching Gurren Lagann.  <br> <br>

Can we get like a panel of hardware engineers to have a discussion with this guy and can I get some popcorn?</p></div>
	</htmltext>
<tokenext>So let me get this straight , the Vice President of a web company is criticizing the hardware guys in two of the world 's biggest chip makers ? You guys do n't get itIs it possible to take out a massive life insurance policy on Jonathan Heiliger ? To build servers for companies like Facebook , and Amazon , and other people who are operating fairly homogeneous applications , the servers have to be cheap , and they have to be super power-efficient.I assure you , despite your misconception that the world revolves around you everyone has those requirements .
From the people who build supercomputers right down to the netbook I am typing on while watching Gurren Lagann .
Can we get like a panel of hardware engineers to have a discussion with this guy and can I get some popcorn ?</tokentext>
<sentencetext>

So let me get this straight, the Vice President of a web company is criticizing the hardware guys in two of the world's biggest chip makers?You guys don't get itIs it possible to take out a massive life insurance policy on Jonathan Heiliger?To build servers for companies like Facebook, and Amazon, and other people who are operating fairly homogeneous applications, the servers have to be cheap, and they have to be super power-efficient.I assure you, despite your misconception that the world revolves around you everyone has those requirements.
From the people who build supercomputers right down to the netbook I am typing on while watching Gurren Lagann.
Can we get like a panel of hardware engineers to have a discussion with this guy and can I get some popcorn?
	</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477085</id>
	<title>Re:Facebook's application is poorly coded</title>
	<author>royallthefourth</author>
	<datestamp>1245948240000</datestamp>
	<modclass>Insightful</modclass>
	<modscore>3</modscore>
	<htmltext>Crap code on faster computer is still going to be faster than it was on a slower computer. He's not saying anything about how efficient their software is, just that buying new processors didn't get him the performance delta that it was supposed to. More advanced hardware should deliver a performance benefit no matter what is running on it.</htmltext>
<tokenext>Crap code on faster computer is still going to be faster than it was on a slower computer .
He 's not saying anything about how efficient their software is , just that buying new processors did n't get him the performance delta that it was supposed to .
More advanced hardware should deliver a performance benefit no matter what is running on it .</tokentext>
<sentencetext>Crap code on faster computer is still going to be faster than it was on a slower computer.
He's not saying anything about how efficient their software is, just that buying new processors didn't get him the performance delta that it was supposed to.
More advanced hardware should deliver a performance benefit no matter what is running on it.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476965</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477831</id>
	<title>Re:Facebook's application is poorly coded</title>
	<author>Bill, Shooter of Bul</author>
	<datestamp>1245954480000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext>Twitter went from  pure ruby on rails, to some ruby messaging class, to a scala written one. The first two were awful, haven't looked too much at the last one. Facebook seems more reliable, and has dealt with its ever increasing userbase pretty well. They seem to use sensible parts including hadoop. They just recently hired away Mysql guru Mark Callaghan from Google. Not sure what they're doing with mysql, but I was never sure what Google did with it either.</htmltext>
<tokenext>Twitter went from pure ruby on rails , to some ruby messaging class , to a scala written one .
The first two were awful , have n't looked too much at the last one .
Facebook seems more reliable , and has dealt with its ever increasing userbase pretty well .
They seem to use sensible parts including hadoop .
They just recently hired away Mysql guru Mark Callaghan from Google .
Not sure what they 're doing with mysql , but I was never sure what Google did with it either .</tokentext>
<sentencetext>Twitter went from  pure ruby on rails, to some ruby messaging class, to a scala written one.
The first two were awful, haven't looked too much at the last one.
Facebook seems more reliable, and has dealt with its ever increasing userbase pretty well.
They seem to use sensible parts including hadoop.
They just recently hired away Mysql guru Mark Callaghan from Google.
Not sure what they're doing with mysql, but I was never sure what Google did with it either.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476965</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477457</id>
	<title>Re:I wonder if he's not thinking about scaling eno</title>
	<author>Anonymous</author>
	<datestamp>1245951120000</datestamp>
	<modclass>None</modclass>
	<modscore>0</modscore>
	<htmltext><p><div class="quote"><p>I wonder if you can run<nobr> <wbr></nobr>.net or Java under OS/390 or MVS...</p></div><p>Yes, actually, though we've been calling it z/OS for about 10 years now, and it's an hybrid mix of MVS and Unix.<br>Java definitely. Not sure if there's a<nobr> <wbr></nobr>.NET engine yet - might need to run Linux for z-Series in an LPAR or under z/VM to get a<nobr> <wbr></nobr>.NET framework</p></div>
	</htmltext>
<tokenext>I wonder if you can run .net or Java under OS/390 or MVS...Yes , actually , though we 've been calling it z/OS for about 10 years now , and it 's an hybrid mix of MVS and Unix.Java definitely .
Not sure if there 's a .NET engine yet - might need to run Linux for z-Series in an LPAR or under z/VM to get a .NET framework</tokentext>
<sentencetext>I wonder if you can run .net or Java under OS/390 or MVS...Yes, actually, though we've been calling it z/OS for about 10 years now, and it's an hybrid mix of MVS and Unix.Java definitely.
Not sure if there's a .NET engine yet - might need to run Linux for z-Series in an LPAR or under z/VM to get a .NET framework
	</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477107</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477625</id>
	<title>ARM?</title>
	<author>sznupi</author>
	<datestamp>1245952740000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext><p>I wonder when we'll see servers with CPUs based on (many...) ARM cores.</p><p>Yes, they are an order of magnitude slower, but three orders of magnitude more power efficient. For the same CPU performance you'd probably be around two orders of magnitude more power efficient (for CPUs at least). If your app runs on a large farm already...</p></htmltext>
<tokenext>I wonder when we 'll see servers with CPUs based on ( many... ) ARM cores.Yes , they are an order of magnitude slower , but three orders of magnitude more power efficient .
For the same CPU performance you 'd probably be around two orders of magnitude more power efficient ( for CPUs at least ) .
If your app runs on a large farm already.. .</tokentext>
<sentencetext>I wonder when we'll see servers with CPUs based on (many...) ARM cores.Yes, they are an order of magnitude slower, but three orders of magnitude more power efficient.
For the same CPU performance you'd probably be around two orders of magnitude more power efficient (for CPUs at least).
If your app runs on a large farm already...</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476967</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28480449</id>
	<title>Re:You're Computin' for a Shootin' Mister</title>
	<author>jmccay</author>
	<datestamp>1246023540000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext>   I wonder if the VP bought multi-core technologies expecting that there would be an instantaneous improvement.  The truth is that multi-core technologies don't guarantee massive improvements if the software was not written to take advantage of more than one processor.  I think this is more a case of a VP seeing too much of the "big picture" and not enough of the details.</htmltext>
<tokenext>I wonder if the VP bought multi-core technologies expecting that there would be an instantaneous improvement .
The truth is that multi-core technologies do n't guarantee massive improvements if the software was not written to take advantage of more than one processor .
I think this is more a case of a VP seeing too much of the " big picture " and not enough of the details .</tokentext>
<sentencetext>   I wonder if the VP bought multi-core technologies expecting that there would be an instantaneous improvement.
The truth is that multi-core technologies don't guarantee massive improvements if the software was not written to take advantage of more than one processor.
I think this is more a case of a VP seeing too much of the "big picture" and not enough of the details.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476905</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477467</id>
	<title>Re:Hm...</title>
	<author>Anonymous</author>
	<datestamp>1245951180000</datestamp>
	<modclass>None</modclass>
	<modscore>0</modscore>
	<htmltext><p>I assume the masses of PHP scripts they have to run aren't CPU intensive? And rather large SQL databases...</p></htmltext>
<tokenext>I assume the masses of PHP scripts they have to run are n't CPU intensive ?
And rather large SQL databases.. .</tokentext>
<sentencetext>I assume the masses of PHP scripts they have to run aren't CPU intensive?
And rather large SQL databases...</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476941</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28488783</id>
	<title>Re:so what about google then?</title>
	<author>LackThereof</author>
	<datestamp>1246013280000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext><p>IIRC, Google doesn't use top-end "server" chips in their servers.  They use consumer grade, midrange chips that they can get at cheap commodity prices, and load balance everything across a ton of machines.</p><p>He implies the problems are due to Intel and AMD not delivering with their <em>server</em> chips; these are not the chips that Google is using.</p></htmltext>
<tokenext>IIRC , Google does n't use top-end " server " chips in their servers .
They use consumer grade , midrange chips that they can get at cheap commodity prices , and load balance everything across a ton of machines.He implies the problems are due to Intel and AMD not delivering with their server chips ; these are not the chips that Google is using .</tokentext>
<sentencetext>IIRC, Google doesn't use top-end "server" chips in their servers.
They use consumer grade, midrange chips that they can get at cheap commodity prices, and load balance everything across a ton of machines.He implies the problems are due to Intel and AMD not delivering with their server chips; these are not the chips that Google is using.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477451</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28490837</id>
	<title>Spoiled Rotten Little Kid</title>
	<author>Anonymous</author>
	<datestamp>1246030680000</datestamp>
	<modclass>None</modclass>
	<modscore>0</modscore>
	<htmltext><p>Start Looking For A New Job Jonathan ?</p><p>I think Jonathan Heiliger, VP of technical operations should first try keeping Facebook from crashing constantly. I can't go 5 minutes without the site locking up on me. Mr. Heiliger, really not impressed ! These chip makers are the one's that are keeping America Safe, right now and into the future. If you want to bad mouth someone Jonathan, bad mouth the people that work under you, that can't even keep Facebook a site worth going to anymore. You are a Spoiled Rotten Little Kid that thinks someone owes you something, the chip makers don't. Keep working harder on Facebook !</p></htmltext>
<tokenext>Start Looking For A New Job Jonathan ? I think Jonathan Heiliger , VP of technical operations should first try keeping Facebook from crashing constantly .
I ca n't go 5 minutes without the site locking up on me .
Mr. Heiliger , really not impressed !
These chip makers are the one 's that are keeping America Safe , right now and into the future .
If you want to bad mouth someone Jonathan , bad mouth the people that work under you , that ca n't even keep Facebook a site worth going to anymore .
You are a Spoiled Rotten Little Kid that thinks someone owes you something , the chip makers do n't .
Keep working harder on Facebook !</tokentext>
<sentencetext>Start Looking For A New Job Jonathan ?I think Jonathan Heiliger, VP of technical operations should first try keeping Facebook from crashing constantly.
I can't go 5 minutes without the site locking up on me.
Mr. Heiliger, really not impressed !
These chip makers are the one's that are keeping America Safe, right now and into the future.
If you want to bad mouth someone Jonathan, bad mouth the people that work under you, that can't even keep Facebook a site worth going to anymore.
You are a Spoiled Rotten Little Kid that thinks someone owes you something, the chip makers don't.
Keep working harder on Facebook !</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28485287</id>
	<title>Re:Facebook's application is poorly coded</title>
	<author>\_avs\_007</author>
	<datestamp>1246040820000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext>So you're saying if I write (for example) an applicaiton to use only a single thread, and use blocking calls for all my I/O operations, I can expect huge performance benefits from migrating to better hardware? You can put in an 8 core system, it won't help your single threaded application if it's using blocking I/O calls.</htmltext>
<tokenext>So you 're saying if I write ( for example ) an applicaiton to use only a single thread , and use blocking calls for all my I/O operations , I can expect huge performance benefits from migrating to better hardware ?
You can put in an 8 core system , it wo n't help your single threaded application if it 's using blocking I/O calls .</tokentext>
<sentencetext>So you're saying if I write (for example) an applicaiton to use only a single thread, and use blocking calls for all my I/O operations, I can expect huge performance benefits from migrating to better hardware?
You can put in an 8 core system, it won't help your single threaded application if it's using blocking I/O calls.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477085</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477531</id>
	<title>Re:Facebook's application is poorly coded</title>
	<author>Anonymous</author>
	<datestamp>1245951780000</datestamp>
	<modclass>None</modclass>
	<modscore>0</modscore>
	<htmltext><p><div class="quote"><p>...</p><p>But I don't care what sort of hardware you through at crap code you are always going to get crap performance.</p></div><p>A wise man once said: A computer's performance is 10\% hardware, 90\% software</p></div>
	</htmltext>
<tokenext>...But I do n't care what sort of hardware you through at crap code you are always going to get crap performance.A wise man once said : A computer 's performance is 10 \ % hardware , 90 \ % software</tokentext>
<sentencetext>...But I don't care what sort of hardware you through at crap code you are always going to get crap performance.A wise man once said: A computer's performance is 10\% hardware, 90\% software
	</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476965</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478081</id>
	<title>Hang on a minute...</title>
	<author>OneSmartFellow</author>
	<datestamp>1245956880000</datestamp>
	<modclass>Funny</modclass>
	<modscore>3</modscore>
	<htmltext>...I'm supposed to care about the comments of the guy who wrote Facebook ?  <br> <br>Hah, hah, hah, hah, hah !At least google needed to actually engineer their solution, but Facebook, come on !  The next time I need to write a PHP script for displaying photos and text, I'll hire my 13 year old daughter.</htmltext>
<tokenext>...I 'm supposed to care about the comments of the guy who wrote Facebook ?
Hah , hah , hah , hah , hah ! At least google needed to actually engineer their solution , but Facebook , come on !
The next time I need to write a PHP script for displaying photos and text , I 'll hire my 13 year old daughter .</tokentext>
<sentencetext>...I'm supposed to care about the comments of the guy who wrote Facebook ?
Hah, hah, hah, hah, hah !At least google needed to actually engineer their solution, but Facebook, come on !
The next time I need to write a PHP script for displaying photos and text, I'll hire my 13 year old daughter.</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478671</id>
	<title>not just the CPU it's overall system performance</title>
	<author>unix\_geek\_512</author>
	<datestamp>1246048560000</datestamp>
	<modclass>Insightful</modclass>
	<modscore>3</modscore>
	<htmltext><p>This isn't just about the CPU, it's about overall system performance.</p><p>Despite improvements in CPU performance, memory and IO performance is lagging behind.</p><p>A modern SATA drive delivers about 90MB/sec ( peak sequential read ).</p><p>Some RAID controllers can do about 600-800MB/sec ( peak sequential read ).</p><p>An average AM2 ( K10 core 65nm ) gets about 34,849MB/sec L1, 12,169MB/sec L2, 6371MB/sec L3, 2,741MB/sec DDR2-800 5-5-5-12.</p><p>Obviously Opterons scale a lot better since they each have an onboard memory controller and additional HT links which greatly increases bandwidth as you add more CPUs. However adding more cores on the same die which have to share a single memory controller can cause starvation.</p><p>Another major issue is software parallelization, writing parallel code is still a difficult problem. If your software doesn't parallelize well it doesn't matter if you have 8, 16 or even 32cores on a single die.</p><p>If you had an equal number of CPU cores and memory controllers you could achieve much better performance, however your relatively very slow storage subsystems would still be a major bottleneck.</p></htmltext>
<tokenext>This is n't just about the CPU , it 's about overall system performance.Despite improvements in CPU performance , memory and IO performance is lagging behind.A modern SATA drive delivers about 90MB/sec ( peak sequential read ) .Some RAID controllers can do about 600-800MB/sec ( peak sequential read ) .An average AM2 ( K10 core 65nm ) gets about 34,849MB/sec L1 , 12,169MB/sec L2 , 6371MB/sec L3 , 2,741MB/sec DDR2-800 5-5-5-12.Obviously Opterons scale a lot better since they each have an onboard memory controller and additional HT links which greatly increases bandwidth as you add more CPUs .
However adding more cores on the same die which have to share a single memory controller can cause starvation.Another major issue is software parallelization , writing parallel code is still a difficult problem .
If your software does n't parallelize well it does n't matter if you have 8 , 16 or even 32cores on a single die.If you had an equal number of CPU cores and memory controllers you could achieve much better performance , however your relatively very slow storage subsystems would still be a major bottleneck .</tokentext>
<sentencetext>This isn't just about the CPU, it's about overall system performance.Despite improvements in CPU performance, memory and IO performance is lagging behind.A modern SATA drive delivers about 90MB/sec ( peak sequential read ).Some RAID controllers can do about 600-800MB/sec ( peak sequential read ).An average AM2 ( K10 core 65nm ) gets about 34,849MB/sec L1, 12,169MB/sec L2, 6371MB/sec L3, 2,741MB/sec DDR2-800 5-5-5-12.Obviously Opterons scale a lot better since they each have an onboard memory controller and additional HT links which greatly increases bandwidth as you add more CPUs.
However adding more cores on the same die which have to share a single memory controller can cause starvation.Another major issue is software parallelization, writing parallel code is still a difficult problem.
If your software doesn't parallelize well it doesn't matter if you have 8, 16 or even 32cores on a single die.If you had an equal number of CPU cores and memory controllers you could achieve much better performance, however your relatively very slow storage subsystems would still be a major bottleneck.</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28483853</id>
	<title>Re:You're Computin' for a Shootin' Mister</title>
	<author>DragonWriter</author>
	<datestamp>1246035480000</datestamp>
	<modclass>Insightful</modclass>
	<modscore>2</modscore>
	<htmltext><blockquote><div><p>So let me get this straight, the Vice President of a web company is criticizing the hardware guys in two of the world's biggest chip makers?</p></div></blockquote><p>Wrong.</p><p>He is criticizing, in the bits in TFS, two groups:<br>1) The <i>marketing</i> guys in two of the world's biggest chip makers (he's not complaining that the chips are flawed from an engineering perspective, he is complaining about the claims, which apparently conflict with Facebooks experience in testing them chips, about the performance of the chips), and<br>2) The people setting the design goals (not, again, the engineers) at the companies making servers, complaining that they are doing a bad job of what he sees as a major need (which is, of course, also the particular thing that Facebook needs), and that Google does a better job of building servers for that need (a complaint which would be more effective at changing behavior at server manufacturers if it was followed up by Facebook going to Google to get Google to build them servers.)</p><blockquote><div><p>Can we get like a panel of hardware engineers to have a discussion with this guy and can I get some popcorn?</p></div></blockquote><p>Why? His complaints aren't directed at engineers.</p></div>
	</htmltext>
<tokenext>So let me get this straight , the Vice President of a web company is criticizing the hardware guys in two of the world 's biggest chip makers ? Wrong.He is criticizing , in the bits in TFS , two groups : 1 ) The marketing guys in two of the world 's biggest chip makers ( he 's not complaining that the chips are flawed from an engineering perspective , he is complaining about the claims , which apparently conflict with Facebooks experience in testing them chips , about the performance of the chips ) , and2 ) The people setting the design goals ( not , again , the engineers ) at the companies making servers , complaining that they are doing a bad job of what he sees as a major need ( which is , of course , also the particular thing that Facebook needs ) , and that Google does a better job of building servers for that need ( a complaint which would be more effective at changing behavior at server manufacturers if it was followed up by Facebook going to Google to get Google to build them servers .
) Can we get like a panel of hardware engineers to have a discussion with this guy and can I get some popcorn ? Why ?
His complaints are n't directed at engineers .</tokentext>
<sentencetext>So let me get this straight, the Vice President of a web company is criticizing the hardware guys in two of the world's biggest chip makers?Wrong.He is criticizing, in the bits in TFS, two groups:1) The marketing guys in two of the world's biggest chip makers (he's not complaining that the chips are flawed from an engineering perspective, he is complaining about the claims, which apparently conflict with Facebooks experience in testing them chips, about the performance of the chips), and2) The people setting the design goals (not, again, the engineers) at the companies making servers, complaining that they are doing a bad job of what he sees as a major need (which is, of course, also the particular thing that Facebook needs), and that Google does a better job of building servers for that need (a complaint which would be more effective at changing behavior at server manufacturers if it was followed up by Facebook going to Google to get Google to build them servers.
)Can we get like a panel of hardware engineers to have a discussion with this guy and can I get some popcorn?Why?
His complaints aren't directed at engineers.
	</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476905</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477785</id>
	<title>Re:You're Computin' for a Shootin' Mister</title>
	<author>Anonymous</author>
	<datestamp>1245954240000</datestamp>
	<modclass>Troll</modclass>
	<modscore>-1</modscore>
	<htmltext>My name is <b>eldavojohn</b>, I am a frequent fellator of slashdot editors and am a seven year old black Labrador dog. My owner is a young man called Rob Malda. He has black hair and wiry legs and I have lived with him since I was three months old. I love him very much and would never bite him. He has a rough sex life and on a couple of occasions I have to make <b>stupid, uninspired, often plagiarized first post</b> jokes with a couple of men to get my bowl of Alpo.<br> <br>

My favorite time is when Rob gets my leash out and takes me for a walk. Awhile back, Rob took me for a walk in the local park where I smelt in the air that lovely smell of a bitch in heat. I followed my nose and sure enough there was a human woman named Kathleen wearing a dress.<br> <br>

I ran round her sniffing as I went, when I got near her butt that lovely smell was at its strongest. It was a gorgeous aroma, and my penis began to unsheath itself before Rob gave me a playful but firm kick in the ribs, letting me know that I shouldn't sniff human women the way I sniff other dogs.<br> <br>

When we got home Rob fed me, and after a drink I laid down and went to sleep, but was awakened shortly afterward by noises coming out of the living room. Rob and Kathleen were on the sofa, Rob bent over and wearing a small pair of Fruit-of-the-Loom chonies and a matching cotton bra, Kathleen sitting next to him wearing a short skirt with no panties and a tank top with no bra. Kathleen had her hand in Rob's chonies and I could see that her hand was moving in them, but I could barely see that her other hand was doing something to Rob's ass. The steamy air carried the same scent I had smelled back at the park. That smell was driving me crazy with lust.<br> <br>

I went towards Kathleen, watching her as she pushed two fingers deep into Rob's behind, and her eyes where closed as she writhed in concert with Rob. I went over to Kathleen right before she opened her eyes and she squealed delightfully while Rob yelled,"<b>Bad boy, eldavojohn</b>!" She pushed my head away before I could get a taste.<br> <br>

I went back to my bed, not understanding what I had done wrong, I was only doing what nature told me to do when a bitch is in heat.<br> <br>

