<article>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#article09_11_23_090213</id>
	<title>Australia's CSIRO To Launch CPU-GPU Supercomputer</title>
	<author>timothy</author>
	<datestamp>1258969200000</datestamp>
	<htmltext>bennyboy64 contributes this excerpt from CRN Australia:  <i>"<a href="http://en.wikipedia.org/wiki/CSIRO">The CSIRO</a> will this week <a href="http://www.crn.com.au/News/161005,csiro-to-launch-gpu-based-supercomputer.aspx">launch  a new supercomputer which uses a cluster of GPUs</a> [pictures] to gain a processing capacity that competes with supercomputers over twice its size.
  The supercomputer is one of the world's first to combine traditional CPUs with the more powerful GPUs.
  It features 100 Intel Xeon CPU chips and 50 Tesla GPU chips, connected to an 80 Terabyte Hitachi Data Systems network attached storage unit. CSIRO science applications <a href="http://www.csiro.au/resources/GPU-cluster.html">have already seen 10-100x