Kathleen kept looking in my direction before she told Rob, "Oh, the poor dog. Why don't you let him join us, Rob?" "Here, boy!" Rob then said. I got up and approached the smell of her scent when she pulled her fingers out of her panties and offered them to me. The taste of her love juice was decadent, the best I have ever tasted. It was better than that time Rob let me have a piece of his dinner steak.<br> <br>

Kathleen then moved her hand back up to Rob's ass then bent over rob as if she was doing him like dogs do. I saw an opportunity and climbed on top of her, mounting her as she mounted Rob. A few wayward thrusts and I was inside her. As this happened she let out a moan of pleasure and her body shuddered.<br> <br>

The further I went into her the more Kathleen thrashed about and moaned in unison with Rob. she had one hand in Rob's ass while her other hand was pinching and pulling her nipples. Their bodies shook and convulsed as all three of us came together. We stayed like this until my cock shrunk and with a loud plop fell out. Kathleen and Rob both got up and Rob put on a robe. Kathleen came across to me and hugged me tight, telling me that I was a good fuck.<br> <br>

Rob said that it wouldn't be the last, and it wasn't.</htmltext>
<tokenext>My name is eldavojohn , I am a frequent fellator of slashdot editors and am a seven year old black Labrador dog .
My owner is a young man called Rob Malda .
He has black hair and wiry legs and I have lived with him since I was three months old .
I love him very much and would never bite him .
He has a rough sex life and on a couple of occasions I have to make stupid , uninspired , often plagiarized first post jokes with a couple of men to get my bowl of Alpo .
My favorite time is when Rob gets my leash out and takes me for a walk .
Awhile back , Rob took me for a walk in the local park where I smelt in the air that lovely smell of a bitch in heat .
I followed my nose and sure enough there was a human woman named Kathleen wearing a dress .
I ran round her sniffing as I went , when I got near her butt that lovely smell was at its strongest .
It was a gorgeous aroma , and my penis began to unsheath itself before Rob gave me a playful but firm kick in the ribs , letting me know that I should n't sniff human women the way I sniff other dogs .
When we got home Rob fed me , and after a drink I laid down and went to sleep , but was awakened shortly afterward by noises coming out of the living room .
Rob and Kathleen were on the sofa , Rob bent over and wearing a small pair of Fruit-of-the-Loom chonies and a matching cotton bra , Kathleen sitting next to him wearing a short skirt with no panties and a tank top with no bra .
Kathleen had her hand in Rob 's chonies and I could see that her hand was moving in them , but I could barely see that her other hand was doing something to Rob 's ass .
The steamy air carried the same scent I had smelled back at the park .
That smell was driving me crazy with lust .
I went towards Kathleen , watching her as she pushed two fingers deep into Rob 's behind , and her eyes where closed as she writhed in concert with Rob .
I went over to Kathleen right before she opened her eyes and she squealed delightfully while Rob yelled , " Bad boy , eldavojohn !
" She pushed my head away before I could get a taste .
I went back to my bed , not understanding what I had done wrong , I was only doing what nature told me to do when a bitch is in heat .
Kathleen kept looking in my direction before she told Rob , " Oh , the poor dog .
Why do n't you let him join us , Rob ?
" " Here , boy !
" Rob then said .
I got up and approached the smell of her scent when she pulled her fingers out of her panties and offered them to me .
The taste of her love juice was decadent , the best I have ever tasted .
It was better than that time Rob let me have a piece of his dinner steak .
Kathleen then moved her hand back up to Rob 's ass then bent over rob as if she was doing him like dogs do .
I saw an opportunity and climbed on top of her , mounting her as she mounted Rob .
A few wayward thrusts and I was inside her .
As this happened she let out a moan of pleasure and her body shuddered .
The further I went into her the more Kathleen thrashed about and moaned in unison with Rob .
she had one hand in Rob 's ass while her other hand was pinching and pulling her nipples .
Their bodies shook and convulsed as all three of us came together .
We stayed like this until my cock shrunk and with a loud plop fell out .
Kathleen and Rob both got up and Rob put on a robe .
Kathleen came across to me and hugged me tight , telling me that I was a good fuck .
Rob said that it would n't be the last , and it was n't .</tokentext>
<sentencetext>My name is eldavojohn, I am a frequent fellator of slashdot editors and am a seven year old black Labrador dog.
My owner is a young man called Rob Malda.
He has black hair and wiry legs and I have lived with him since I was three months old.
I love him very much and would never bite him.
He has a rough sex life and on a couple of occasions I have to make stupid, uninspired, often plagiarized first post jokes with a couple of men to get my bowl of Alpo.
My favorite time is when Rob gets my leash out and takes me for a walk.
Awhile back, Rob took me for a walk in the local park where I smelt in the air that lovely smell of a bitch in heat.
I followed my nose and sure enough there was a human woman named Kathleen wearing a dress.
I ran round her sniffing as I went, when I got near her butt that lovely smell was at its strongest.
It was a gorgeous aroma, and my penis began to unsheath itself before Rob gave me a playful but firm kick in the ribs, letting me know that I shouldn't sniff human women the way I sniff other dogs.
When we got home Rob fed me, and after a drink I laid down and went to sleep, but was awakened shortly afterward by noises coming out of the living room.
Rob and Kathleen were on the sofa, Rob bent over and wearing a small pair of Fruit-of-the-Loom chonies and a matching cotton bra, Kathleen sitting next to him wearing a short skirt with no panties and a tank top with no bra.
Kathleen had her hand in Rob's chonies and I could see that her hand was moving in them, but I could barely see that her other hand was doing something to Rob's ass.
The steamy air carried the same scent I had smelled back at the park.
That smell was driving me crazy with lust.
I went towards Kathleen, watching her as she pushed two fingers deep into Rob's behind, and her eyes where closed as she writhed in concert with Rob.
I went over to Kathleen right before she opened her eyes and she squealed delightfully while Rob yelled,"Bad boy, eldavojohn!
" She pushed my head away before I could get a taste.
I went back to my bed, not understanding what I had done wrong, I was only doing what nature told me to do when a bitch is in heat.
Kathleen kept looking in my direction before she told Rob, "Oh, the poor dog.
Why don't you let him join us, Rob?
" "Here, boy!
" Rob then said.
I got up and approached the smell of her scent when she pulled her fingers out of her panties and offered them to me.
The taste of her love juice was decadent, the best I have ever tasted.
It was better than that time Rob let me have a piece of his dinner steak.
Kathleen then moved her hand back up to Rob's ass then bent over rob as if she was doing him like dogs do.
I saw an opportunity and climbed on top of her, mounting her as she mounted Rob.
A few wayward thrusts and I was inside her.
As this happened she let out a moan of pleasure and her body shuddered.
The further I went into her the more Kathleen thrashed about and moaned in unison with Rob.
she had one hand in Rob's ass while her other hand was pinching and pulling her nipples.
Their bodies shook and convulsed as all three of us came together.
We stayed like this until my cock shrunk and with a loud plop fell out.
Kathleen and Rob both got up and Rob put on a robe.
Kathleen came across to me and hugged me tight, telling me that I was a good fuck.
Rob said that it wouldn't be the last, and it wasn't.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476905</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479573</id>
	<title>Facebook should buy SiCortex</title>
	<author>White Flame</author>
	<datestamp>1246013640000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext><p>They're on the block, and produce incredibly power-efficient, inexpensive (complete systems for ~$200/core to the end user) machines.  Their sales were continually gaining momentum, but their investors hit troubled times and had to pull out.</p></htmltext>
<tokenext>They 're on the block , and produce incredibly power-efficient , inexpensive ( complete systems for ~ $ 200/core to the end user ) machines .
Their sales were continually gaining momentum , but their investors hit troubled times and had to pull out .</tokentext>
<sentencetext>They're on the block, and produce incredibly power-efficient, inexpensive (complete systems for ~$200/core to the end user) machines.
Their sales were continually gaining momentum, but their investors hit troubled times and had to pull out.</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28483819</id>
	<title>Re:Facebook's application is poorly coded</title>
	<author>MikeBabcock</author>
	<datestamp>1246035360000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext><p>I've seen crap code that doesn't improve much with better hardware.  Things like:</p><p>for i in blah.length()<br>{<br>
&nbsp; &nbsp; data += blah<br>
&nbsp; &nbsp; f.write(data)<br>
&nbsp; &nbsp; while (f.writing()) sleep(500)<br>}</p><p>As another example, I was reading the driver code for a PCI serial device I have and it loops through all 128 ports looking for waiting data then rests and does it again and ignores the hardware interrupts.</p></htmltext>
<tokenext>I 've seen crap code that does n't improve much with better hardware .
Things like : for i in blah.length ( ) {     data + = blah     f.write ( data )     while ( f.writing ( ) ) sleep ( 500 ) } As another example , I was reading the driver code for a PCI serial device I have and it loops through all 128 ports looking for waiting data then rests and does it again and ignores the hardware interrupts .</tokentext>
<sentencetext>I've seen crap code that doesn't improve much with better hardware.
Things like:for i in blah.length(){
    data += blah
    f.write(data)
    while (f.writing()) sleep(500)}As another example, I was reading the driver code for a PCI serial device I have and it loops through all 128 ports looking for waiting data then rests and does it again and ignores the hardware interrupts.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477167</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476943</id>
	<title>YouTube...</title>
	<author>Anonymous</author>
	<datestamp>1245947280000</datestamp>
	<modclass>None</modclass>
	<modscore>0</modscore>
	<htmltext><p>YouTube has been constantly crashing and showing server errors more frequently over the last month. I'm not sure how much more bandwidth that YouTube demands over Facebook or Amazon (if it does) but whatever issues with servers and reliability that the Facebook team is having with Intel and AMD might be even worse over at YouTube.</p><p>Anyone experiencing major outages and slow movement and loading times over at YouTube recently? Can't say I've had a problem with Amazon or Facebook other than intended outages for updates or repairs in the last month.</p></htmltext>
<tokenext>YouTube has been constantly crashing and showing server errors more frequently over the last month .
I 'm not sure how much more bandwidth that YouTube demands over Facebook or Amazon ( if it does ) but whatever issues with servers and reliability that the Facebook team is having with Intel and AMD might be even worse over at YouTube.Anyone experiencing major outages and slow movement and loading times over at YouTube recently ?
Ca n't say I 've had a problem with Amazon or Facebook other than intended outages for updates or repairs in the last month .</tokentext>
<sentencetext>YouTube has been constantly crashing and showing server errors more frequently over the last month.
I'm not sure how much more bandwidth that YouTube demands over Facebook or Amazon (if it does) but whatever issues with servers and reliability that the Facebook team is having with Intel and AMD might be even worse over at YouTube.Anyone experiencing major outages and slow movement and loading times over at YouTube recently?
Can't say I've had a problem with Amazon or Facebook other than intended outages for updates or repairs in the last month.</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28480043</id>
	<title>Re:Hm...</title>
	<author>Anonymous</author>
	<datestamp>1246019700000</datestamp>
	<modclass>None</modclass>
	<modscore>0</modscore>
	<htmltext><p>NSA domestic spying programs don't need a revenue model.  It's called taxes.  Why try to build a mind-reading machine when you can do the next-best thing?  People are only too happy to dump their entire lives into Facebook, where the NSA mines the data for otherwise obscure relationships.  When you disappear suddenly, no one will ask where you went.  They will know the answer, but will be too terrified to verbalize it.  If you follow this one rule, though, you can avoid winding up in a Wackenhut gulag: Dissent is only patriotic if it's not directed at our dear mullah Obama, peace be upon him.</p></htmltext>
<tokenext>NSA domestic spying programs do n't need a revenue model .
It 's called taxes .
Why try to build a mind-reading machine when you can do the next-best thing ?
People are only too happy to dump their entire lives into Facebook , where the NSA mines the data for otherwise obscure relationships .
When you disappear suddenly , no one will ask where you went .
They will know the answer , but will be too terrified to verbalize it .
If you follow this one rule , though , you can avoid winding up in a Wackenhut gulag : Dissent is only patriotic if it 's not directed at our dear mullah Obama , peace be upon him .</tokentext>
<sentencetext>NSA domestic spying programs don't need a revenue model.
It's called taxes.
Why try to build a mind-reading machine when you can do the next-best thing?
People are only too happy to dump their entire lives into Facebook, where the NSA mines the data for otherwise obscure relationships.
When you disappear suddenly, no one will ask where you went.
They will know the answer, but will be too terrified to verbalize it.
If you follow this one rule, though, you can avoid winding up in a Wackenhut gulag: Dissent is only patriotic if it's not directed at our dear mullah Obama, peace be upon him.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477211</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478127</id>
	<title>Re:Facebook's application is poorly coded</title>
	<author>grantek</author>
	<datestamp>1245957180000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext><p>Not necessarily - for example, if you look at the PC boot sequence, people have been screaming that there are too many hard-coded sleep periods that don't get smaller as processors get faster. It's feasible there's similar locking problems (especially if it's associated with I/O) in a web app.</p></htmltext>
<tokenext>Not necessarily - for example , if you look at the PC boot sequence , people have been screaming that there are too many hard-coded sleep periods that do n't get smaller as processors get faster .
It 's feasible there 's similar locking problems ( especially if it 's associated with I/O ) in a web app .</tokentext>
<sentencetext>Not necessarily - for example, if you look at the PC boot sequence, people have been screaming that there are too many hard-coded sleep periods that don't get smaller as processors get faster.
It's feasible there's similar locking problems (especially if it's associated with I/O) in a web app.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477085</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28482821</id>
	<title>Re:Something about his arguement doesn't work</title>
	<author>happyhangone</author>
	<datestamp>1246032060000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext><p>Well, if they are bitching about AMD and Intel, he can call Dell's DCS and buy <a href="http://www.greenm3.com/2009/05/xs11-vx8-dells-via-nano-server-anti-vmware-and-intel.html" title="greenm3.com" rel="nofollow">VIA systems</a> [greenm3.com] or just buy the systems that Amazon, Google and others are purchasing (<a href="http://en.community.dell.com/blogs/direct2dell/archive/2009/03/16/dell-s-cloud-servers-keep-getting-stronger.aspx" title="dell.com" rel="nofollow">using Intel and AMD</a> [dell.com])</p></htmltext>
<tokenext>Well , if they are bitching about AMD and Intel , he can call Dell 's DCS and buy VIA systems [ greenm3.com ] or just buy the systems that Amazon , Google and others are purchasing ( using Intel and AMD [ dell.com ] )</tokentext>
<sentencetext>Well, if they are bitching about AMD and Intel, he can call Dell's DCS and buy VIA systems [greenm3.com] or just buy the systems that Amazon, Google and others are purchasing (using Intel and AMD [dell.com])</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476989</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476967</id>
	<title>Well I suppose...</title>
	<author>cptnapalm</author>
	<datestamp>1245947460000</datestamp>
	<modclass>Funny</modclass>
	<modscore>3</modscore>
	<htmltext><p>Well, I suppose that if he does not like the offerings from Intel and AMD, they could always go with...</p><p>Uh..</p><p>Oh.</p></htmltext>
<tokenext>Well , I suppose that if he does not like the offerings from Intel and AMD , they could always go with...Uh..Oh .</tokentext>
<sentencetext>Well, I suppose that if he does not like the offerings from Intel and AMD, they could always go with...Uh..Oh.</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28480007</id>
	<title>Re:A Familiar Tune from Facebook</title>
	<author>cheekyboy</author>
	<datestamp>1246019160000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext><p>lets see , 1 motherboard + cpu &amp; ram, and one disk (cheapest).</p><p>Not bad, just 4 items to make a PC, the power can be worked out later.</p><p>Design custom cases/racks, outsource the tool work, bingo, not hard at all, each PC could be under $150.</p></htmltext>
<tokenext>lets see , 1 motherboard + cpu &amp; ram , and one disk ( cheapest ) .Not bad , just 4 items to make a PC , the power can be worked out later.Design custom cases/racks , outsource the tool work , bingo , not hard at all , each PC could be under $ 150 .</tokentext>
<sentencetext>lets see , 1 motherboard + cpu &amp; ram, and one disk (cheapest).Not bad, just 4 items to make a PC, the power can be worked out later.Design custom cases/racks, outsource the tool work, bingo, not hard at all, each PC could be under $150.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477189</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477411</id>
	<title>Re:Hm...</title>
	<author>HockeyPuck</author>
	<datestamp>1245950700000</datestamp>
	<modclass>Informativ</modclass>
	<modscore>5</modscore>
	<htmltext><blockquote><div><p>The average server is going to be much more I/O intensive than CPU intensive unless you do cluster computing or render a lot of stuff.</p></div></blockquote><p>As someone who designs and deploys large storage environments for a living, I call BS.  While the current generation of HBAs are 8Gb FibreChannel, I would say that the "average server" (as you put it) could happily live on a 1Gb HBA.  Recall that almost all servers, or atleast those you care about, have DUAL HBA connections to their respective storage.  So that's actually 2Gb of storage connectivity.  Sure there are servers which have multiple HBAs, or use a higher utilization of the HBAs, such as database servers or backup/media servers.  Most servers today are deployed with dual 4Gb HBAs as the 8Gb SFPs/optics are still quite pricey, and you cannot, in all seriousness, purchase 1 or 2Gb FC HBAs.</p><p>Even as we deploy VMware based servers, the VMware servers themselves tend to be more memory/cpu strapped than IO.</p><p>It would be very rare, or almost impossible for a server to be driving linerate HBAs, with still plenty of headroom left in the CPU.  Even basic test tools like IOmeter require significant CPU usage to drive an HBA to capacity.  And that is when it's writing/reading all zeros. It's doesn't actually need to do anything with the data. As would be the case if a database server was requesting 2Gb/s from a disk array, and then had to join/sort/add/whatever the tables retrieved.</p><p>
&nbsp;</p></div>
	</htmltext>
<tokenext>The average server is going to be much more I/O intensive than CPU intensive unless you do cluster computing or render a lot of stuff.As someone who designs and deploys large storage environments for a living , I call BS .
While the current generation of HBAs are 8Gb FibreChannel , I would say that the " average server " ( as you put it ) could happily live on a 1Gb HBA .
Recall that almost all servers , or atleast those you care about , have DUAL HBA connections to their respective storage .
So that 's actually 2Gb of storage connectivity .
Sure there are servers which have multiple HBAs , or use a higher utilization of the HBAs , such as database servers or backup/media servers .
Most servers today are deployed with dual 4Gb HBAs as the 8Gb SFPs/optics are still quite pricey , and you can not , in all seriousness , purchase 1 or 2Gb FC HBAs.Even as we deploy VMware based servers , the VMware servers themselves tend to be more memory/cpu strapped than IO.It would be very rare , or almost impossible for a server to be driving linerate HBAs , with still plenty of headroom left in the CPU .
Even basic test tools like IOmeter require significant CPU usage to drive an HBA to capacity .
And that is when it 's writing/reading all zeros .
It 's does n't actually need to do anything with the data .
As would be the case if a database server was requesting 2Gb/s from a disk array , and then had to join/sort/add/whatever the tables retrieved .
 </tokentext>
<sentencetext>The average server is going to be much more I/O intensive than CPU intensive unless you do cluster computing or render a lot of stuff.As someone who designs and deploys large storage environments for a living, I call BS.
While the current generation of HBAs are 8Gb FibreChannel, I would say that the "average server" (as you put it) could happily live on a 1Gb HBA.
Recall that almost all servers, or atleast those you care about, have DUAL HBA connections to their respective storage.
So that's actually 2Gb of storage connectivity.
Sure there are servers which have multiple HBAs, or use a higher utilization of the HBAs, such as database servers or backup/media servers.
Most servers today are deployed with dual 4Gb HBAs as the 8Gb SFPs/optics are still quite pricey, and you cannot, in all seriousness, purchase 1 or 2Gb FC HBAs.Even as we deploy VMware based servers, the VMware servers themselves tend to be more memory/cpu strapped than IO.It would be very rare, or almost impossible for a server to be driving linerate HBAs, with still plenty of headroom left in the CPU.
Even basic test tools like IOmeter require significant CPU usage to drive an HBA to capacity.
And that is when it's writing/reading all zeros.
It's doesn't actually need to do anything with the data.
As would be the case if a database server was requesting 2Gb/s from a disk array, and then had to join/sort/add/whatever the tables retrieved.
 
	</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476941</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478305</id>
	<title>Re:WTF?</title>
	<author>Anonymous</author>
	<datestamp>1245959040000</datestamp>
	<modclass>Interestin</modclass>
	<modscore>1</modscore>
	<htmltext><p>At the end of your long comment u said "If I was Intel/AMD I'd be chiming in right about now and opening a dialog with Facebook and looking to see what the issues are. Facebook is a big customer with huge name recognition and you want to be able to use them as an example of your solution delivering the promised performance for your marketing."</p><p>U know the funny thing is...back in the day, lets say Pentium 1, 2 and 3 days. AMD had awesome products, and no one was buying because they had no idea who these guys were. (Example. AMD was first to make 1 GHz CPU and it was faster per clock then Intel's offerings). So what AMD did is they went out of their way to work with customers, to get sales, and they slowly built up a name for themselves...its really funny that after they reached the top (remember athlon VS Pentium when AMD was without a doubt better), they "forget" to do the things that got them where they are, and wonder why they are dieing.</p></htmltext>
<tokenext>At the end of your long comment u said " If I was Intel/AMD I 'd be chiming in right about now and opening a dialog with Facebook and looking to see what the issues are .
Facebook is a big customer with huge name recognition and you want to be able to use them as an example of your solution delivering the promised performance for your marketing .
" U know the funny thing is...back in the day , lets say Pentium 1 , 2 and 3 days .
AMD had awesome products , and no one was buying because they had no idea who these guys were .
( Example. AMD was first to make 1 GHz CPU and it was faster per clock then Intel 's offerings ) .
So what AMD did is they went out of their way to work with customers , to get sales , and they slowly built up a name for themselves...its really funny that after they reached the top ( remember athlon VS Pentium when AMD was without a doubt better ) , they " forget " to do the things that got them where they are , and wonder why they are dieing .</tokentext>
<sentencetext>At the end of your long comment u said "If I was Intel/AMD I'd be chiming in right about now and opening a dialog with Facebook and looking to see what the issues are.
Facebook is a big customer with huge name recognition and you want to be able to use them as an example of your solution delivering the promised performance for your marketing.
"U know the funny thing is...back in the day, lets say Pentium 1, 2 and 3 days.
AMD had awesome products, and no one was buying because they had no idea who these guys were.
(Example. AMD was first to make 1 GHz CPU and it was faster per clock then Intel's offerings).
So what AMD did is they went out of their way to work with customers, to get sales, and they slowly built up a name for themselves...its really funny that after they reached the top (remember athlon VS Pentium when AMD was without a doubt better), they "forget" to do the things that got them where they are, and wonder why they are dieing.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477361</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28481227</id>
	<title>Re:Hm...</title>
	<author>MightyDrunken</author>
	<datestamp>1246027080000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext><p><div class="quote"><p>Even basic test tools like IOmeter require significant CPU usage to drive an HBA to capacity. And that is when it's writing/reading all zeros. It's doesn't actually need to do anything with the data. As would be the case if a database server was requesting 2Gb/s from a disk array, and then had to join/sort/add/whatever the tables retrieved.</p></div><p>What is the CPU doing? Is the memory controller/CPU overwhelmed with data?

With say 1Gb of I/O how does the seek time affect the performance of the work these servers are doing?</p></div>
	</htmltext>
<tokenext>Even basic test tools like IOmeter require significant CPU usage to drive an HBA to capacity .
And that is when it 's writing/reading all zeros .
It 's does n't actually need to do anything with the data .
As would be the case if a database server was requesting 2Gb/s from a disk array , and then had to join/sort/add/whatever the tables retrieved.What is the CPU doing ?
Is the memory controller/CPU overwhelmed with data ?
With say 1Gb of I/O how does the seek time affect the performance of the work these servers are doing ?</tokentext>
<sentencetext>Even basic test tools like IOmeter require significant CPU usage to drive an HBA to capacity.
And that is when it's writing/reading all zeros.
It's doesn't actually need to do anything with the data.
As would be the case if a database server was requesting 2Gb/s from a disk array, and then had to join/sort/add/whatever the tables retrieved.What is the CPU doing?
Is the memory controller/CPU overwhelmed with data?
With say 1Gb of I/O how does the seek time affect the performance of the work these servers are doing?
	</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477411</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477675</id>
	<title>Would you expect otherwise</title>
	<author>Chuck Chunder</author>
	<datestamp>1245953220000</datestamp>
	<modclass>Funny</modclass>
	<modscore>4</modscore>
	<htmltext><blockquote><div><p>If that's his argument, then it would seem that the real conclusion is that Facebook can't build systems as good as Google's</p></div></blockquote><p>

Google's core business is intelligence.<br>
Facebooks core business is stupidity.</p></div>
	</htmltext>
<tokenext>If that 's his argument , then it would seem that the real conclusion is that Facebook ca n't build systems as good as Google 's Google 's core business is intelligence .
Facebooks core business is stupidity .</tokentext>
<sentencetext>If that's his argument, then it would seem that the real conclusion is that Facebook can't build systems as good as Google's

Google's core business is intelligence.
Facebooks core business is stupidity.
	</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476989</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28486667</id>
	<title>Re:Strange...</title>
	<author>Anonymous</author>
	<datestamp>1246046880000</datestamp>
	<modclass>None</modclass>
	<modscore>0</modscore>
	<htmltext><p>This is Facebook we're talking about. Their code is a mess of PHP on top of MySQL that simply doesn't scale. The presentation on their infrastructure a few months ago was almost funny. IIRC, they have 5000 servers to handle 30,000 requests per second. And they've even had to jump through hoops to get that to work. For instance, they don't do database joins because they can't get replication to work in a timely fashion, so different data resides on different database clusters. I would be surprised if they were able to simulate their actual production environment on a smaller scale. They probably did do testing, but not with their actual application code, so machines that may have passed their stress testing could be under-performing once they reach the production environment.</p><p>So they're left in a situation where they have no revenue to speak of and run heavily in the red and can only scale by buying more and more hardware. They've got a decent size chunk of cash in the bank, but operational costs are eating through it quite quickly. So rather than (or possibly in addition to) re-architecting their application to scale better, they're publicly chastising server makers for not advancing at a rate that will bail them out of their current mess.</p><p>What Facebook needs to do is re-write their application in a technology that scales and run it on technology that scales. It would be a difficult process, but in the long run, it's what they need to survive. The logical choice would be, IMHO Java. I may be biased since I work on a Java-based application, but I know it can scale since we've reached 5000 requests per second on 5 app servers and 2 database servers while maintaining sub-second page generation times on all but the most involved pages. The load at the time was about 50\% on the database servers and around 30\% on the app servers. FWIW, our application is considerably more complex that Facebook's is, though less bandwidth intensive.</p></htmltext>
<tokenext>This is Facebook we 're talking about .
Their code is a mess of PHP on top of MySQL that simply does n't scale .
The presentation on their infrastructure a few months ago was almost funny .
IIRC , they have 5000 servers to handle 30,000 requests per second .
And they 've even had to jump through hoops to get that to work .
For instance , they do n't do database joins because they ca n't get replication to work in a timely fashion , so different data resides on different database clusters .
I would be surprised if they were able to simulate their actual production environment on a smaller scale .
They probably did do testing , but not with their actual application code , so machines that may have passed their stress testing could be under-performing once they reach the production environment.So they 're left in a situation where they have no revenue to speak of and run heavily in the red and can only scale by buying more and more hardware .
They 've got a decent size chunk of cash in the bank , but operational costs are eating through it quite quickly .
So rather than ( or possibly in addition to ) re-architecting their application to scale better , they 're publicly chastising server makers for not advancing at a rate that will bail them out of their current mess.What Facebook needs to do is re-write their application in a technology that scales and run it on technology that scales .
It would be a difficult process , but in the long run , it 's what they need to survive .
The logical choice would be , IMHO Java .
I may be biased since I work on a Java-based application , but I know it can scale since we 've reached 5000 requests per second on 5 app servers and 2 database servers while maintaining sub-second page generation times on all but the most involved pages .
The load at the time was about 50 \ % on the database servers and around 30 \ % on the app servers .
FWIW , our application is considerably more complex that Facebook 's is , though less bandwidth intensive .</tokentext>
<sentencetext>This is Facebook we're talking about.
Their code is a mess of PHP on top of MySQL that simply doesn't scale.
The presentation on their infrastructure a few months ago was almost funny.
IIRC, they have 5000 servers to handle 30,000 requests per second.
And they've even had to jump through hoops to get that to work.
For instance, they don't do database joins because they can't get replication to work in a timely fashion, so different data resides on different database clusters.
I would be surprised if they were able to simulate their actual production environment on a smaller scale.
They probably did do testing, but not with their actual application code, so machines that may have passed their stress testing could be under-performing once they reach the production environment.So they're left in a situation where they have no revenue to speak of and run heavily in the red and can only scale by buying more and more hardware.
They've got a decent size chunk of cash in the bank, but operational costs are eating through it quite quickly.
So rather than (or possibly in addition to) re-architecting their application to scale better, they're publicly chastising server makers for not advancing at a rate that will bail them out of their current mess.What Facebook needs to do is re-write their application in a technology that scales and run it on technology that scales.
It would be a difficult process, but in the long run, it's what they need to survive.
The logical choice would be, IMHO Java.
I may be biased since I work on a Java-based application, but I know it can scale since we've reached 5000 requests per second on 5 app servers and 2 database servers while maintaining sub-second page generation times on all but the most involved pages.
The load at the time was about 50\% on the database servers and around 30\% on the app servers.
FWIW, our application is considerably more complex that Facebook's is, though less bandwidth intensive.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477581</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28480349</id>
	<title>Re:Hm...</title>
	<author>ezzzD55J</author>
	<datestamp>1246022880000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext><p><div class="quote"><p>Tell me again what Facebook's revenue model is...??</p></div><p>i can't answer that very well, but i know they have a lot of money.</p></div>
	</htmltext>
<tokenext>Tell me again what Facebook 's revenue model is... ?
? i ca n't answer that very well , but i know they have a lot of money .</tokentext>
<sentencetext>Tell me again what Facebook's revenue model is...?
?i can't answer that very well, but i know they have a lot of money.
	</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477211</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477699</id>
	<title>co34</title>
	<author>Anonymous</author>
	<datestamp>1245953460000</datestamp>
	<modclass>Redundant</modclass>
	<modscore>-1</modscore>
	<htmltext>From a techn1cal</htmltext>
<tokenext>From a techn1cal</tokentext>
<sentencetext>From a techn1cal</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477187</id>
	<title>Re:Facebook's application is poorly coded</title>
	<author>Anonymous</author>
	<datestamp>1245948960000</datestamp>
	<modclass>None</modclass>
	<modscore>0</modscore>
	<htmltext><p>Crap code on faster computer is still going to be faster than it was on a slower computer</p><p>That is only partially true.  But what if they have algs in there that are O(N^N)?  Say N is the number of users?  And they are adding 1000 new users per day?</p><p>So as they add more users the 'better' hardware does NOTHING for them.  It sounds like Intel or AMD reps promised them the moon and they bought the story.  I would say perhaps he was sold a bill of goods.  But guess what without metrics to back it up.  Like how does it run with HIS software I would say he didnt do his job.</p><p>I would also say a bad alg on better hardware will eventually perform just as bad once you add more users.  They are trying to throw more hardware at the problem without understanding the problem.  Or perhaps they do but are unwilling/unable to change something?</p></htmltext>
<tokenext>Crap code on faster computer is still going to be faster than it was on a slower computerThat is only partially true .
But what if they have algs in there that are O ( N ^ N ) ?
Say N is the number of users ?
And they are adding 1000 new users per day ? So as they add more users the 'better ' hardware does NOTHING for them .
It sounds like Intel or AMD reps promised them the moon and they bought the story .
I would say perhaps he was sold a bill of goods .
But guess what without metrics to back it up .
Like how does it run with HIS software I would say he didnt do his job.I would also say a bad alg on better hardware will eventually perform just as bad once you add more users .
They are trying to throw more hardware at the problem without understanding the problem .
Or perhaps they do but are unwilling/unable to change something ?</tokentext>
<sentencetext>Crap code on faster computer is still going to be faster than it was on a slower computerThat is only partially true.
But what if they have algs in there that are O(N^N)?
Say N is the number of users?
And they are adding 1000 new users per day?So as they add more users the 'better' hardware does NOTHING for them.
It sounds like Intel or AMD reps promised them the moon and they bought the story.
I would say perhaps he was sold a bill of goods.
But guess what without metrics to back it up.
Like how does it run with HIS software I would say he didnt do his job.I would also say a bad alg on better hardware will eventually perform just as bad once you add more users.
They are trying to throw more hardware at the problem without understanding the problem.
Or perhaps they do but are unwilling/unable to change something?</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477085</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479953</id>
	<title>Re:Sounds like a bunch of excuses to me</title>
	<author>cheekyboy</author>
	<datestamp>1246018680000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext><p>I pick Utter Performance, then I pick Cheap Power + Cheap Computers.</p><p>Who cares if it eats tonnes of cabon if its CHEAP, that equals profits.</p></htmltext>
<tokenext>I pick Utter Performance , then I pick Cheap Power + Cheap Computers.Who cares if it eats tonnes of cabon if its CHEAP , that equals profits .</tokentext>
<sentencetext>I pick Utter Performance, then I pick Cheap Power + Cheap Computers.Who cares if it eats tonnes of cabon if its CHEAP, that equals profits.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477115</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479487</id>
	<title>Performance issues, eh.</title>
	<author>jjgm</author>
	<datestamp>1246012980000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext>PALO ALTO (Reuters). PHP-based website reports scalability problems. Blames server hardware. Film at 11.</htmltext>
<tokenext>PALO ALTO ( Reuters ) .
PHP-based website reports scalability problems .
Blames server hardware .
Film at 11 .</tokentext>
<sentencetext>PALO ALTO (Reuters).
PHP-based website reports scalability problems.
Blames server hardware.
Film at 11.</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477081</id>
	<title>Re:You're Computin' for a Shootin' Mister</title>
	<author>Anonymous</author>
	<datestamp>1245948240000</datestamp>
	<modclass>Insightful</modclass>
	<modscore>2</modscore>
	<htmltext>You've basically agreed with him. Whats your bitch? I don't get it.</htmltext>
<tokenext>You 've basically agreed with him .
Whats your bitch ?
I do n't get it .</tokentext>
<sentencetext>You've basically agreed with him.
Whats your bitch?
I don't get it.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476905</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477751</id>
	<title>Re:WTF?</title>
	<author>lukas84</author>
	<datestamp>1245954060000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext><p>You used slow SATA Disks and complain about low IOPS when doing random IO?</p><p>You'll be sure to get a job at Facebook<nobr> <wbr></nobr>:)</p></htmltext>
<tokenext>You used slow SATA Disks and complain about low IOPS when doing random IO ? You 'll be sure to get a job at Facebook : )</tokentext>
<sentencetext>You used slow SATA Disks and complain about low IOPS when doing random IO?You'll be sure to get a job at Facebook :)</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477335</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479977</id>
	<title>Re:Facebook's application is poorly coded</title>
	<author>Big\_Mamma</author>
	<datestamp>1246018920000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext>This really makes me doubt their ability to benchmark / scale things properly. In the article, he sounds like facebook is completely CPU bound, and yet he's slamming the latest generation server processors by Intel / AMD?<br> <br>

From all the benchmarks I've seen, like <a href="http://www.anandtech.com/cpuchipsets/intel/showdoc.aspx?i=3536&amp;p=9" title="anandtech.com" rel="nofollow">Anandtech's</a> [anandtech.com] and from personal experience, web servers scale pretty much linearly with clock speed * IPC and the amount of cores present in the system. The addition of HT is good for another 20\% throughput.<br> <br>

What they need to do is to look at their setup, and make sure there isn't another bottleneck - have you spawned enough threads and processes to utilize the system completely? PHP may be "thread safe", but that usually means that there's a huge lock around everything that could be dangerous and one process refuses to use more than 100\% cpu on 1 core, so serve it with apache-prefork + load balancer + separate static file server. Same thing for Python - fork off more copies via mod\_wsgi even in threaded mode, as many as you can afford within the available RAM, or the Global Interpreter Lock will limit the CPU usage to 1 core.<br> <br>

If you have setup the environment well and there are no other bottlenecks, web services scale perfectly with the available CPU power. And that has increased by an insane amount for the Xeon 54xx to 55xx, it's almost doubled the performance in most server apps (OLTP, VM), but even the PHP test case which failed to scale to 16 cores in a single process was good for +39\%.</htmltext>
<tokenext>This really makes me doubt their ability to benchmark / scale things properly .
In the article , he sounds like facebook is completely CPU bound , and yet he 's slamming the latest generation server processors by Intel / AMD ?
From all the benchmarks I 've seen , like Anandtech 's [ anandtech.com ] and from personal experience , web servers scale pretty much linearly with clock speed * IPC and the amount of cores present in the system .
The addition of HT is good for another 20 \ % throughput .
What they need to do is to look at their setup , and make sure there is n't another bottleneck - have you spawned enough threads and processes to utilize the system completely ?
PHP may be " thread safe " , but that usually means that there 's a huge lock around everything that could be dangerous and one process refuses to use more than 100 \ % cpu on 1 core , so serve it with apache-prefork + load balancer + separate static file server .
Same thing for Python - fork off more copies via mod \ _wsgi even in threaded mode , as many as you can afford within the available RAM , or the Global Interpreter Lock will limit the CPU usage to 1 core .
If you have setup the environment well and there are no other bottlenecks , web services scale perfectly with the available CPU power .
And that has increased by an insane amount for the Xeon 54xx to 55xx , it 's almost doubled the performance in most server apps ( OLTP , VM ) , but even the PHP test case which failed to scale to 16 cores in a single process was good for + 39 \ % .</tokentext>
<sentencetext>This really makes me doubt their ability to benchmark / scale things properly.
In the article, he sounds like facebook is completely CPU bound, and yet he's slamming the latest generation server processors by Intel / AMD?
From all the benchmarks I've seen, like Anandtech's [anandtech.com] and from personal experience, web servers scale pretty much linearly with clock speed * IPC and the amount of cores present in the system.
The addition of HT is good for another 20\% throughput.
What they need to do is to look at their setup, and make sure there isn't another bottleneck - have you spawned enough threads and processes to utilize the system completely?
PHP may be "thread safe", but that usually means that there's a huge lock around everything that could be dangerous and one process refuses to use more than 100\% cpu on 1 core, so serve it with apache-prefork + load balancer + separate static file server.
Same thing for Python - fork off more copies via mod\_wsgi even in threaded mode, as many as you can afford within the available RAM, or the Global Interpreter Lock will limit the CPU usage to 1 core.
If you have setup the environment well and there are no other bottlenecks, web services scale perfectly with the available CPU power.
And that has increased by an insane amount for the Xeon 54xx to 55xx, it's almost doubled the performance in most server apps (OLTP, VM), but even the PHP test case which failed to scale to 16 cores in a single process was good for +39\%.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476965</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477851</id>
	<title>Re:WTF?</title>
	<author>macshit</author>
	<datestamp>1245954720000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext><p>I've no idea what Intel/AMD claimed, but you'd have to be utterly naive to believe <em>any</em> claim involving a single fixed performance increase.  Modern systems are very complex, and performance is <em>insanely</em> context dependent.  Facebook's server apps are not going to be precisely like anybody else's apps, and there's simply no way to know how they'll perform without testing them.
</p><p>...and of course you wouldn't expect somebody with the title "VP of technical operations" to be so technically naive<nobr> <wbr></nobr>...<br>oh... wait.  Facebook.
</p><p>Never mind.</p></htmltext>
<tokenext>I 've no idea what Intel/AMD claimed , but you 'd have to be utterly naive to believe any claim involving a single fixed performance increase .
Modern systems are very complex , and performance is insanely context dependent .
Facebook 's server apps are not going to be precisely like anybody else 's apps , and there 's simply no way to know how they 'll perform without testing them .
...and of course you would n't expect somebody with the title " VP of technical operations " to be so technically naive ...oh... wait. Facebook .
Never mind .</tokentext>
<sentencetext>I've no idea what Intel/AMD claimed, but you'd have to be utterly naive to believe any claim involving a single fixed performance increase.
Modern systems are very complex, and performance is insanely context dependent.
Facebook's server apps are not going to be precisely like anybody else's apps, and there's simply no way to know how they'll perform without testing them.
...and of course you wouldn't expect somebody with the title "VP of technical operations" to be so technically naive ...oh... wait.  Facebook.
Never mind.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477361</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477115</id>
	<title>Sounds like a bunch of excuses to me</title>
	<author>Anonymous</author>
	<datestamp>1245948420000</datestamp>
	<modclass>Insightful</modclass>
	<modscore>4</modscore>
	<htmltext><p>Assuming that a solution was properly engineered, this should not have been a surprise.</p><p>Cheap. power efficient, performance. Pick two.</p></htmltext>
<tokenext>Assuming that a solution was properly engineered , this should not have been a surprise.Cheap .
power efficient , performance .
Pick two .</tokentext>
<sentencetext>Assuming that a solution was properly engineered, this should not have been a surprise.Cheap.
power efficient, performance.
Pick two.</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477373</id>
	<title>Re:Facebook's application is poorly coded</title>
	<author>RightSaidFred99</author>
	<datestamp>1245950340000</datestamp>
	<modclass>Troll</modclass>
	<modscore>-1</modscore>
	<htmltext><p>No, the facebook guy is just stupid, that's all.  There are plenty of other performance bottlenecks in a system.  Hard drive performance is currently the biggest one.  This asshole is like a gamer who updates from a 2.2GHz processor to a next-gen 3.0Ghz faster per clock CPU then complains when Crysis doesn't run any faster.</p><p>Yeah, stupid, Crysis isn't faster because your $150 graphics card was the bottleneck.</p></htmltext>
<tokenext>No , the facebook guy is just stupid , that 's all .
There are plenty of other performance bottlenecks in a system .
Hard drive performance is currently the biggest one .
This asshole is like a gamer who updates from a 2.2GHz processor to a next-gen 3.0Ghz faster per clock CPU then complains when Crysis does n't run any faster.Yeah , stupid , Crysis is n't faster because your $ 150 graphics card was the bottleneck .</tokentext>
<sentencetext>No, the facebook guy is just stupid, that's all.
There are plenty of other performance bottlenecks in a system.
Hard drive performance is currently the biggest one.
This asshole is like a gamer who updates from a 2.2GHz processor to a next-gen 3.0Ghz faster per clock CPU then complains when Crysis doesn't run any faster.Yeah, stupid, Crysis isn't faster because your $150 graphics card was the bottleneck.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477085</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477967</id>
	<title>Re:Sounds like a bunch of excuses to me</title>
	<author>rossifer</author>
	<datestamp>1245955920000</datestamp>
	<modclass>Interestin</modclass>
	<modscore>2</modscore>
	<htmltext><blockquote><div><p>Assuming that a solution was properly engineered, this should not have been a surprise.</p><p>Cheap. power efficient, performance. Pick two.</p></div></blockquote><p>Actually, Google got all three of those in their system-level design (when cheap is measured per CPU).  What they didn't get was per CPU reliability.  That's pretty miserable by the standards of commercial servers.  Luckily, all Google software is architected, designed, and written to work around frequent hardware failures, so that's ultimately covered.</p></div>
	</htmltext>
<tokenext>Assuming that a solution was properly engineered , this should not have been a surprise.Cheap .
power efficient , performance .
Pick two.Actually , Google got all three of those in their system-level design ( when cheap is measured per CPU ) .
What they did n't get was per CPU reliability .
That 's pretty miserable by the standards of commercial servers .
Luckily , all Google software is architected , designed , and written to work around frequent hardware failures , so that 's ultimately covered .</tokentext>
<sentencetext>Assuming that a solution was properly engineered, this should not have been a surprise.Cheap.
power efficient, performance.
Pick two.Actually, Google got all three of those in their system-level design (when cheap is measured per CPU).
What they didn't get was per CPU reliability.
That's pretty miserable by the standards of commercial servers.
Luckily, all Google software is architected, designed, and written to work around frequent hardware failures, so that's ultimately covered.
	</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477115</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28480433</id>
	<title>Better Software helps</title>
	<author>happy\_place</author>
	<datestamp>1246023420000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext>No matter how much faster hardware gets, the software has to take advantage of the improvements.  If you're using a bloated interpreted web-based language running on an OS that's not fine-tuned to your given piece of hardware and don't see huge improvements, perhaps one should evaluate the layers of innefficient code that you've rested your apps on... The Gaming Industry has had a lot of success in eeking out every cycle of performance possible, but they spend time tuning their products to hardware solutions.</htmltext>
<tokenext>No matter how much faster hardware gets , the software has to take advantage of the improvements .
If you 're using a bloated interpreted web-based language running on an OS that 's not fine-tuned to your given piece of hardware and do n't see huge improvements , perhaps one should evaluate the layers of innefficient code that you 've rested your apps on... The Gaming Industry has had a lot of success in eeking out every cycle of performance possible , but they spend time tuning their products to hardware solutions .</tokentext>
<sentencetext>No matter how much faster hardware gets, the software has to take advantage of the improvements.
If you're using a bloated interpreted web-based language running on an OS that's not fine-tuned to your given piece of hardware and don't see huge improvements, perhaps one should evaluate the layers of innefficient code that you've rested your apps on... The Gaming Industry has had a lot of success in eeking out every cycle of performance possible, but they spend time tuning their products to hardware solutions.</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28482249</id>
	<title>Yeah ok!</title>
	<author>kuei12</author>
	<datestamp>1246030560000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext>I'm thinkin this fella Jonathan Heiliger is probably disappointed because he is probably running Dell Power Edge servers he bought off someone on Craigslist. That would explain the lack of chip performance. Does he have any clue that google is probably running intel/amd chips? They may have designed their own servers, but not the chips.</htmltext>
<tokenext>I 'm thinkin this fella Jonathan Heiliger is probably disappointed because he is probably running Dell Power Edge servers he bought off someone on Craigslist .
That would explain the lack of chip performance .
Does he have any clue that google is probably running intel/amd chips ?
They may have designed their own servers , but not the chips .</tokentext>
<sentencetext>I'm thinkin this fella Jonathan Heiliger is probably disappointed because he is probably running Dell Power Edge servers he bought off someone on Craigslist.
That would explain the lack of chip performance.
Does he have any clue that google is probably running intel/amd chips?
They may have designed their own servers, but not the chips.</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28482555</id>
	<title>Re:Hm...</title>
	<author>Anonymous</author>
	<datestamp>1246031340000</datestamp>
	<modclass>None</modclass>
	<modscore>0</modscore>
	<htmltext><p>the obvious conclusion is that your hbas + vmware are not<br>an efficient combination.</p><p>i get linerate AoE on 2gbe nics at 5\% cpu on some pretty<br>pitiful linux machines.</p></htmltext>
<tokenext>the obvious conclusion is that your hbas + vmware are notan efficient combination.i get linerate AoE on 2gbe nics at 5 \ % cpu on some prettypitiful linux machines .</tokentext>
<sentencetext>the obvious conclusion is that your hbas + vmware are notan efficient combination.i get linerate AoE on 2gbe nics at 5\% cpu on some prettypitiful linux machines.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477411</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477487</id>
	<title>And yet...</title>
	<author>Junta</author>
	<datestamp>1245951420000</datestamp>
	<modclass>Interestin</modclass>
	<modscore>5</modscore>
	<htmltext><p>Every major server vendor has jumped on the bandwagon of 'look how efficient we are, and 'cheap'.  Three years ago, by and large the tier ones wouldn't bother designing systems without forcing even the cheap design to have parts included to facilitate purchase of redundant add-ons (i.e. power distribution cards designed for dual power supplies regardless of one being bought or not).  They would always put a high end storage controller on the planar.  They would always make their 'entry' platform be burdened with expensive components to make it easier to option it up.</p><p>Now, we have tons of 'internet scale', or 'cloud', or whatever buzzword you feel like.  They tend to stress energy efficiency, low cost components, with sales and management strategies targeted at thousands of servers  (i.e. IBM iDataplex, HP SL6000).  Basically, precisely what he prescribes, though probably not as 'cheap' as he wants.  The incentive he gives is that the vendors should have zero margin, which is not particularly compelling for companies to work toward.  Google's situation works because they brought it in-house and thus have fewer middle-men.  Honestly, from all the rumours I hear, it's the logical thing to do when your server consumption is larger than some respectable computer companies' entire production.  If he thinks the volume of servers is high enough to pull a google, by all means do it.  Otherwise, be prepared for people not jump at the chance to give their designs to him at zero margin.</p><p>Of course, if he is calling them out on performance per-watt by avoiding non-x86 solutions, including ARM, that might be a fair criticism.  However, I think company forays into 'exotic' architectures have not panned out in the market recently.  Sun's niagra, despite all the worthy praise, couldn't attract a mass-market required to subsidize it for those who benefited most from it.  Last year, IBM seemed to be saying Cell architecture would light the world on fire, but have been a lot quieter about it now.   The message their buisness leaders have probably taken in is that while these things have their target market, that market isn't worth the expense of developing products that are refused by the larger market and focus instead on leveraging commonly accepted building blocks to do as best they can for that niche, even if it means skipping the 'perfect' solution.  Sure, IBM still sells plenty of POWER, but I haven't heard that be *particularly* praised on the performance/watt category like I hear a lot for Niagra, Cell, and ARM.  And if not for POWER's legacy, it probably would be still born in the market today.  The PA-RISC-&gt;Itanium decision for HP probably sank their HP-UX product line faster than banking on legacy of PA-RISC installs, and it seems IBM won't make that mistake, but at the same time I don't hear much about *new* POWER customers.</p></htmltext>
<tokenext>Every major server vendor has jumped on the bandwagon of 'look how efficient we are , and 'cheap' .
Three years ago , by and large the tier ones would n't bother designing systems without forcing even the cheap design to have parts included to facilitate purchase of redundant add-ons ( i.e .
power distribution cards designed for dual power supplies regardless of one being bought or not ) .
They would always put a high end storage controller on the planar .
They would always make their 'entry ' platform be burdened with expensive components to make it easier to option it up.Now , we have tons of 'internet scale ' , or 'cloud ' , or whatever buzzword you feel like .
They tend to stress energy efficiency , low cost components , with sales and management strategies targeted at thousands of servers ( i.e .
IBM iDataplex , HP SL6000 ) .
Basically , precisely what he prescribes , though probably not as 'cheap ' as he wants .
The incentive he gives is that the vendors should have zero margin , which is not particularly compelling for companies to work toward .
Google 's situation works because they brought it in-house and thus have fewer middle-men .
Honestly , from all the rumours I hear , it 's the logical thing to do when your server consumption is larger than some respectable computer companies ' entire production .
If he thinks the volume of servers is high enough to pull a google , by all means do it .
Otherwise , be prepared for people not jump at the chance to give their designs to him at zero margin.Of course , if he is calling them out on performance per-watt by avoiding non-x86 solutions , including ARM , that might be a fair criticism .
However , I think company forays into 'exotic ' architectures have not panned out in the market recently .
Sun 's niagra , despite all the worthy praise , could n't attract a mass-market required to subsidize it for those who benefited most from it .
Last year , IBM seemed to be saying Cell architecture would light the world on fire , but have been a lot quieter about it now .
The message their buisness leaders have probably taken in is that while these things have their target market , that market is n't worth the expense of developing products that are refused by the larger market and focus instead on leveraging commonly accepted building blocks to do as best they can for that niche , even if it means skipping the 'perfect ' solution .
Sure , IBM still sells plenty of POWER , but I have n't heard that be * particularly * praised on the performance/watt category like I hear a lot for Niagra , Cell , and ARM .
And if not for POWER 's legacy , it probably would be still born in the market today .
The PA-RISC- &gt; Itanium decision for HP probably sank their HP-UX product line faster than banking on legacy of PA-RISC installs , and it seems IBM wo n't make that mistake , but at the same time I do n't hear much about * new * POWER customers .</tokentext>
<sentencetext>Every major server vendor has jumped on the bandwagon of 'look how efficient we are, and 'cheap'.
Three years ago, by and large the tier ones wouldn't bother designing systems without forcing even the cheap design to have parts included to facilitate purchase of redundant add-ons (i.e.
power distribution cards designed for dual power supplies regardless of one being bought or not).
They would always put a high end storage controller on the planar.
They would always make their 'entry' platform be burdened with expensive components to make it easier to option it up.Now, we have tons of 'internet scale', or 'cloud', or whatever buzzword you feel like.
They tend to stress energy efficiency, low cost components, with sales and management strategies targeted at thousands of servers  (i.e.
IBM iDataplex, HP SL6000).
Basically, precisely what he prescribes, though probably not as 'cheap' as he wants.
The incentive he gives is that the vendors should have zero margin, which is not particularly compelling for companies to work toward.
Google's situation works because they brought it in-house and thus have fewer middle-men.
Honestly, from all the rumours I hear, it's the logical thing to do when your server consumption is larger than some respectable computer companies' entire production.
If he thinks the volume of servers is high enough to pull a google, by all means do it.
Otherwise, be prepared for people not jump at the chance to give their designs to him at zero margin.Of course, if he is calling them out on performance per-watt by avoiding non-x86 solutions, including ARM, that might be a fair criticism.
However, I think company forays into 'exotic' architectures have not panned out in the market recently.
Sun's niagra, despite all the worthy praise, couldn't attract a mass-market required to subsidize it for those who benefited most from it.
Last year, IBM seemed to be saying Cell architecture would light the world on fire, but have been a lot quieter about it now.
The message their buisness leaders have probably taken in is that while these things have their target market, that market isn't worth the expense of developing products that are refused by the larger market and focus instead on leveraging commonly accepted building blocks to do as best they can for that niche, even if it means skipping the 'perfect' solution.
Sure, IBM still sells plenty of POWER, but I haven't heard that be *particularly* praised on the performance/watt category like I hear a lot for Niagra, Cell, and ARM.
And if not for POWER's legacy, it probably would be still born in the market today.
The PA-RISC-&gt;Itanium decision for HP probably sank their HP-UX product line faster than banking on legacy of PA-RISC installs, and it seems IBM won't make that mistake, but at the same time I don't hear much about *new* POWER customers.</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479893</id>
	<title>Re:You're Computin' for a Shootin' Mister</title>
	<author>KibibyteBrain</author>
	<datestamp>1246017720000</datestamp>
	<modclass>Insightful</modclass>
	<modscore>2</modscore>
	<htmltext>Yes, this reads like "Guy with huge ego upset that engineers can't use magic to conjure up ideal device at his command." to me.</htmltext>
<tokenext>Yes , this reads like " Guy with huge ego upset that engineers ca n't use magic to conjure up ideal device at his command .
" to me .</tokentext>
<sentencetext>Yes, this reads like "Guy with huge ego upset that engineers can't use magic to conjure up ideal device at his command.
" to me.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476905</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477977</id>
	<title>Re:WTF?</title>
	<author>Runaway1956</author>
	<datestamp>1245955980000</datestamp>
	<modclass>Interestin</modclass>
	<modscore>3</modscore>
	<htmltext><p>Uhhh, correct me if I'm wrong.  I've been looking at after market bolt on parts for my car.  The headers claim increase fuel mileage, the spark plugs, the air filter, the tires, as does a turbocharger. The glass pack mufflers, and the resonator.  Oh yeah, the aerodynamic rims, the hood, and spoiler.  Don't forget the carbon fiber body panels.  Taken all together, those increased MPG's add up to about 150 MPG.  You're saying I may not see that much improvement on my 1968 Chevy Malibu?  It's just hype?  Man - you just saved me about $5,000!!!</p></htmltext>
<tokenext>Uhhh , correct me if I 'm wrong .
I 've been looking at after market bolt on parts for my car .
The headers claim increase fuel mileage , the spark plugs , the air filter , the tires , as does a turbocharger .
The glass pack mufflers , and the resonator .
Oh yeah , the aerodynamic rims , the hood , and spoiler .
Do n't forget the carbon fiber body panels .
Taken all together , those increased MPG 's add up to about 150 MPG .
You 're saying I may not see that much improvement on my 1968 Chevy Malibu ?
It 's just hype ?
Man - you just saved me about $ 5,000 ! !
!</tokentext>
<sentencetext>Uhhh, correct me if I'm wrong.
I've been looking at after market bolt on parts for my car.
The headers claim increase fuel mileage, the spark plugs, the air filter, the tires, as does a turbocharger.
The glass pack mufflers, and the resonator.
Oh yeah, the aerodynamic rims, the hood, and spoiler.
Don't forget the carbon fiber body panels.
Taken all together, those increased MPG's add up to about 150 MPG.
You're saying I may not see that much improvement on my 1968 Chevy Malibu?
It's just hype?
Man - you just saved me about $5,000!!
!</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477361</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479135</id>
	<title>Re:Hm...</title>
	<author>drsmithy</author>
	<datestamp>1246009320000</datestamp>
	<modclass>Informativ</modclass>
	<modscore>2</modscore>
	<htmltext><p> <i>As someone who designs and deploys large storage environments for a living, </i>
</p><p>Then you should know that throughput is not the only (or - typically - the most important) measure of IO performance.
</p><p>Typical computing tasks tend to be I/O bound - specifically by random I/O performance.  To a large degree, this is due to the massive disparity in performance improvements between CPUa and storage.</p></htmltext>
<tokenext>As someone who designs and deploys large storage environments for a living , Then you should know that throughput is not the only ( or - typically - the most important ) measure of IO performance .
Typical computing tasks tend to be I/O bound - specifically by random I/O performance .
To a large degree , this is due to the massive disparity in performance improvements between CPUa and storage .</tokentext>
<sentencetext> As someone who designs and deploys large storage environments for a living, 
Then you should know that throughput is not the only (or - typically - the most important) measure of IO performance.
Typical computing tasks tend to be I/O bound - specifically by random I/O performance.
To a large degree, this is due to the massive disparity in performance improvements between CPUa and storage.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477411</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28491535</id>
	<title>facebook vp must know they are designed for vm's</title>
	<author>Anonymous</author>
	<datestamp>1246038000000</datestamp>
	<modclass>None</modclass>
	<modscore>0</modscore>
	<htmltext><p>LOL the new ones were designed for VM's really... haha thats why they do all the hardware virtulization and Virtulized i/o and they pimp it showing it. They could take advantage of super computers and drain there drops out of them if they built on the vm cloud. but it's just easier to do googles approach and much cheaper. But for paying customers they will want to buy the warranty and support because they are selling a service. Facebook guy needs a life.</p></htmltext>
<tokenext>LOL the new ones were designed for VM 's really... haha thats why they do all the hardware virtulization and Virtulized i/o and they pimp it showing it .
They could take advantage of super computers and drain there drops out of them if they built on the vm cloud .
but it 's just easier to do googles approach and much cheaper .
But for paying customers they will want to buy the warranty and support because they are selling a service .
Facebook guy needs a life .</tokentext>
<sentencetext>LOL the new ones were designed for VM's really... haha thats why they do all the hardware virtulization and Virtulized i/o and they pimp it showing it.
They could take advantage of super computers and drain there drops out of them if they built on the vm cloud.
but it's just easier to do googles approach and much cheaper.
But for paying customers they will want to buy the warranty and support because they are selling a service.
Facebook guy needs a life.</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477169</id>
	<title>Re:Well I suppose...</title>
	<author>NervousNerd</author>
	<datestamp>1245948780000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext>I heard that Cyrix CPU's had bitching performance.</htmltext>
<tokenext>I heard that Cyrix CPU 's had bitching performance .</tokentext>
<sentencetext>I heard that Cyrix CPU's had bitching performance.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476967</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477077</id>
	<title>Re:Well I suppose...</title>
	<author>WilliamBaughman</author>
	<datestamp>1245948120000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext>Well, Sun's Niagara 2 processor seems pretty good at number of network I/O operations per Watt.  And if he doesn't like that, he can try IBM's Power6, although I hear it's better for floating-point than integer (relative to Intel and AMD's offerings.)  If he doesn't like that, he can sell his company to Google and ask them to build servers for him.</htmltext>
<tokenext>Well , Sun 's Niagara 2 processor seems pretty good at number of network I/O operations per Watt .
And if he does n't like that , he can try IBM 's Power6 , although I hear it 's better for floating-point than integer ( relative to Intel and AMD 's offerings .
) If he does n't like that , he can sell his company to Google and ask them to build servers for him .</tokentext>
<sentencetext>Well, Sun's Niagara 2 processor seems pretty good at number of network I/O operations per Watt.
And if he doesn't like that, he can try IBM's Power6, although I hear it's better for floating-point than integer (relative to Intel and AMD's offerings.
)  If he doesn't like that, he can sell his company to Google and ask them to build servers for him.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476967</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477335</id>
	<title>Re:WTF?</title>
	<author>whoever57</author>
	<datestamp>1245950040000</datestamp>
	<modclass>Informativ</modclass>
	<modscore>1</modscore>
	<htmltext><p>I have some sympathy for this guy. Some years ago, I built a fileserver using the best SATA RAID (hardware RAID) cards I could find (~$300) from major manufacturers and enterprise disks (specified for use in RAID systems)</p><p>Performance absolutely sucked. The cards were fast enough it I tried to read/write single large files, but when reading/writing large numbers of small files, they were very slow. The first manufacturer's card was appallingly slow. I replaced it with another manufacturer's card and performance was merely slow.</p><p>I followed all the manufacturer's recommendations, I communicated with one manufacturer on a Linux RAID mailing list, but was never able to get anything remotely like acceptable performance. For compariso, later I built a fileserver around an old (sub 1GHz) PC, using software RAID and was able to get at least the same performance.</p><p>I was only building one machine, so I did not have the luxury of benchmarking it.</p></htmltext>
<tokenext>I have some sympathy for this guy .
Some years ago , I built a fileserver using the best SATA RAID ( hardware RAID ) cards I could find ( ~ $ 300 ) from major manufacturers and enterprise disks ( specified for use in RAID systems ) Performance absolutely sucked .
The cards were fast enough it I tried to read/write single large files , but when reading/writing large numbers of small files , they were very slow .
The first manufacturer 's card was appallingly slow .
I replaced it with another manufacturer 's card and performance was merely slow.I followed all the manufacturer 's recommendations , I communicated with one manufacturer on a Linux RAID mailing list , but was never able to get anything remotely like acceptable performance .
For compariso , later I built a fileserver around an old ( sub 1GHz ) PC , using software RAID and was able to get at least the same performance.I was only building one machine , so I did not have the luxury of benchmarking it .</tokentext>
<sentencetext>I have some sympathy for this guy.
Some years ago, I built a fileserver using the best SATA RAID (hardware RAID) cards I could find (~$300) from major manufacturers and enterprise disks (specified for use in RAID systems)Performance absolutely sucked.
The cards were fast enough it I tried to read/write single large files, but when reading/writing large numbers of small files, they were very slow.
The first manufacturer's card was appallingly slow.
I replaced it with another manufacturer's card and performance was merely slow.I followed all the manufacturer's recommendations, I communicated with one manufacturer on a Linux RAID mailing list, but was never able to get anything remotely like acceptable performance.
For compariso, later I built a fileserver around an old (sub 1GHz) PC, using software RAID and was able to get at least the same performance.I was only building one machine, so I did not have the luxury of benchmarking it.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476921</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478723</id>
	<title>PHP "extension"</title>
	<author>RGRistroph</author>
	<datestamp>1246048860000</datestamp>
	<modclass>Insightful</modclass>
	<modscore>4</modscore>
	<htmltext><p>I once did a large project in which I took a large, slow site in PHP (it was pretty complecated, it was a CRM with a lot of custom business logic) and rewrote all the core functionality from PHP to C / C++, and made it a "module" of PHP.  The rewriting was mostly simple translation -- litterally removing all dollar signs, adding some types, and attempting to compile, and just fixing the compile errors until it would build.  Then going back through it with a fine-tooth comb to track down all the memory leaks.</p><p>The speed increase from doing that is pretty surprising.  Simple loops that do a bit of math or something speed up by 100 times, and a loop that creates and destroys an object within the loop will be 100,000 times faster.  This is without actually trying to write fast C/C++ code, and not create and delete the same thing over and over in a loop -- just pure dumb translation of the code.</p><p>At that point, the web site guys can keep tweaking and changing the web page in PHP just like before; but they load that module in the php.ini and then they have a basic library of stuff, like login\_user() or get\_user\_balance() and etc, that are really fast and do all the heavy lifting.</p><p>I would be surprised if Facebook has not already done this.  How to do it is well documented in several books, and there are lots of PHP modules written in C/C++ to look at for examples.</p><p>I suspect that Facebook's VP is right that AMD and Intel exaggerate their claims, but is also generally true that most computer programs are more IO bound that you expect.  This is not a reason to avoid something like I describe above; once you have the more complete control of programming in C, IO issues may be easier to find and address.</p><p>He also mentions that the servers offered by Dell and others aren't very power efficient or practicle for him, and he mentions Google designing their own servers.  Nothing google did was really rocket science, from what we know, and Facebook probably doesn't have to go as far as they did to get a reasonable benefit.  It's not that hard to set up motherboards to run without a case, booting off the network with no harddrive attached.</p></htmltext>
<tokenext>I once did a large project in which I took a large , slow site in PHP ( it was pretty complecated , it was a CRM with a lot of custom business logic ) and rewrote all the core functionality from PHP to C / C + + , and made it a " module " of PHP .
The rewriting was mostly simple translation -- litterally removing all dollar signs , adding some types , and attempting to compile , and just fixing the compile errors until it would build .
Then going back through it with a fine-tooth comb to track down all the memory leaks.The speed increase from doing that is pretty surprising .
Simple loops that do a bit of math or something speed up by 100 times , and a loop that creates and destroys an object within the loop will be 100,000 times faster .
This is without actually trying to write fast C/C + + code , and not create and delete the same thing over and over in a loop -- just pure dumb translation of the code.At that point , the web site guys can keep tweaking and changing the web page in PHP just like before ; but they load that module in the php.ini and then they have a basic library of stuff , like login \ _user ( ) or get \ _user \ _balance ( ) and etc , that are really fast and do all the heavy lifting.I would be surprised if Facebook has not already done this .
How to do it is well documented in several books , and there are lots of PHP modules written in C/C + + to look at for examples.I suspect that Facebook 's VP is right that AMD and Intel exaggerate their claims , but is also generally true that most computer programs are more IO bound that you expect .
This is not a reason to avoid something like I describe above ; once you have the more complete control of programming in C , IO issues may be easier to find and address.He also mentions that the servers offered by Dell and others are n't very power efficient or practicle for him , and he mentions Google designing their own servers .
Nothing google did was really rocket science , from what we know , and Facebook probably does n't have to go as far as they did to get a reasonable benefit .
It 's not that hard to set up motherboards to run without a case , booting off the network with no harddrive attached .</tokentext>
<sentencetext>I once did a large project in which I took a large, slow site in PHP (it was pretty complecated, it was a CRM with a lot of custom business logic) and rewrote all the core functionality from PHP to C / C++, and made it a "module" of PHP.
The rewriting was mostly simple translation -- litterally removing all dollar signs, adding some types, and attempting to compile, and just fixing the compile errors until it would build.
Then going back through it with a fine-tooth comb to track down all the memory leaks.The speed increase from doing that is pretty surprising.
Simple loops that do a bit of math or something speed up by 100 times, and a loop that creates and destroys an object within the loop will be 100,000 times faster.
This is without actually trying to write fast C/C++ code, and not create and delete the same thing over and over in a loop -- just pure dumb translation of the code.At that point, the web site guys can keep tweaking and changing the web page in PHP just like before; but they load that module in the php.ini and then they have a basic library of stuff, like login\_user() or get\_user\_balance() and etc, that are really fast and do all the heavy lifting.I would be surprised if Facebook has not already done this.
How to do it is well documented in several books, and there are lots of PHP modules written in C/C++ to look at for examples.I suspect that Facebook's VP is right that AMD and Intel exaggerate their claims, but is also generally true that most computer programs are more IO bound that you expect.
This is not a reason to avoid something like I describe above; once you have the more complete control of programming in C, IO issues may be easier to find and address.He also mentions that the servers offered by Dell and others aren't very power efficient or practicle for him, and he mentions Google designing their own servers.
Nothing google did was really rocket science, from what we know, and Facebook probably doesn't have to go as far as they did to get a reasonable benefit.
It's not that hard to set up motherboards to run without a case, booting off the network with no harddrive attached.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477011</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477063</id>
	<title>Atom</title>
	<author>Anonymous</author>
	<datestamp>1245948120000</datestamp>
	<modclass>None</modclass>
	<modscore>0</modscore>
	<htmltext>"the servers have to be cheap, and they have to be super power-efficient."

So aren't Atom-based nettops using like 5 watts and dual core versions selling for $150, you supply the drive?

<a href="http://www.newegg.com/Product/Product.aspx?Item=N82E16856167037" title="newegg.com" rel="nofollow">http://www.newegg.com/Product/Product.aspx?Item=N82E16856167037</a> [newegg.com]</htmltext>
<tokenext>" the servers have to be cheap , and they have to be super power-efficient .
" So are n't Atom-based nettops using like 5 watts and dual core versions selling for $ 150 , you supply the drive ?
http : //www.newegg.com/Product/Product.aspx ? Item = N82E16856167037 [ newegg.com ]</tokentext>
<sentencetext>"the servers have to be cheap, and they have to be super power-efficient.
"

So aren't Atom-based nettops using like 5 watts and dual core versions selling for $150, you supply the drive?
http://www.newegg.com/Product/Product.aspx?Item=N82E16856167037 [newegg.com]</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28487137</id>
	<title>Re:WTF?</title>
	<author>Anonymous</author>
	<datestamp>1246049220000</datestamp>
	<modclass>None</modclass>
	<modscore>0</modscore>
	<htmltext><p>Run XP on a new Core2Duo system. Using his logic, it should be twice as fast if not faster than XP on a single core P4 from 2006. It isn't. Why? XP was not built to use two cores efficiently, as is the case with most software on the market today.</p><p>Software plays a huge role in hardware speeds. Poorly written software even more so.</p></htmltext>
<tokenext>Run XP on a new Core2Duo system .
Using his logic , it should be twice as fast if not faster than XP on a single core P4 from 2006 .
It is n't .
Why ? XP was not built to use two cores efficiently , as is the case with most software on the market today.Software plays a huge role in hardware speeds .
Poorly written software even more so .</tokentext>
<sentencetext>Run XP on a new Core2Duo system.
Using his logic, it should be twice as fast if not faster than XP on a single core P4 from 2006.
It isn't.
Why? XP was not built to use two cores efficiently, as is the case with most software on the market today.Software plays a huge role in hardware speeds.
Poorly written software even more so.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477763</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479535</id>
	<title>Re:Would you expect otherwise</title>
	<author>vic-traill</author>
	<datestamp>1246013400000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext><p><div class="quote"><p>Google's core business is intelligence. Facebooks core business is stupidity.</p></div><p>.

</p><p>That's a god-damned beaut, brother.  I actually spewed some yogurt into my nose, I started laughing so hard in mid-swallow.</p><p>If I had my way around here - which of course I don't - you'd get some bonus/recognition cool award thing.  Not mod points, but like an instant order of magnitude reduction in your<nobr> <wbr></nobr>/. UID.</p><p>Not that you really need it, now that I've looked at your UID, but hey, it's the thought that counts, right?</p></div>
	</htmltext>
<tokenext>Google 's core business is intelligence .
Facebooks core business is stupidity. . That 's a god-damned beaut , brother .
I actually spewed some yogurt into my nose , I started laughing so hard in mid-swallow.If I had my way around here - which of course I do n't - you 'd get some bonus/recognition cool award thing .
Not mod points , but like an instant order of magnitude reduction in your / .
UID.Not that you really need it , now that I 've looked at your UID , but hey , it 's the thought that counts , right ?</tokentext>
<sentencetext>Google's core business is intelligence.
Facebooks core business is stupidity..

That's a god-damned beaut, brother.
I actually spewed some yogurt into my nose, I started laughing so hard in mid-swallow.If I had my way around here - which of course I don't - you'd get some bonus/recognition cool award thing.
Not mod points, but like an instant order of magnitude reduction in your /.
UID.Not that you really need it, now that I've looked at your UID, but hey, it's the thought that counts, right?
	</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477675</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28480163</id>
	<title>Super power efficiency?</title>
	<author>Tragedy4u</author>
	<datestamp>1246021020000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext>What grown man, let alone a VP of technology would use the word SUPER to describe something technical?</htmltext>
<tokenext>What grown man , let alone a VP of technology would use the word SUPER to describe something technical ?</tokentext>
<sentencetext>What grown man, let alone a VP of technology would use the word SUPER to describe something technical?</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28489715</id>
	<title>Re:Hm...</title>
	<author>Anonymous</author>
	<datestamp>1246019340000</datestamp>
	<modclass>None</modclass>
	<modscore>0</modscore>
	<htmltext><p>You may have a point, but I'm going to pick on your example.</p><p><div class="quote"><p>Even as we deploy VMware based servers, the VMware servers themselves tend to be more memory/cpu strapped than IO.</p></div><p>The whole POINT of consolidating servers into a virtual environment is to increase memory/cpu utilization of the host.  I would think they are not "average" servers at this point.</p><p>+1 Used IT professional acronyms that 95\% of<nobr> <wbr></nobr>/. "Nerds" couldn't point out in a lineup.</p></div>
	</htmltext>
<tokenext>You may have a point , but I 'm going to pick on your example.Even as we deploy VMware based servers , the VMware servers themselves tend to be more memory/cpu strapped than IO.The whole POINT of consolidating servers into a virtual environment is to increase memory/cpu utilization of the host .
I would think they are not " average " servers at this point. + 1 Used IT professional acronyms that 95 \ % of / .
" Nerds " could n't point out in a lineup .</tokentext>
<sentencetext>You may have a point, but I'm going to pick on your example.Even as we deploy VMware based servers, the VMware servers themselves tend to be more memory/cpu strapped than IO.The whole POINT of consolidating servers into a virtual environment is to increase memory/cpu utilization of the host.
I would think they are not "average" servers at this point.+1 Used IT professional acronyms that 95\% of /.
"Nerds" couldn't point out in a lineup.
	</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477411</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478473</id>
	<title>Re:WTF?</title>
	<author>Hecatonchires</author>
	<datestamp>1246046760000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext>+1 Use of a car example</htmltext>
<tokenext>+ 1 Use of a car example</tokentext>
<sentencetext>+1 Use of a car example</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477977</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28480495</id>
	<title>Re:Something about his arguement doesn't work</title>
	<author>ceoyoyo</author>
	<datestamp>1246023780000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext><p>It doesn't add up the other way either:</p><p>1) Facebook needs cheap, power efficient systems.<br>2) This guy is complaining because the chips aren't fast enough.</p></htmltext>
<tokenext>It does n't add up the other way either : 1 ) Facebook needs cheap , power efficient systems.2 ) This guy is complaining because the chips are n't fast enough .</tokentext>
<sentencetext>It doesn't add up the other way either:1) Facebook needs cheap, power efficient systems.2) This guy is complaining because the chips aren't fast enough.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476989</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28494677</id>
	<title>Re:You're Computin' for a Shootin' Mister</title>
	<author>lsatenstein</author>
	<datestamp>1246117440000</datestamp>
	<modclass>None</modclass>
	<modscore>0</modscore>
	<htmltext>Actually, the VP is right. And the problem is the architecture. 64bit AMD/Intel architecture is very adequate for the desktop, but not for servers.

Their microcode does not pipeline enough, and the dma needs to transfer 128bytes at a time, not just the 64bits at a time.
Also, need the disk system bios to really be a subsystem, part of a disk controller, not part of the main bios.  Offload work where possible, or switch to Mainframe systems which give more bang for the buck.</htmltext>
<tokenext>Actually , the VP is right .
And the problem is the architecture .
64bit AMD/Intel architecture is very adequate for the desktop , but not for servers .
Their microcode does not pipeline enough , and the dma needs to transfer 128bytes at a time , not just the 64bits at a time .
Also , need the disk system bios to really be a subsystem , part of a disk controller , not part of the main bios .
Offload work where possible , or switch to Mainframe systems which give more bang for the buck .</tokentext>
<sentencetext>Actually, the VP is right.
And the problem is the architecture.
64bit AMD/Intel architecture is very adequate for the desktop, but not for servers.
Their microcode does not pipeline enough, and the dma needs to transfer 128bytes at a time, not just the 64bits at a time.
Also, need the disk system bios to really be a subsystem, part of a disk controller, not part of the main bios.
Offload work where possible, or switch to Mainframe systems which give more bang for the buck.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476905</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478981</id>
	<title>Re:You're Computin' for a Shootin' Mister</title>
	<author>metacell</author>
	<datestamp>1246007700000</datestamp>
	<modclass>Insightful</modclass>
	<modscore>2</modscore>
	<htmltext><p><div class="quote"><p>So let me get this straight, the Vice President of a web company is criticizing the hardware guys in two of the world's biggest chip makers?</p></div><p>He's not criticising their technical know-how, he's criticising them for not knowing what their web company customers want.</p><p>Since he himself is one of those customers, it's not too unlikely that he knows what he's talking about.</p></div>
	</htmltext>
<tokenext>So let me get this straight , the Vice President of a web company is criticizing the hardware guys in two of the world 's biggest chip makers ? He 's not criticising their technical know-how , he 's criticising them for not knowing what their web company customers want.Since he himself is one of those customers , it 's not too unlikely that he knows what he 's talking about .</tokentext>
<sentencetext>So let me get this straight, the Vice President of a web company is criticizing the hardware guys in two of the world's biggest chip makers?He's not criticising their technical know-how, he's criticising them for not knowing what their web company customers want.Since he himself is one of those customers, it's not too unlikely that he knows what he's talking about.
	</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476905</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477523</id>
	<title>Re:WTF?</title>
	<author>Anonymous</author>
	<datestamp>1245951720000</datestamp>
	<modclass>Interestin</modclass>
	<modscore>2</modscore>
	<htmltext><p>You can better identify your bottlenecks by benchmarking. Facebook's scalability is likely not as cpu-bound as predicted, thus the dude's angst on discovering that CPU upgrades weren't a silver bullet.</p><p>In your case, you haven't looked past the RAID configuration for the root-cause of your performance issues. Without benchmarking you don't really know if it was an issue with: the filesystem, the block size, stripe size, or a caching tunable.</p><p>Systems architecture isn't as easy as PC builders would have you believe.</p></htmltext>
<tokenext>You can better identify your bottlenecks by benchmarking .
Facebook 's scalability is likely not as cpu-bound as predicted , thus the dude 's angst on discovering that CPU upgrades were n't a silver bullet.In your case , you have n't looked past the RAID configuration for the root-cause of your performance issues .
Without benchmarking you do n't really know if it was an issue with : the filesystem , the block size , stripe size , or a caching tunable.Systems architecture is n't as easy as PC builders would have you believe .</tokentext>
<sentencetext>You can better identify your bottlenecks by benchmarking.
Facebook's scalability is likely not as cpu-bound as predicted, thus the dude's angst on discovering that CPU upgrades weren't a silver bullet.In your case, you haven't looked past the RAID configuration for the root-cause of your performance issues.
Without benchmarking you don't really know if it was an issue with: the filesystem, the block size, stripe size, or a caching tunable.Systems architecture isn't as easy as PC builders would have you believe.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477335</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479875</id>
	<title>Re:WTF?</title>
	<author>somenickname</author>
	<datestamp>1246017600000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext><p>Actually, if the application is purely CPU bound and well tuned for CPU X, then switching to the superior CPU Y may not give you an performance benefits at all until it's recompiled and possibly tuned for CPU Y.  That assumes the compilers have been updated to understand how to optimally schedule instructions for CPU Y, what new instructions may be available on CPU Y, etc.  Hardware vendors know these things whereas hardware consumers think that CPU Y is supposed to be faster and so their software should be running faster.</p><p>An example of this was when Sun was switching from UltraSPARC II to UltraSPARC III.  Even though the UltraSPARC III chips were clocked 50\% higher and had all sorts of goodies that made it seem like a better chip, when the first internal machines appeared performance was dismal and was often below UltraSPARC II in benchmarks.  Once the compilers became smart enough to emit optimized code for the UltraSPARC III and once the systems engineers doing benchmarking understood the quirks of the new CPU, performance started to get more inline with expectations.</p></htmltext>
<tokenext>Actually , if the application is purely CPU bound and well tuned for CPU X , then switching to the superior CPU Y may not give you an performance benefits at all until it 's recompiled and possibly tuned for CPU Y. That assumes the compilers have been updated to understand how to optimally schedule instructions for CPU Y , what new instructions may be available on CPU Y , etc .
Hardware vendors know these things whereas hardware consumers think that CPU Y is supposed to be faster and so their software should be running faster.An example of this was when Sun was switching from UltraSPARC II to UltraSPARC III .
Even though the UltraSPARC III chips were clocked 50 \ % higher and had all sorts of goodies that made it seem like a better chip , when the first internal machines appeared performance was dismal and was often below UltraSPARC II in benchmarks .
Once the compilers became smart enough to emit optimized code for the UltraSPARC III and once the systems engineers doing benchmarking understood the quirks of the new CPU , performance started to get more inline with expectations .</tokentext>
<sentencetext>Actually, if the application is purely CPU bound and well tuned for CPU X, then switching to the superior CPU Y may not give you an performance benefits at all until it's recompiled and possibly tuned for CPU Y.  That assumes the compilers have been updated to understand how to optimally schedule instructions for CPU Y, what new instructions may be available on CPU Y, etc.
Hardware vendors know these things whereas hardware consumers think that CPU Y is supposed to be faster and so their software should be running faster.An example of this was when Sun was switching from UltraSPARC II to UltraSPARC III.
Even though the UltraSPARC III chips were clocked 50\% higher and had all sorts of goodies that made it seem like a better chip, when the first internal machines appeared performance was dismal and was often below UltraSPARC II in benchmarks.
Once the compilers became smart enough to emit optimized code for the UltraSPARC III and once the systems engineers doing benchmarking understood the quirks of the new CPU, performance started to get more inline with expectations.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477763</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477763</id>
	<title>Re:WTF?</title>
	<author>edmudama</author>
	<datestamp>1245954060000</datestamp>
	<modclass>Insightful</modclass>
	<modscore>5</modscore>
	<htmltext><p><div class="quote"><p>I think we read different articles.  He's not saying he didn't plan well enough, he's saying that Intel and AMD promise that Gen Y processor is 35\% faster than Gen X processor, and he's not seeing anywhere near 35\% in real world performance. </p></div><p>If the application was purely CPU bound, and Y wasn't giving me 35\% more than X, I'd complain.</p><p>However, if it's a complex system like almost everything else, why would they expect their application to get 35\% faster when there's probably 6 or 8 critical subsystems that could all be bottlenecks as well?</p></div>
	</htmltext>
<tokenext>I think we read different articles .
He 's not saying he did n't plan well enough , he 's saying that Intel and AMD promise that Gen Y processor is 35 \ % faster than Gen X processor , and he 's not seeing anywhere near 35 \ % in real world performance .
If the application was purely CPU bound , and Y was n't giving me 35 \ % more than X , I 'd complain.However , if it 's a complex system like almost everything else , why would they expect their application to get 35 \ % faster when there 's probably 6 or 8 critical subsystems that could all be bottlenecks as well ?</tokentext>
<sentencetext>I think we read different articles.
He's not saying he didn't plan well enough, he's saying that Intel and AMD promise that Gen Y processor is 35\% faster than Gen X processor, and he's not seeing anywhere near 35\% in real world performance.
If the application was purely CPU bound, and Y wasn't giving me 35\% more than X, I'd complain.However, if it's a complex system like almost everything else, why would they expect their application to get 35\% faster when there's probably 6 or 8 critical subsystems that could all be bottlenecks as well?
	</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477361</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478707</id>
	<title>Re:Facebook's application is poorly coded</title>
	<author>tyrione</author>
	<datestamp>1246048800000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext>Your math skills are crap.</htmltext>
<tokenext>Your math skills are crap .</tokentext>
<sentencetext>Your math skills are crap.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477167</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477441</id>
	<title>Re:I wonder if he's not thinking about scaling eno</title>
	<author>dirtyhippie</author>
	<datestamp>1245951000000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext><p>Java? Yes, absolutely. Not sure about<nobr> <wbr></nobr>.net tho.</p></htmltext>
<tokenext>Java ?
Yes , absolutely .
Not sure about .net tho .</tokentext>
<sentencetext>Java?
Yes, absolutely.
Not sure about .net tho.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477107</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476921</id>
	<title>WTF?</title>
	<author>Anonymous</author>
	<datestamp>1245947100000</datestamp>
	<modclass>Insightful</modclass>
	<modscore>1</modscore>
	<htmltext><p>Maybe the dude should have benchmarked before committing. How does he scope his projects, with brochures?</p></htmltext>
<tokenext>Maybe the dude should have benchmarked before committing .
How does he scope his projects , with brochures ?</tokentext>
<sentencetext>Maybe the dude should have benchmarked before committing.
How does he scope his projects, with brochures?</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28480021</id>
	<title>seriously guys...</title>
	<author>nimbius</author>
	<datestamp>1246019280000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext>how many mips does it take to run a database full of one-liner updates about someones cat, or boyfriends girlfriends ex fiance?

<br> <br>unless bejewelled is actually running as a computational cluster app, you folks have far LESS to bitch about than a web-business or the worlds largest search engine in my opinion.  find a more efficient operating system, and insist more efficient code.  you dont bitch at vendors for performance, you find other ones if it gets bad enough.  im sure Cray would love to hear from you guys, or perhaps maybe SGI?</htmltext>
<tokenext>how many mips does it take to run a database full of one-liner updates about someones cat , or boyfriends girlfriends ex fiance ?
unless bejewelled is actually running as a computational cluster app , you folks have far LESS to bitch about than a web-business or the worlds largest search engine in my opinion .
find a more efficient operating system , and insist more efficient code .
you dont bitch at vendors for performance , you find other ones if it gets bad enough .
im sure Cray would love to hear from you guys , or perhaps maybe SGI ?</tokentext>
<sentencetext>how many mips does it take to run a database full of one-liner updates about someones cat, or boyfriends girlfriends ex fiance?
unless bejewelled is actually running as a computational cluster app, you folks have far LESS to bitch about than a web-business or the worlds largest search engine in my opinion.
find a more efficient operating system, and insist more efficient code.
you dont bitch at vendors for performance, you find other ones if it gets bad enough.
im sure Cray would love to hear from you guys, or perhaps maybe SGI?</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479401</id>
	<title>Re:WTF?</title>
	<author>TheRaven64</author>
	<datestamp>1246012200000</datestamp>
	<modclass>Interestin</modclass>
	<modscore>4</modscore>
	<htmltext><p>One of the fun toys Intel has to play with is a complete system simulator, which simulates every single component in a computer for early testing.  This lets them vary parameters that aren't feasible yet while they're working on their design goals.  A few years ago they did a test; what happens to the system performance if you make the CPU infinitely fast?  They adjusted the simulator so that every CPU operation took zero simulated time and ran their benchmark suite.  It ran twice as fast (in simulated time) as it was before.  </p><p>
A CPU-bound workload can quickly become a RAM-speed bound or a disk-speed bound workload if you make the CPU faster but don't upgrade anything else.</p></htmltext>
<tokenext>One of the fun toys Intel has to play with is a complete system simulator , which simulates every single component in a computer for early testing .
This lets them vary parameters that are n't feasible yet while they 're working on their design goals .
A few years ago they did a test ; what happens to the system performance if you make the CPU infinitely fast ?
They adjusted the simulator so that every CPU operation took zero simulated time and ran their benchmark suite .
It ran twice as fast ( in simulated time ) as it was before .
A CPU-bound workload can quickly become a RAM-speed bound or a disk-speed bound workload if you make the CPU faster but do n't upgrade anything else .</tokentext>
<sentencetext>One of the fun toys Intel has to play with is a complete system simulator, which simulates every single component in a computer for early testing.
This lets them vary parameters that aren't feasible yet while they're working on their design goals.
A few years ago they did a test; what happens to the system performance if you make the CPU infinitely fast?
They adjusted the simulator so that every CPU operation took zero simulated time and ran their benchmark suite.
It ran twice as fast (in simulated time) as it was before.
A CPU-bound workload can quickly become a RAM-speed bound or a disk-speed bound workload if you make the CPU faster but don't upgrade anything else.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477763</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477383</id>
	<title>The Technical VP of Facebook?????</title>
	<author>Anonymous</author>
	<datestamp>1245950460000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext>Let me file his opinion with my next door neighbor the plumber.</htmltext>
<tokenext>Let me file his opinion with my next door neighbor the plumber .</tokentext>
<sentencetext>Let me file his opinion with my next door neighbor the plumber.</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478637</id>
	<title>Wrong tool for the job?</title>
	<author>Ihlosi</author>
	<datestamp>1246048200000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext><p>I don't think that Amazon, Facebook, etc, need all the floating point performance that current desktop and server CPUs can deliver. If they're really looking for power efficiency, they shouldn't use CPUs with features that are never used, but draw tons of power nonetheless.</p></htmltext>
<tokenext>I do n't think that Amazon , Facebook , etc , need all the floating point performance that current desktop and server CPUs can deliver .
If they 're really looking for power efficiency , they should n't use CPUs with features that are never used , but draw tons of power nonetheless .</tokentext>
<sentencetext>I don't think that Amazon, Facebook, etc, need all the floating point performance that current desktop and server CPUs can deliver.
If they're really looking for power efficiency, they shouldn't use CPUs with features that are never used, but draw tons of power nonetheless.</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477069</id>
	<title>Re:WTF?</title>
	<author>Anonymous</author>
	<datestamp>1245948120000</datestamp>
	<modclass>Insightful</modclass>
	<modscore>5</modscore>
	<htmltext>I agree I think this was writing his own resignation with this crap.  The guy is basically telling everyone that he is incapable of finding an acceptable solution for his company and blaming intel and amd because he has committed a great deal of money on something that he didn't plan well enough to know exactly what the long term costs vs performance was.   In the very article he says to not be cheap,  but in many more words than necessary, probably to try to disguise what he is saying like most politicians, that they were not only too cheap, but made bad decisions on what to be cheap with.

Its as if he's already in a public office, hes telling everyone he screwed up, why he screwed up, and trying to make it look like hes teaching everyone lesson to make his mistake to be less of a disaster.</htmltext>
<tokenext>I agree I think this was writing his own resignation with this crap .
The guy is basically telling everyone that he is incapable of finding an acceptable solution for his company and blaming intel and amd because he has committed a great deal of money on something that he did n't plan well enough to know exactly what the long term costs vs performance was .
In the very article he says to not be cheap , but in many more words than necessary , probably to try to disguise what he is saying like most politicians , that they were not only too cheap , but made bad decisions on what to be cheap with .
Its as if he 's already in a public office , hes telling everyone he screwed up , why he screwed up , and trying to make it look like hes teaching everyone lesson to make his mistake to be less of a disaster .</tokentext>
<sentencetext>I agree I think this was writing his own resignation with this crap.
The guy is basically telling everyone that he is incapable of finding an acceptable solution for his company and blaming intel and amd because he has committed a great deal of money on something that he didn't plan well enough to know exactly what the long term costs vs performance was.
In the very article he says to not be cheap,  but in many more words than necessary, probably to try to disguise what he is saying like most politicians, that they were not only too cheap, but made bad decisions on what to be cheap with.
Its as if he's already in a public office, hes telling everyone he screwed up, why he screwed up, and trying to make it look like hes teaching everyone lesson to make his mistake to be less of a disaster.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476921</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28481355</id>
	<title>Re:WTF?</title>
	<author>buddyglass</author>
	<datestamp>1246027500000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext>It doesn't seem obvious to me from the article that he bought a gillion processors and is now disappointed with their performance.  He just says that one of the unexpected barriers they've run into in their quest to continually upgrade their infrastructure performance is the fact that Intel &amp; AMD's server CPUs don't live up to their marketing when it comes to Facebook's particular workload.  Given that he also touts "testing" as a key to success in the very same article, I'm assuming they tested the performance of next-gen processors before buying a ton of them.</htmltext>
<tokenext>It does n't seem obvious to me from the article that he bought a gillion processors and is now disappointed with their performance .
He just says that one of the unexpected barriers they 've run into in their quest to continually upgrade their infrastructure performance is the fact that Intel &amp; AMD 's server CPUs do n't live up to their marketing when it comes to Facebook 's particular workload .
Given that he also touts " testing " as a key to success in the very same article , I 'm assuming they tested the performance of next-gen processors before buying a ton of them .</tokentext>
<sentencetext>It doesn't seem obvious to me from the article that he bought a gillion processors and is now disappointed with their performance.
He just says that one of the unexpected barriers they've run into in their quest to continually upgrade their infrastructure performance is the fact that Intel &amp; AMD's server CPUs don't live up to their marketing when it comes to Facebook's particular workload.
Given that he also touts "testing" as a key to success in the very same article, I'm assuming they tested the performance of next-gen processors before buying a ton of them.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477069</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479411</id>
	<title>What's your experience?</title>
	<author>jernejk</author>
	<datestamp>1246012320000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext>I'm just about to buy a new server and I'd like to hear from slashdot what's your experience with the new CPUs? Should I buy the new 55 or the old 54 series?</htmltext>
<tokenext>I 'm just about to buy a new server and I 'd like to hear from slashdot what 's your experience with the new CPUs ?
Should I buy the new 55 or the old 54 series ?</tokentext>
<sentencetext>I'm just about to buy a new server and I'd like to hear from slashdot what's your experience with the new CPUs?
Should I buy the new 55 or the old 54 series?</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478273</id>
	<title>Re:You're Computin' for a Shootin' Mister</title>
	<author>hairyfeet</author>
	<datestamp>1245958740000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext><p>If that is what he is looking for I wonder why he didn't look into <a href="http://www.pcworld.com/businesscenter/article/164209/vias\_nano\_may\_challenge\_intel\_amd\_in\_servers.html" title="pcworld.com">Nano servers</a> [pcworld.com] for his needs? Low power? Yep. Easy to cool? Yep. I don't know about the price but I bet buying in bulk like Facebook would do he would get a good price.</p><p>

It sounds to me like somebody didn't do his homework. As an added bonus the Nano has built in hardware crypto support which would help even more for things like HTTPS. He sounds like he would have been better off with Nano blades to me.</p></htmltext>
<tokenext>If that is what he is looking for I wonder why he did n't look into Nano servers [ pcworld.com ] for his needs ?
Low power ?
Yep. Easy to cool ?
Yep. I do n't know about the price but I bet buying in bulk like Facebook would do he would get a good price .
It sounds to me like somebody did n't do his homework .
As an added bonus the Nano has built in hardware crypto support which would help even more for things like HTTPS .
He sounds like he would have been better off with Nano blades to me .</tokentext>
<sentencetext>If that is what he is looking for I wonder why he didn't look into Nano servers [pcworld.com] for his needs?
Low power?
Yep. Easy to cool?
Yep. I don't know about the price but I bet buying in bulk like Facebook would do he would get a good price.
It sounds to me like somebody didn't do his homework.
As an added bonus the Nano has built in hardware crypto support which would help even more for things like HTTPS.
He sounds like he would have been better off with Nano blades to me.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476905</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477919</id>
	<title>Re:Hm...</title>
	<author>Mozk</author>
	<datestamp>1245955380000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext><p>Secretly selling all the data it accumulates on its users?</p></htmltext>
<tokenext>Secretly selling all the data it accumulates on its users ?</tokentext>
<sentencetext>Secretly selling all the data it accumulates on its users?</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477211</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28480557</id>
	<title>Re:You're Computin' for a Shootin' Mister</title>
	<author>Andy Dodd</author>
	<datestamp>1246024200000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext><p>One thing I wonder:</p><p>Many of the new CPUs DO have huge performance gains when running multimedia applications (video encoding/decoding, and especially 3D gaming).</p><p>Maybe they just aren't as good for Facebook's systems, which probably don't do much floating point, if at all.  Similarly, almost nothing FB does (except for possibly resizing images when they are uploaded by users) can take advantage of SIMD instructions such as SSE4 (which delivered huge performance gains for some relatively common desktop multimedia apps but isn't going to help an HTTP server or database server.)</p></htmltext>
<tokenext>One thing I wonder : Many of the new CPUs DO have huge performance gains when running multimedia applications ( video encoding/decoding , and especially 3D gaming ) .Maybe they just are n't as good for Facebook 's systems , which probably do n't do much floating point , if at all .
Similarly , almost nothing FB does ( except for possibly resizing images when they are uploaded by users ) can take advantage of SIMD instructions such as SSE4 ( which delivered huge performance gains for some relatively common desktop multimedia apps but is n't going to help an HTTP server or database server .
)</tokentext>
<sentencetext>One thing I wonder:Many of the new CPUs DO have huge performance gains when running multimedia applications (video encoding/decoding, and especially 3D gaming).Maybe they just aren't as good for Facebook's systems, which probably don't do much floating point, if at all.
Similarly, almost nothing FB does (except for possibly resizing images when they are uploaded by users) can take advantage of SIMD instructions such as SSE4 (which delivered huge performance gains for some relatively common desktop multimedia apps but isn't going to help an HTTP server or database server.
)</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476905</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28482179</id>
	<title>Re:Well I suppose...</title>
	<author>Anonymous</author>
	<datestamp>1246030380000</datestamp>
	<modclass>None</modclass>
	<modscore>0</modscore>
	<htmltext><p><div class="quote"><p>Well, I suppose that if he does not like the offerings from Intel and AMD, they could always go with...</p></div><p>ARM - works well enough for the best selling video game console. It's all about cycles per watt.</p></div>
	</htmltext>
<tokenext>Well , I suppose that if he does not like the offerings from Intel and AMD , they could always go with...ARM - works well enough for the best selling video game console .
It 's all about cycles per watt .</tokentext>
<sentencetext>Well, I suppose that if he does not like the offerings from Intel and AMD, they could always go with...ARM - works well enough for the best selling video game console.
It's all about cycles per watt.
	</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476967</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28484189</id>
	<title>Re:Hm...</title>
	<author>Anonymous</author>
	<datestamp>1246036800000</datestamp>
	<modclass>None</modclass>
	<modscore>0</modscore>
	<htmltext><p>Maybe he meant that the average server spends most of its time waiting for the mechanical heads of the hard disks to deliver requested data instead of worrying that it can't process more than 100 MB/s of data. ie limited by IO-ops/s instead of bandwidth<br>Most disk access is random especially for servers running multiple virtual machines.</p></htmltext>
<tokenext>Maybe he meant that the average server spends most of its time waiting for the mechanical heads of the hard disks to deliver requested data instead of worrying that it ca n't process more than 100 MB/s of data .
ie limited by IO-ops/s instead of bandwidthMost disk access is random especially for servers running multiple virtual machines .</tokentext>
<sentencetext>Maybe he meant that the average server spends most of its time waiting for the mechanical heads of the hard disks to deliver requested data instead of worrying that it can't process more than 100 MB/s of data.
ie limited by IO-ops/s instead of bandwidthMost disk access is random especially for servers running multiple virtual machines.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477411</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477881</id>
	<title>Improve the effieciency of the back-end!</title>
	<author>achacha</author>
	<datestamp>1245955020000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext><p>I bet if they rewrote the CPU intensive parts in C/C++ (from PHP) they could reduce the overall machine count and stop worrying about which CPU manufacturer lied to them that week<nobr> <wbr></nobr>:)</p></htmltext>
<tokenext>I bet if they rewrote the CPU intensive parts in C/C + + ( from PHP ) they could reduce the overall machine count and stop worrying about which CPU manufacturer lied to them that week : )</tokentext>
<sentencetext>I bet if they rewrote the CPU intensive parts in C/C++ (from PHP) they could reduce the overall machine count and stop worrying about which CPU manufacturer lied to them that week :)</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28480303</id>
	<title>Re:WTF?</title>
	<author>GaryOlson</author>
	<datestamp>1246022580000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext>Now spend that savings on a few classes in critical thinking and analysis....if the government will let you.</htmltext>
<tokenext>Now spend that savings on a few classes in critical thinking and analysis....if the government will let you .</tokentext>
<sentencetext>Now spend that savings on a few classes in critical thinking and analysis....if the government will let you.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477977</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478669</id>
	<title>Re:Hm...</title>
	<author>tyrione</author>
	<datestamp>1246048560000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext><p><div class="quote"><p><div class="quote"><p> To build servers for companies like Facebook, and Amazon, and other people who are operating fairly homogeneous applications, the servers have to be cheap, and they have to be super power-efficient.</p></div><p>

Hm, lets see... perhaps because Facebook and Amazon are niche markets? The average server isn't going to even need all the computing horsepower and the power efficiency is simply a drop in the bucket for most companies electrical bills. The average server is going to be much more I/O intensive than CPU intensive unless you do cluster computing or render a lot of stuff. The average server such as a web server or a file server doesn't use that much CPU and usually you are running 1-3 servers, not the hundreds that Facebook or Amazon would run.



And really, why is a VP complaining about this stuff? That he can't either afford custom solutions or spend the money buying more servers?</p></div><p>Agreed. Imagine this guy deploying regional Telco solutions. Facebook doesn't touch the level of tiering required for what people bitch about regarding their 3G options. Enterprise Data Centers, Call Center Suites, etc., are massive and when we really start moving everything to parallel blocks of code he might want to look at the architecture of the actual software.</p></div>
	</htmltext>
<tokenext>To build servers for companies like Facebook , and Amazon , and other people who are operating fairly homogeneous applications , the servers have to be cheap , and they have to be super power-efficient .
Hm , lets see... perhaps because Facebook and Amazon are niche markets ?
The average server is n't going to even need all the computing horsepower and the power efficiency is simply a drop in the bucket for most companies electrical bills .
The average server is going to be much more I/O intensive than CPU intensive unless you do cluster computing or render a lot of stuff .
The average server such as a web server or a file server does n't use that much CPU and usually you are running 1-3 servers , not the hundreds that Facebook or Amazon would run .
And really , why is a VP complaining about this stuff ?
That he ca n't either afford custom solutions or spend the money buying more servers ? Agreed .
Imagine this guy deploying regional Telco solutions .
Facebook does n't touch the level of tiering required for what people bitch about regarding their 3G options .
Enterprise Data Centers , Call Center Suites , etc. , are massive and when we really start moving everything to parallel blocks of code he might want to look at the architecture of the actual software .</tokentext>
<sentencetext> To build servers for companies like Facebook, and Amazon, and other people who are operating fairly homogeneous applications, the servers have to be cheap, and they have to be super power-efficient.
Hm, lets see... perhaps because Facebook and Amazon are niche markets?
The average server isn't going to even need all the computing horsepower and the power efficiency is simply a drop in the bucket for most companies electrical bills.
The average server is going to be much more I/O intensive than CPU intensive unless you do cluster computing or render a lot of stuff.
The average server such as a web server or a file server doesn't use that much CPU and usually you are running 1-3 servers, not the hundreds that Facebook or Amazon would run.
And really, why is a VP complaining about this stuff?
That he can't either afford custom solutions or spend the money buying more servers?Agreed.
Imagine this guy deploying regional Telco solutions.
Facebook doesn't touch the level of tiering required for what people bitch about regarding their 3G options.
Enterprise Data Centers, Call Center Suites, etc., are massive and when we really start moving everything to parallel blocks of code he might want to look at the architecture of the actual software.
	</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476941</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479699</id>
	<title>Re:Sun....</title>
	<author>zefrer</author>
	<datestamp>1246015440000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext><p>Uh not quite. Web servers sure, database? I think not. Having actually tried to do what you're suggesting, all those threads are no good if each of them sucks in terms of performance.</p><p>Webservers/fileservers sure but it is too expensive to consider unless there aren't any other viable, cheaper options - which there are plenty of</p></htmltext>
<tokenext>Uh not quite .
Web servers sure , database ?
I think not .
Having actually tried to do what you 're suggesting , all those threads are no good if each of them sucks in terms of performance.Webservers/fileservers sure but it is too expensive to consider unless there are n't any other viable , cheaper options - which there are plenty of</tokentext>
<sentencetext>Uh not quite.
Web servers sure, database?
I think not.
Having actually tried to do what you're suggesting, all those threads are no good if each of them sucks in terms of performance.Webservers/fileservers sure but it is too expensive to consider unless there aren't any other viable, cheaper options - which there are plenty of</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477079</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477571</id>
	<title>Depends on 'headroom' of other subsystems.</title>
	<author>Kadin2048</author>
	<datestamp>1245952200000</datestamp>
	<modclass>Informativ</modclass>
	<modscore>2</modscore>
	<htmltext><p>Not necessarily, no.</p><p>It's all about <b>how</b> CPU limited the workload is.</p><p>You might be running a program that's CPU limited on one processor, then upgrade the processor and discover that it's suddenly discover that instead of being CPU-bound, now you're memory-bound.  Or I/O bound. Or whatever.</p><p>Point is, just because you've hit the wall in terms of CPU doesn't mean you'll get a 50\% improvement with a 50\% increase in CPU<nobr> <wbr></nobr>... you'll only get that <b>if</b> all the rest of the server's systems have 50\% overhead to spare.  And in most cases they don't.  One of them will hit the performance wall before you return to being CPU-bound with the shiny new processor.</p><p>There are exceptions to this -- renderfarms, for instance, or some distributed HPC stuff -- where you really can reasonably expect to get 50\% more performance out of 50\% more CPU, but they're exceptions not the rule.</p></htmltext>
<tokenext>Not necessarily , no.It 's all about how CPU limited the workload is.You might be running a program that 's CPU limited on one processor , then upgrade the processor and discover that it 's suddenly discover that instead of being CPU-bound , now you 're memory-bound .
Or I/O bound .
Or whatever.Point is , just because you 've hit the wall in terms of CPU does n't mean you 'll get a 50 \ % improvement with a 50 \ % increase in CPU ... you 'll only get that if all the rest of the server 's systems have 50 \ % overhead to spare .
And in most cases they do n't .
One of them will hit the performance wall before you return to being CPU-bound with the shiny new processor.There are exceptions to this -- renderfarms , for instance , or some distributed HPC stuff -- where you really can reasonably expect to get 50 \ % more performance out of 50 \ % more CPU , but they 're exceptions not the rule .</tokentext>
<sentencetext>Not necessarily, no.It's all about how CPU limited the workload is.You might be running a program that's CPU limited on one processor, then upgrade the processor and discover that it's suddenly discover that instead of being CPU-bound, now you're memory-bound.
Or I/O bound.
Or whatever.Point is, just because you've hit the wall in terms of CPU doesn't mean you'll get a 50\% improvement with a 50\% increase in CPU ... you'll only get that if all the rest of the server's systems have 50\% overhead to spare.
And in most cases they don't.
One of them will hit the performance wall before you return to being CPU-bound with the shiny new processor.There are exceptions to this -- renderfarms, for instance, or some distributed HPC stuff -- where you really can reasonably expect to get 50\% more performance out of 50\% more CPU, but they're exceptions not the rule.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477167</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478733</id>
	<title>overblown company claims</title>
	<author>Anonymous</author>
	<datestamp>1246049040000</datestamp>
	<modclass>None</modclass>
	<modscore>0</modscore>
	<htmltext><p>I bet intel/amds performance claims are closer to reality than facebooks book value.</p></htmltext>
<tokenext>I bet intel/amds performance claims are closer to reality than facebooks book value .</tokentext>
<sentencetext>I bet intel/amds performance claims are closer to reality than facebooks book value.</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477581</id>
	<title>Strange...</title>
	<author>spire3661</author>
	<datestamp>1245952500000</datestamp>
	<modclass>Informativ</modclass>
	<modscore>4</modscore>
	<htmltext><p>Since when do we listen to manufacturer's claims? You take the new hardware, stress test it with your custom software, record results, plan servers accordingly. How hard is it really to commission a server design that meets your needs and then QA some prototypes?</p></htmltext>
<tokenext>Since when do we listen to manufacturer 's claims ?
You take the new hardware , stress test it with your custom software , record results , plan servers accordingly .
How hard is it really to commission a server design that meets your needs and then QA some prototypes ?</tokentext>
<sentencetext>Since when do we listen to manufacturer's claims?
You take the new hardware, stress test it with your custom software, record results, plan servers accordingly.
How hard is it really to commission a server design that meets your needs and then QA some prototypes?</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476973</id>
	<title>Well which is it?</title>
	<author>Anonymous</author>
	<datestamp>1245947520000</datestamp>
	<modclass>None</modclass>
	<modscore>0</modscore>
	<htmltext><p>Do you want your servers to be cheap or do you want them to be good?</p></htmltext>
<tokenext>Do you want your servers to be cheap or do you want them to be good ?</tokentext>
<sentencetext>Do you want your servers to be cheap or do you want them to be good?</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477743</id>
	<title>Re:Facebook's application is poorly coded</title>
	<author>dbcad7</author>
	<datestamp>1245954000000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext>I suppose that is true to a point.. but perhaps there is more to the story than processors.. for example, I imagine you would have similar experiences in surfing the internet when comparing the latest processor and something 5 years old<nobr> <wbr></nobr>.. when they are both connected using a 300 baud modem.. A faster processor doesn't necessarily mean a faster "system".. and software is part of that system.. mp3's don't play faster with a faster processor (by design) perhaps there is something in the design of their software that is limiting the performance.</htmltext>
<tokenext>I suppose that is true to a point.. but perhaps there is more to the story than processors.. for example , I imagine you would have similar experiences in surfing the internet when comparing the latest processor and something 5 years old .. when they are both connected using a 300 baud modem.. A faster processor does n't necessarily mean a faster " system " .. and software is part of that system.. mp3 's do n't play faster with a faster processor ( by design ) perhaps there is something in the design of their software that is limiting the performance .</tokentext>
<sentencetext>I suppose that is true to a point.. but perhaps there is more to the story than processors.. for example, I imagine you would have similar experiences in surfing the internet when comparing the latest processor and something 5 years old .. when they are both connected using a 300 baud modem.. A faster processor doesn't necessarily mean a faster "system".. and software is part of that system.. mp3's don't play faster with a faster processor (by design) perhaps there is something in the design of their software that is limiting the performance.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477085</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477207</id>
	<title>Re:Well I suppose...</title>
	<author>kzieli</author>
	<datestamp>1245949140000</datestamp>
	<modclass>Informativ</modclass>
	<modscore>3</modscore>
	<htmltext>There's actually 2 seperate points here

<ol>
<li>the latest CPU's don't seem to be any better in practice then the previous model.</li>

<li>Server OEM's are not delivering power efficient servers.</li>
</ol><p>

the two points are somewhat independent of each other. The second I suspect is due to their being a lack of any standard for power efficent servers. Google did it by running single voltage power supplies. A standard around something like this would be useful, and not just for servers I suspect.</p></htmltext>
<tokenext>There 's actually 2 seperate points here the latest CPU 's do n't seem to be any better in practice then the previous model .
Server OEM 's are not delivering power efficient servers .
the two points are somewhat independent of each other .
The second I suspect is due to their being a lack of any standard for power efficent servers .
Google did it by running single voltage power supplies .
A standard around something like this would be useful , and not just for servers I suspect .</tokentext>
<sentencetext>There's actually 2 seperate points here


the latest CPU's don't seem to be any better in practice then the previous model.
Server OEM's are not delivering power efficient servers.
the two points are somewhat independent of each other.
The second I suspect is due to their being a lack of any standard for power efficent servers.
Google did it by running single voltage power supplies.
A standard around something like this would be useful, and not just for servers I suspect.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476967</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477189</id>
	<title>A Familiar Tune from Facebook</title>
	<author>1sockchuck</author>
	<datestamp>1245948960000</datestamp>
	<modclass>Interestin</modclass>
	<modscore>3</modscore>
	<htmltext>This is becoming an annual event for Heiliger, who also <a href="http://www.datacenterknowledge.com/archives/2008/06/26/facebook-better-cloud-servers-needed/" title="datacenterknowledge.com">complained about server vendors</a> [datacenterknowledge.com] at GigaOm's Structure 08 conference last year. Facebook used to buy a lot of cloud-optimized gear from Rackable/SGI, but no longer appears on the list of their largest customers. Makes you wonder if they're not going to follow Google's lead and build their own servers.</htmltext>
<tokenext>This is becoming an annual event for Heiliger , who also complained about server vendors [ datacenterknowledge.com ] at GigaOm 's Structure 08 conference last year .
Facebook used to buy a lot of cloud-optimized gear from Rackable/SGI , but no longer appears on the list of their largest customers .
Makes you wonder if they 're not going to follow Google 's lead and build their own servers .</tokentext>
<sentencetext>This is becoming an annual event for Heiliger, who also complained about server vendors [datacenterknowledge.com] at GigaOm's Structure 08 conference last year.
Facebook used to buy a lot of cloud-optimized gear from Rackable/SGI, but no longer appears on the list of their largest customers.
Makes you wonder if they're not going to follow Google's lead and build their own servers.</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478019</id>
	<title>Jonathan Heiliger can kiss my rosy ass...</title>
	<author>hyades1</author>
	<datestamp>1245956340000</datestamp>
	<modclass>Troll</modclass>
	<modscore>0</modscore>
	<htmltext><p> Ever since I accepted their invitation to use my (very unusual) name openly on Facebook, every single variation on that name has been spammed. This Heilinger jackass can take his "harsh words" about the performance of people who do real hardware work and shove them deep and hard.  You can't bullshit a chip, you can merely mis-state its performance.  When you lie about how you'll protect information people give you in trust, you're pretty much a douchebag. Mr. Heiliger, you are a douchebag. </p></htmltext>
<tokenext>Ever since I accepted their invitation to use my ( very unusual ) name openly on Facebook , every single variation on that name has been spammed .
This Heilinger jackass can take his " harsh words " about the performance of people who do real hardware work and shove them deep and hard .
You ca n't bullshit a chip , you can merely mis-state its performance .
When you lie about how you 'll protect information people give you in trust , you 're pretty much a douchebag .
Mr. Heiliger , you are a douchebag .</tokentext>
<sentencetext> Ever since I accepted their invitation to use my (very unusual) name openly on Facebook, every single variation on that name has been spammed.
This Heilinger jackass can take his "harsh words" about the performance of people who do real hardware work and shove them deep and hard.
You can't bullshit a chip, you can merely mis-state its performance.
When you lie about how you'll protect information people give you in trust, you're pretty much a douchebag.
Mr. Heiliger, you are a douchebag. </sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479295</id>
	<title>Hey! Leave lying about harware</title>
	<author>Anonymous</author>
	<datestamp>1246010940000</datestamp>
	<modclass>None</modclass>
	<modscore>0</modscore>
	<htmltext><p>to Apple. That is the only thing that they are good at.</p></htmltext>
<tokenext>to Apple .
That is the only thing that they are good at .</tokentext>
<sentencetext>to Apple.
That is the only thing that they are good at.</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477107</id>
	<title>I wonder if he's not thinking about scaling enough</title>
	<author>filesiteguy</author>
	<datestamp>1245948360000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext>Honestly, I think that his arguments fall flat. Though I like FB as much as anybody, and I feel for someone dealing with massive performance needs - I only have a paltry 30 servers running my main application - I wonder if he's being realistic.<p>

<i>There's a pretty simple answer for scaling infrastructure. It's, 'Don't be cheap,'" Heiliger said. He added that Facebook does drive hard bargains with its hardware and software infrastructure suppliers, and is careful not to overbuy.</i> </p><p>

I remember reading about how Amazon does it. They have clusters of servers running whatever OS suits the particular person having written the portion of code being used and will blow through something like 100 dead servers a day. IIRC, when you load a page from Amazon you get content delivered by 20+ servers onto one web page.</p><p>

Maybe he just needs to scale out.</p><p>

Or - I just noticed an unused AS390 in the server room today. Apparently the Z890 that replaced it is also going to be replaced by a new z9 machine. He could bundle some apps on the z890 or the 390.</p><p>

I wonder if you can run<nobr> <wbr></nobr>.net or Java under OS/390 or MVS...</p></htmltext>
<tokenext>Honestly , I think that his arguments fall flat .
Though I like FB as much as anybody , and I feel for someone dealing with massive performance needs - I only have a paltry 30 servers running my main application - I wonder if he 's being realistic .
There 's a pretty simple answer for scaling infrastructure .
It 's , 'Do n't be cheap, ' " Heiliger said .
He added that Facebook does drive hard bargains with its hardware and software infrastructure suppliers , and is careful not to overbuy .
I remember reading about how Amazon does it .
They have clusters of servers running whatever OS suits the particular person having written the portion of code being used and will blow through something like 100 dead servers a day .
IIRC , when you load a page from Amazon you get content delivered by 20 + servers onto one web page .
Maybe he just needs to scale out .
Or - I just noticed an unused AS390 in the server room today .
Apparently the Z890 that replaced it is also going to be replaced by a new z9 machine .
He could bundle some apps on the z890 or the 390 .
I wonder if you can run .net or Java under OS/390 or MVS.. .</tokentext>
<sentencetext>Honestly, I think that his arguments fall flat.
Though I like FB as much as anybody, and I feel for someone dealing with massive performance needs - I only have a paltry 30 servers running my main application - I wonder if he's being realistic.
There's a pretty simple answer for scaling infrastructure.
It's, 'Don't be cheap,'" Heiliger said.
He added that Facebook does drive hard bargains with its hardware and software infrastructure suppliers, and is careful not to overbuy.
I remember reading about how Amazon does it.
They have clusters of servers running whatever OS suits the particular person having written the portion of code being used and will blow through something like 100 dead servers a day.
IIRC, when you load a page from Amazon you get content delivered by 20+ servers onto one web page.
Maybe he just needs to scale out.
Or - I just noticed an unused AS390 in the server room today.
Apparently the Z890 that replaced it is also going to be replaced by a new z9 machine.
He could bundle some apps on the z890 or the 390.
I wonder if you can run .net or Java under OS/390 or MVS...</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28485751</id>
	<title>Re:You're Computin' for a Shootin' Mister</title>
	<author>AmberBlackCat</author>
	<datestamp>1246042560000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext>I've heard enough "hardware engineers" criticizing "software engineers" that I don't care if a software guy criticizes hardware guys for a while. Neither is really qualified to judge the other.</htmltext>
<tokenext>I 've heard enough " hardware engineers " criticizing " software engineers " that I do n't care if a software guy criticizes hardware guys for a while .
Neither is really qualified to judge the other .</tokentext>
<sentencetext>I've heard enough "hardware engineers" criticizing "software engineers" that I don't care if a software guy criticizes hardware guys for a while.
Neither is really qualified to judge the other.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476905</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477011</id>
	<title>PHP</title>
	<author>Anonymous</author>
	<datestamp>1245947700000</datestamp>
	<modclass>Interestin</modclass>
	<modscore>3</modscore>
	<htmltext><blockquote><div><p>And we're, literally in real time right now, trying to figure out why that is,' Heiliger said.</p></div></blockquote><p>
It's because your shitty website doesn't have a single line of compiled code. PHP only goes so far.</p></div>
	</htmltext>
<tokenext>And we 're , literally in real time right now , trying to figure out why that is, ' Heiliger said .
It 's because your shitty website does n't have a single line of compiled code .
PHP only goes so far .</tokentext>
<sentencetext>And we're, literally in real time right now, trying to figure out why that is,' Heiliger said.
It's because your shitty website doesn't have a single line of compiled code.
PHP only goes so far.
	</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28483113</id>
	<title>Re:WTF?</title>
	<author>ITJC68</author>
	<datestamp>1246033020000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext>Excellent post. From someone who obviously read the article and understood what the complaint was. I too believe that Intel and AMD stretch what their "new and improved" CPUs deliver. It is the same in the desktop arena. Dual core versus Quad core. People buy the cheapest quad thinking they will be better off versus the faster (GHZ) dual core. Until the market can make good utilization of dual cores for the desktop (they are but not enough to warrant 3 or 4 cores) it is pointless. I always use Toms Hardware for real world comparison and it has always been a very good resource.</htmltext>
<tokenext>Excellent post .
From someone who obviously read the article and understood what the complaint was .
I too believe that Intel and AMD stretch what their " new and improved " CPUs deliver .
It is the same in the desktop arena .
Dual core versus Quad core .
People buy the cheapest quad thinking they will be better off versus the faster ( GHZ ) dual core .
Until the market can make good utilization of dual cores for the desktop ( they are but not enough to warrant 3 or 4 cores ) it is pointless .
I always use Toms Hardware for real world comparison and it has always been a very good resource .</tokentext>
<sentencetext>Excellent post.
From someone who obviously read the article and understood what the complaint was.
I too believe that Intel and AMD stretch what their "new and improved" CPUs deliver.
It is the same in the desktop arena.
Dual core versus Quad core.
People buy the cheapest quad thinking they will be better off versus the faster (GHZ) dual core.
Until the market can make good utilization of dual cores for the desktop (they are but not enough to warrant 3 or 4 cores) it is pointless.
I always use Toms Hardware for real world comparison and it has always been a very good resource.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477361</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28486019</id>
	<title>Coolthreads</title>
	<author>Anonymous</author>
	<datestamp>1246043820000</datestamp>
	<modclass>None</modclass>
	<modscore>0</modscore>
	<htmltext><p>If they want performance per watt, why don't they use sun ultrasparc t1 or t2?  Massive throughput specifically optimized for web applications.</p></htmltext>
<tokenext>If they want performance per watt , why do n't they use sun ultrasparc t1 or t2 ?
Massive throughput specifically optimized for web applications .</tokentext>
<sentencetext>If they want performance per watt, why don't they use sun ultrasparc t1 or t2?
Massive throughput specifically optimized for web applications.</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478683</id>
	<title>99.99\%</title>
	<author>Colin Smith</author>
	<datestamp>1246048620000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext><p>Of server manufacturers customers are not Google, or Amazon or Facebook. The VP doesn't get that he's just not that important....<br>
&nbsp; &nbsp;</p></htmltext>
<tokenext>Of server manufacturers customers are not Google , or Amazon or Facebook .
The VP does n't get that he 's just not that important... .    </tokentext>
<sentencetext>Of server manufacturers customers are not Google, or Amazon or Facebook.
The VP doesn't get that he's just not that important....
   </sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477819</id>
	<title>Re:Facebook's application is poorly coded</title>
	<author>Stormie</author>
	<datestamp>1245954360000</datestamp>
	<modclass>Interestin</modclass>
	<modscore>4</modscore>
	<htmltext><blockquote><div><p>I have heard from some reliable sources that Facebook and Twitter's backend applications are poorly written.</p></div></blockquote><p>Given the quality of Facebook's developer API (it's horrible), I'd be amazed if the back-end of the actual site <i>wasn't</i> poorly written.</p></div>
	</htmltext>
<tokenext>I have heard from some reliable sources that Facebook and Twitter 's backend applications are poorly written.Given the quality of Facebook 's developer API ( it 's horrible ) , I 'd be amazed if the back-end of the actual site was n't poorly written .</tokentext>
<sentencetext>I have heard from some reliable sources that Facebook and Twitter's backend applications are poorly written.Given the quality of Facebook's developer API (it's horrible), I'd be amazed if the back-end of the actual site wasn't poorly written.
	</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476965</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477489</id>
	<title>Re:PHP</title>
	<author>Anonymous</author>
	<datestamp>1245951420000</datestamp>
	<modclass>None</modclass>
	<modscore>0</modscore>
	<htmltext><p>You're a tool if you think they aren't using the Zend optimizer or APC keep the compiled byte-code cached.  In the end, it's the same as running Java on a JVM.</p></htmltext>
<tokenext>You 're a tool if you think they are n't using the Zend optimizer or APC keep the compiled byte-code cached .
In the end , it 's the same as running Java on a JVM .</tokentext>
<sentencetext>You're a tool if you think they aren't using the Zend optimizer or APC keep the compiled byte-code cached.
In the end, it's the same as running Java on a JVM.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477011</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477211</id>
	<title>Re:Hm...</title>
	<author>vux984</author>
	<datestamp>1245949140000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext><p><i>That he can't either afford custom solutions or spend the money buying more servers?</i></p><p>Tell me again what Facebook's revenue model is...??</p></htmltext>
<tokenext>That he ca n't either afford custom solutions or spend the money buying more servers ? Tell me again what Facebook 's revenue model is... ?
?</tokentext>
<sentencetext>That he can't either afford custom solutions or spend the money buying more servers?Tell me again what Facebook's revenue model is...?
?</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476941</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28480667</id>
	<title>Re:Facebook's application is poorly coded</title>
	<author>Anonymous</author>
	<datestamp>1246024800000</datestamp>
	<modclass>None</modclass>
	<modscore>0</modscore>
	<htmltext><p>No, depends on where the bottleneck is.  If it's in Memory or I/O no amount of CPU speed increase will help.</p><p>Also it's entirely possible Facebook is growing beyond rational means. There is a limit to technology today, and it's not a matter of not trying hard enough to invent faster computers, it's just what are tech is capable of.</p><p>Maybe if Facebook's home page didn't have 3 million images, apps, and other random fucking annoying popups they'd have a lower hit.</p></htmltext>
<tokenext>No , depends on where the bottleneck is .
If it 's in Memory or I/O no amount of CPU speed increase will help.Also it 's entirely possible Facebook is growing beyond rational means .
There is a limit to technology today , and it 's not a matter of not trying hard enough to invent faster computers , it 's just what are tech is capable of.Maybe if Facebook 's home page did n't have 3 million images , apps , and other random fucking annoying popups they 'd have a lower hit .</tokentext>
<sentencetext>No, depends on where the bottleneck is.
If it's in Memory or I/O no amount of CPU speed increase will help.Also it's entirely possible Facebook is growing beyond rational means.
There is a limit to technology today, and it's not a matter of not trying hard enough to invent faster computers, it's just what are tech is capable of.Maybe if Facebook's home page didn't have 3 million images, apps, and other random fucking annoying popups they'd have a lower hit.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477085</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478623</id>
	<title>and..</title>
	<author>Anonymous</author>
	<datestamp>1246048080000</datestamp>
	<modclass>None</modclass>
	<modscore>0</modscore>
	<htmltext><p><div class="quote"><p>the servers have to be cheap, and they have to be super power-efficient.</p></div><p>yeah, and self-replicating too!</p></div>
	</htmltext>
<tokenext>the servers have to be cheap , and they have to be super power-efficient.yeah , and self-replicating too !</tokentext>
<sentencetext>the servers have to be cheap, and they have to be super power-efficient.yeah, and self-replicating too!
	</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476941</id>
	<title>Hm...</title>
	<author>Darkness404</author>
	<datestamp>1245947280000</datestamp>
	<modclass>Insightful</modclass>
	<modscore>3</modscore>
	<htmltext><p><div class="quote"><p> To build servers for companies like Facebook, and Amazon, and other people who are operating fairly homogeneous applications, the servers have to be cheap, and they have to be super power-efficient.</p> </div><p>

Hm, lets see... perhaps because Facebook and Amazon are niche markets? The average server isn't going to even need all the computing horsepower and the power efficiency is simply a drop in the bucket for most companies electrical bills. The average server is going to be much more I/O intensive than CPU intensive unless you do cluster computing or render a lot of stuff. The average server such as a web server or a file server doesn't use that much CPU and usually you are running 1-3 servers, not the hundreds that Facebook or Amazon would run. <br> <br>

And really, why is a VP complaining about this stuff? That he can't either afford custom solutions or spend the money buying more servers?</p></div>
	</htmltext>
<tokenext>To build servers for companies like Facebook , and Amazon , and other people who are operating fairly homogeneous applications , the servers have to be cheap , and they have to be super power-efficient .
Hm , lets see... perhaps because Facebook and Amazon are niche markets ?
The average server is n't going to even need all the computing horsepower and the power efficiency is simply a drop in the bucket for most companies electrical bills .
The average server is going to be much more I/O intensive than CPU intensive unless you do cluster computing or render a lot of stuff .
The average server such as a web server or a file server does n't use that much CPU and usually you are running 1-3 servers , not the hundreds that Facebook or Amazon would run .
And really , why is a VP complaining about this stuff ?
That he ca n't either afford custom solutions or spend the money buying more servers ?</tokentext>
<sentencetext> To build servers for companies like Facebook, and Amazon, and other people who are operating fairly homogeneous applications, the servers have to be cheap, and they have to be super power-efficient.
Hm, lets see... perhaps because Facebook and Amazon are niche markets?
The average server isn't going to even need all the computing horsepower and the power efficiency is simply a drop in the bucket for most companies electrical bills.
The average server is going to be much more I/O intensive than CPU intensive unless you do cluster computing or render a lot of stuff.
The average server such as a web server or a file server doesn't use that much CPU and usually you are running 1-3 servers, not the hundreds that Facebook or Amazon would run.
And really, why is a VP complaining about this stuff?
That he can't either afford custom solutions or spend the money buying more servers?
	</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477079</id>
	<title>Sun....</title>
	<author>Anonymous</author>
	<datestamp>1245948180000</datestamp>
	<modclass>Insightful</modclass>
	<modscore>2</modscore>
	<htmltext>Its the next logical solution... Those T5440 servers with 256 processing threads are MONSTERS in terms of handling simultaneous connections which make them very good web servers, database servers, and file servers, all of which means they are very good for a company who's product is a website.</htmltext>
<tokenext>Its the next logical solution... Those T5440 servers with 256 processing threads are MONSTERS in terms of handling simultaneous connections which make them very good web servers , database servers , and file servers , all of which means they are very good for a company who 's product is a website .</tokentext>
<sentencetext>Its the next logical solution... Those T5440 servers with 256 processing threads are MONSTERS in terms of handling simultaneous connections which make them very good web servers, database servers, and file servers, all of which means they are very good for a company who's product is a website.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476967</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28484715</id>
	<title>When does it make sense...?</title>
	<author>Targon</author>
	<datestamp>1246038600000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext><p>AMD and Intel are in the position where they need to make products used by a large number of customers.   As a result of this, their primary focus will be on making products that will draw in the greatest number of customers.    As we saw with Itanium, if the customer base is not large enough, the R&amp;D costs will never be made up.</p><p>So, when does it make sense for a chip and motherboard supplier to make a product with only one or two POTENTIAL customers in mind?    Never is the answer that comes to mind.     Both AMD and Intel MUST spend their resources on making products that will result in a net profit.</p><p>So, Facebook and web servers, and database engines...it should be possible for AMD or Intel to make a platform with these specific applications in mind, but the cost for such a specific product to be developed when there are very few potential customers that would want it would be small.    Potential is the key, because I am sure that if Facebook approached AMD or Intel and wanted a fixed-purpose product to be developed, they would be happy to do it for the right price.</p><p>When a company makes a motherboard, the focus is to make a product that will get enough interest and sales to make a profit.   As a result, we see motherboards with extra PCI, PCI Express, memory, USB, SATA, and other connectors than most people would actually need.    If it gets cut back to only what would be needed for a specific customer, then the machine would probably perform better.    Expecting a product aimed at a large number of people to be perfectly optimized and customized for any one specific purpose is foolish.</p><p>And of course, you have the limitations inherent in any system, including bandwidth between components, ethernet controllers, and how much CPU power may be used for things like USB, SATA, and ethernet.    When you buy a $75 motherboard and expect the performance of a $250 motherboard, you are pretty much guaranteed to be disappointed.</p></htmltext>
<tokenext>AMD and Intel are in the position where they need to make products used by a large number of customers .
As a result of this , their primary focus will be on making products that will draw in the greatest number of customers .
As we saw with Itanium , if the customer base is not large enough , the R&amp;D costs will never be made up.So , when does it make sense for a chip and motherboard supplier to make a product with only one or two POTENTIAL customers in mind ?
Never is the answer that comes to mind .
Both AMD and Intel MUST spend their resources on making products that will result in a net profit.So , Facebook and web servers , and database engines...it should be possible for AMD or Intel to make a platform with these specific applications in mind , but the cost for such a specific product to be developed when there are very few potential customers that would want it would be small .
Potential is the key , because I am sure that if Facebook approached AMD or Intel and wanted a fixed-purpose product to be developed , they would be happy to do it for the right price.When a company makes a motherboard , the focus is to make a product that will get enough interest and sales to make a profit .
As a result , we see motherboards with extra PCI , PCI Express , memory , USB , SATA , and other connectors than most people would actually need .
If it gets cut back to only what would be needed for a specific customer , then the machine would probably perform better .
Expecting a product aimed at a large number of people to be perfectly optimized and customized for any one specific purpose is foolish.And of course , you have the limitations inherent in any system , including bandwidth between components , ethernet controllers , and how much CPU power may be used for things like USB , SATA , and ethernet .
When you buy a $ 75 motherboard and expect the performance of a $ 250 motherboard , you are pretty much guaranteed to be disappointed .</tokentext>
<sentencetext>AMD and Intel are in the position where they need to make products used by a large number of customers.
As a result of this, their primary focus will be on making products that will draw in the greatest number of customers.
As we saw with Itanium, if the customer base is not large enough, the R&amp;D costs will never be made up.So, when does it make sense for a chip and motherboard supplier to make a product with only one or two POTENTIAL customers in mind?
Never is the answer that comes to mind.
Both AMD and Intel MUST spend their resources on making products that will result in a net profit.So, Facebook and web servers, and database engines...it should be possible for AMD or Intel to make a platform with these specific applications in mind, but the cost for such a specific product to be developed when there are very few potential customers that would want it would be small.
Potential is the key, because I am sure that if Facebook approached AMD or Intel and wanted a fixed-purpose product to be developed, they would be happy to do it for the right price.When a company makes a motherboard, the focus is to make a product that will get enough interest and sales to make a profit.
As a result, we see motherboards with extra PCI, PCI Express, memory, USB, SATA, and other connectors than most people would actually need.
If it gets cut back to only what would be needed for a specific customer, then the machine would probably perform better.
Expecting a product aimed at a large number of people to be perfectly optimized and customized for any one specific purpose is foolish.And of course, you have the limitations inherent in any system, including bandwidth between components, ethernet controllers, and how much CPU power may be used for things like USB, SATA, and ethernet.
When you buy a $75 motherboard and expect the performance of a $250 motherboard, you are pretty much guaranteed to be disappointed.</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477337</id>
	<title>Re:Facebook's application is poorly coded</title>
	<author>Necroman</author>
	<datestamp>1245950040000</datestamp>
	<modclass>Interestin</modclass>
	<modscore>4</modscore>
	<htmltext><p>One of the server techs from Twitter was at SXSW 2 years and gave some details about how their backend servers worked.  If I remember correctly (there were 4 sites on the panel, so I may be confusing them with someone else), the original code was written in Ruby on Rails which did not scale well to the multi-server systems that they had setup.  They have spent a lot of time improving their code over the years, but from what I could tell, their initial implementation wasn't the most thought out thing in the world.</p></htmltext>
<tokenext>One of the server techs from Twitter was at SXSW 2 years and gave some details about how their backend servers worked .
If I remember correctly ( there were 4 sites on the panel , so I may be confusing them with someone else ) , the original code was written in Ruby on Rails which did not scale well to the multi-server systems that they had setup .
They have spent a lot of time improving their code over the years , but from what I could tell , their initial implementation was n't the most thought out thing in the world .</tokentext>
<sentencetext>One of the server techs from Twitter was at SXSW 2 years and gave some details about how their backend servers worked.
If I remember correctly (there were 4 sites on the panel, so I may be confusing them with someone else), the original code was written in Ruby on Rails which did not scale well to the multi-server systems that they had setup.
They have spent a lot of time improving their code over the years, but from what I could tell, their initial implementation wasn't the most thought out thing in the world.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476965</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477245</id>
	<title>Having a Bad Day?</title>
	<author>FranTaylor</author>
	<datestamp>1245949380000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext><p>He woke up on the wrong side of the bed, and then he had to sign the check for the electric bill.</p><p>He's just grumpy.</p></htmltext>
<tokenext>He woke up on the wrong side of the bed , and then he had to sign the check for the electric bill.He 's just grumpy .</tokentext>
<sentencetext>He woke up on the wrong side of the bed, and then he had to sign the check for the electric bill.He's just grumpy.</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476965</id>
	<title>Facebook's application is poorly coded</title>
	<author>jsimon12</author>
	<datestamp>1245947460000</datestamp>
	<modclass>Insightful</modclass>
	<modscore>3</modscore>
	<htmltext><p>I have heard from some reliable sources that Facebook and Twitter's backend applications are poorly written.</p><p>Are Intel and AMD's claims overblown, sure what hardware manufacter doesn't cherry pick performance claims.</p><p>But I don't care what sort of hardware you through at crap code you are always going to get crap performance.</p></htmltext>
<tokenext>I have heard from some reliable sources that Facebook and Twitter 's backend applications are poorly written.Are Intel and AMD 's claims overblown , sure what hardware manufacter does n't cherry pick performance claims.But I do n't care what sort of hardware you through at crap code you are always going to get crap performance .</tokentext>
<sentencetext>I have heard from some reliable sources that Facebook and Twitter's backend applications are poorly written.Are Intel and AMD's claims overblown, sure what hardware manufacter doesn't cherry pick performance claims.But I don't care what sort of hardware you through at crap code you are always going to get crap performance.</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28480937</id>
	<title>Re:WTF?</title>
	<author>maraist</author>
	<datestamp>1246026000000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext>"If the application was purely CPU bound, and Y wasn't giving me 35\% more than X, I'd complain."<br><br>Huh?  Why should a specific APP be expected to achieve that max theoretical performance increase that a contrived benchmark (specInt, etc) rates?  While this class of bench marks is based on 'real world apps', nobody uses those apps anymore.<br><br>You can write two 'CPU bound' apps that have completely different branch-patterns and memory-patterns.  One with great instruction-sequentially and mem-locality (and thus deep-pipeline utilization and highly cacheable even though it exceeds the cache size) and one which makes effectively random walks through memory with contingent random code-jumps (such that DRAM banks are constantly powered up and down).  You'll totally trash the memory performance and thus your memory bus bandwidth is bumpkiss.<br><br>So if you are currently achieving say 60\% max-theoretical throughput on CPU-1, then what the author is suggesting is that CPU-2 should achieve 135\% * (60\% * CPU-1).  But the design decisions of the new CPU might be such that the particular code-flow of the target app are actually WORSE (i.e. a longer pipeline visa v. the Pentium 4).  The issue might be more subtle, such as fooling the branch-predictor logic (which is complex).  Or it could be that the data-dependent random memory walks are such that the parallel memory channels CAN'T be utilized effectively. So BW and parallel chips don't make up for fundamentally rate-limited random-access latency.<br><br>The point is, ALWAYS sample the new hardware and make mass purchasing choices based on your personal cost-benefit analysis.. Bitching to the vendor is unfruitful - choose alternate purchasing strategies.</htmltext>
<tokenext>" If the application was purely CPU bound , and Y was n't giving me 35 \ % more than X , I 'd complain. " Huh ?
Why should a specific APP be expected to achieve that max theoretical performance increase that a contrived benchmark ( specInt , etc ) rates ?
While this class of bench marks is based on 'real world apps ' , nobody uses those apps anymore.You can write two 'CPU bound ' apps that have completely different branch-patterns and memory-patterns .
One with great instruction-sequentially and mem-locality ( and thus deep-pipeline utilization and highly cacheable even though it exceeds the cache size ) and one which makes effectively random walks through memory with contingent random code-jumps ( such that DRAM banks are constantly powered up and down ) .
You 'll totally trash the memory performance and thus your memory bus bandwidth is bumpkiss.So if you are currently achieving say 60 \ % max-theoretical throughput on CPU-1 , then what the author is suggesting is that CPU-2 should achieve 135 \ % * ( 60 \ % * CPU-1 ) .
But the design decisions of the new CPU might be such that the particular code-flow of the target app are actually WORSE ( i.e .
a longer pipeline visa v. the Pentium 4 ) .
The issue might be more subtle , such as fooling the branch-predictor logic ( which is complex ) .
Or it could be that the data-dependent random memory walks are such that the parallel memory channels CA N'T be utilized effectively .
So BW and parallel chips do n't make up for fundamentally rate-limited random-access latency.The point is , ALWAYS sample the new hardware and make mass purchasing choices based on your personal cost-benefit analysis.. Bitching to the vendor is unfruitful - choose alternate purchasing strategies .</tokentext>
<sentencetext>"If the application was purely CPU bound, and Y wasn't giving me 35\% more than X, I'd complain."Huh?
Why should a specific APP be expected to achieve that max theoretical performance increase that a contrived benchmark (specInt, etc) rates?
While this class of bench marks is based on 'real world apps', nobody uses those apps anymore.You can write two 'CPU bound' apps that have completely different branch-patterns and memory-patterns.
One with great instruction-sequentially and mem-locality (and thus deep-pipeline utilization and highly cacheable even though it exceeds the cache size) and one which makes effectively random walks through memory with contingent random code-jumps (such that DRAM banks are constantly powered up and down).
You'll totally trash the memory performance and thus your memory bus bandwidth is bumpkiss.So if you are currently achieving say 60\% max-theoretical throughput on CPU-1, then what the author is suggesting is that CPU-2 should achieve 135\% * (60\% * CPU-1).
But the design decisions of the new CPU might be such that the particular code-flow of the target app are actually WORSE (i.e.
a longer pipeline visa v. the Pentium 4).
The issue might be more subtle, such as fooling the branch-predictor logic (which is complex).
Or it could be that the data-dependent random memory walks are such that the parallel memory channels CAN'T be utilized effectively.
So BW and parallel chips don't make up for fundamentally rate-limited random-access latency.The point is, ALWAYS sample the new hardware and make mass purchasing choices based on your personal cost-benefit analysis.. Bitching to the vendor is unfruitful - choose alternate purchasing strategies.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477763</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476989</id>
	<title>Something about his arguement doesn't work</title>
	<author>joeflies</author>
	<datestamp>1245947580000</datestamp>
	<modclass>Insightful</modclass>
	<modscore>5</modscore>
	<htmltext><p>1) Facebook &amp; Amazon need cheap, power efficient systems<br>2) Intel and AMD aren't measuring up with processors to power these systems<br>3) However, Google has systems appropriate for this use (presumably using Intel or AMD processors)</p><p>If that's his argument, then it would seem that the real conclusion is that Facebook can't build systems as good as Google's, even though they are using the same processor technology.</p></htmltext>
<tokenext>1 ) Facebook &amp; Amazon need cheap , power efficient systems2 ) Intel and AMD are n't measuring up with processors to power these systems3 ) However , Google has systems appropriate for this use ( presumably using Intel or AMD processors ) If that 's his argument , then it would seem that the real conclusion is that Facebook ca n't build systems as good as Google 's , even though they are using the same processor technology .</tokentext>
<sentencetext>1) Facebook &amp; Amazon need cheap, power efficient systems2) Intel and AMD aren't measuring up with processors to power these systems3) However, Google has systems appropriate for this use (presumably using Intel or AMD processors)If that's his argument, then it would seem that the real conclusion is that Facebook can't build systems as good as Google's, even though they are using the same processor technology.</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479537</id>
	<title>Re:Facebook's application is poorly coded</title>
	<author>Anonymous</author>
	<datestamp>1246013400000</datestamp>
	<modclass>Insightful</modclass>
	<modscore>2</modscore>
	<htmltext><p>Not "unreasonable", but possibly naive and inexperienced, depending on the details.</p><p>Crap code that bottlenecks a CPU often will not scale as well as good code. It involved bad synchronization, other contention, spinning loops, and memory bandwidth limits. It is often NUMA-unfriendly. It often interacts poorly with the other resources, such as I/O.</p></htmltext>
<tokenext>Not " unreasonable " , but possibly naive and inexperienced , depending on the details.Crap code that bottlenecks a CPU often will not scale as well as good code .
It involved bad synchronization , other contention , spinning loops , and memory bandwidth limits .
It is often NUMA-unfriendly .
It often interacts poorly with the other resources , such as I/O .</tokentext>
<sentencetext>Not "unreasonable", but possibly naive and inexperienced, depending on the details.Crap code that bottlenecks a CPU often will not scale as well as good code.
It involved bad synchronization, other contention, spinning loops, and memory bandwidth limits.
It is often NUMA-unfriendly.
It often interacts poorly with the other resources, such as I/O.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477167</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28487125</id>
	<title>Re:Facebook's application is poorly coded</title>
	<author>relguj9</author>
	<datestamp>1246049160000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext><p><div class="quote"><p>is it unreasonable to think that buying CPUs marketed as 50\% faster should give a 50\% performance increase?</p></div><p>Yes, it is unreasonable, entirely.  Even if the CPU is the only physical bottleneck, you still can't expect X\% clock speed to correlate to X\% performance increase.  That's pretty basic computer/software engineering.  You <i>can</i> make reasonable estimations, but the only way to know what your gains will be is to test it.<br> <br>

Anymore, 35\% faster can mean any number of things as well, it's not just straight speed of operation execution.  35\% faster could mean that they increased the efficiency of some operations, bumped up the 1st level cache, increased clock speed or any combination of those plus other things.  The 35\% is probably just an attempt to put all of the gains into a single number that really can't be presumed to be all that accurate.<br> <br>

Basically, you'd need to know exactly what was changed or upgraded from your previous CPU and figure out how each of those upgrades will affect your particular configuration before making any reasonable estimation of performance gains.  It would be absolutely <i>impossible</i> for AMD or Intel to do this for you.</p></div>
	</htmltext>
<tokenext>is it unreasonable to think that buying CPUs marketed as 50 \ % faster should give a 50 \ % performance increase ? Yes , it is unreasonable , entirely .
Even if the CPU is the only physical bottleneck , you still ca n't expect X \ % clock speed to correlate to X \ % performance increase .
That 's pretty basic computer/software engineering .
You can make reasonable estimations , but the only way to know what your gains will be is to test it .
Anymore , 35 \ % faster can mean any number of things as well , it 's not just straight speed of operation execution .
35 \ % faster could mean that they increased the efficiency of some operations , bumped up the 1st level cache , increased clock speed or any combination of those plus other things .
The 35 \ % is probably just an attempt to put all of the gains into a single number that really ca n't be presumed to be all that accurate .
Basically , you 'd need to know exactly what was changed or upgraded from your previous CPU and figure out how each of those upgrades will affect your particular configuration before making any reasonable estimation of performance gains .
It would be absolutely impossible for AMD or Intel to do this for you .</tokentext>
<sentencetext>is it unreasonable to think that buying CPUs marketed as 50\% faster should give a 50\% performance increase?Yes, it is unreasonable, entirely.
Even if the CPU is the only physical bottleneck, you still can't expect X\% clock speed to correlate to X\% performance increase.
That's pretty basic computer/software engineering.
You can make reasonable estimations, but the only way to know what your gains will be is to test it.
Anymore, 35\% faster can mean any number of things as well, it's not just straight speed of operation execution.
35\% faster could mean that they increased the efficiency of some operations, bumped up the 1st level cache, increased clock speed or any combination of those plus other things.
The 35\% is probably just an attempt to put all of the gains into a single number that really can't be presumed to be all that accurate.
Basically, you'd need to know exactly what was changed or upgraded from your previous CPU and figure out how each of those upgrades will affect your particular configuration before making any reasonable estimation of performance gains.
It would be absolutely impossible for AMD or Intel to do this for you.
	</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477167</parent>
</comment>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_3</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476921
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477069
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477361
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477977
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478473
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_22</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476921
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477069
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477361
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478839
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_13</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476905
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28480449
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_73</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477107
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477441
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_87</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476921
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477069
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477361
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477851
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_12</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476965
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477167
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479537
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_64</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477115
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477967
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_37</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476905
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478981
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_40</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476965
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477337
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_9</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476905
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28480675
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_42</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476905
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478273
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_2</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476965
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477085
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478095
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_67</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476965
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477085
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477845
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_28</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476941
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477211
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28480349
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_88</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477107
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477457
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_61</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476965
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477085
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478127
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_32</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476989
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477675
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479535
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_18</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476941
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477411
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28484189
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_11</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476941
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477411
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477873
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_25</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477451
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28488783
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_85</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476941
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477467
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_59</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476941
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478669
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_62</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476965
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477085
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28480457
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_53</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476965
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477819
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_8</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477115
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479953
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_49</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476905
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477785
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_52</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476941
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477411
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28481227
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_7</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476941
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477411
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479461
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_31</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476905
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28483853
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_26</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476965
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477085
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479831
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_54</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476941
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477411
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479135
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_17</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476921
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477069
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477361
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478305
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_77</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476941
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477411
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28482555
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_30</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476921
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477069
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477361
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477763
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479401
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_79</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476905
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28480465
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_16</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476965
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477831
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_82</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476965
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477085
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477373
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_78</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476965
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477965
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_23</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476941
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477411
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28489715
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_72</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476965
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477999
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_46</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476905
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28485751
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_6</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476905
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479893
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_51</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476921
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477069
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477361
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477763
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479875
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_65</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476921
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477335
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477751
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_36</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476921
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477069
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28481355
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_41</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476905
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477081
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_15</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476921
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477069
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477361
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477977
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28480303
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_43</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476921
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477335
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477523
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_1</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476905
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28480557
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_29</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476989
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478265
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_20</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476941
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477535
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_80</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476967
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28482179
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_71</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476989
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28482821
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_10</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477011
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477489
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_68</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476965
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477167
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28487125
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_70</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476967
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477625
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_44</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476965
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477085
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28480667
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_35</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476965
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477085
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477187
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_58</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477581
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28486667
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_63</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476921
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477069
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477361
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477763
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28480937
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_34</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476967
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477207
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_0</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476965
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477167
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28512067
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_86</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476921
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477069
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477361
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477763
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28487137
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_27</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476941
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477211
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28480043
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_76</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477011
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478723
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_83</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476967
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477079
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479699
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_66</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476905
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28494677
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_57</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476965
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477085
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28485287
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_60</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476921
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477069
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477361
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28483113
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_33</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476965
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477167
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478707
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_56</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476941
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477211
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477919
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_19</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476921
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477069
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477475
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_47</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476941
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477185
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_5</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476965
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479977
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_50</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476905
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479055
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_24</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476965
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477531
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_84</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476965
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477167
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477571
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_75</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476965
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477167
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479041
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_14</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476921
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477069
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477361
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478709
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_74</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476967
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477169
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_48</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476965
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477167
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28483819
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_39</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476989
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28480495
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_21</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476921
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477069
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477361
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477977
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28485891
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_81</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476965
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477085
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477743
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_55</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476967
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477077
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_38</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477189
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28480007
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_69</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477011
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478529
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_4</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476921
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477069
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477361
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478901
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_06_26_028218_45</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477011
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477425
</commentlist>
</thread>
<conversation>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#conversation09_06_26_028218.8</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476905
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28480557
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479055
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478273
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479893
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28483853
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28480449
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477081
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28494677
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478981
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28480465
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28480675
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477785
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28485751
</commentlist>
</conversation>
<conversation>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#conversation09_06_26_028218.13</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477107
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477441
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477457
</commentlist>
</conversation>
<conversation>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#conversation09_06_26_028218.16</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477063
</commentlist>
</conversation>
<conversation>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#conversation09_06_26_028218.14</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477011
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478529
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477425
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477489
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478723
</commentlist>
</conversation>
<conversation>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#conversation09_06_26_028218.5</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476921
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477069
--http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477475
--http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28481355
--http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477361
---http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478305
---http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477851
---http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478901
---http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478709
---http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28483113
---http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478839
---http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477763
----http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479401
----http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479875
----http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28480937
----http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28487137
---http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477977
----http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28480303
----http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28485891
----http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478473
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477335
--http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477751
--http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477523
</commentlist>
</conversation>
<conversation>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#conversation09_06_26_028218.9</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478019
</commentlist>
</conversation>
<conversation>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#conversation09_06_26_028218.2</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477383
</commentlist>
</conversation>
<conversation>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#conversation09_06_26_028218.17</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476967
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477079
--http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479699
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477625
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477077
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477207
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477169
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28482179
</commentlist>
</conversation>
<conversation>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#conversation09_06_26_028218.3</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476941
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477467
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477185
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478669
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477535
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477211
--http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28480349
--http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477919
--http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28480043
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477411
--http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28489715
--http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28481227
--http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477873
--http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28482555
--http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479135
--http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28484189
--http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479461
</commentlist>
</conversation>
<conversation>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#conversation09_06_26_028218.0</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476943
</commentlist>
</conversation>
<conversation>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#conversation09_06_26_028218.11</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476989
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477675
--http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479535
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28480495
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28482821
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478265
</commentlist>
</conversation>
<conversation>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#conversation09_06_26_028218.12</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478671
</commentlist>
</conversation>
<conversation>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#conversation09_06_26_028218.1</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28476965
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479977
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477999
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477531
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477337
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477831
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477819
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477167
--http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28512067
--http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479041
--http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479537
--http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477571
--http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478707
--http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28483819
--http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28487125
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477965
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477085
--http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28480457
--http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477187
--http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478095
--http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28485287
--http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477845
--http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28478127
--http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477743
--http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479831
--http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477373
--http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28480667
</commentlist>
</conversation>
<conversation>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#conversation09_06_26_028218.10</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477451
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28488783
</commentlist>
</conversation>
<conversation>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#conversation09_06_26_028218.6</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477115
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477967
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28479953
</commentlist>
</conversation>
<conversation>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#conversation09_06_26_028218.4</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477581
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28486667
</commentlist>
</conversation>
<conversation>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#conversation09_06_26_028218.7</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477189
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28480007
</commentlist>
</conversation>
<conversation>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#conversation09_06_26_028218.15</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_06_26_028218.28477305
</commentlist>
</conversation>
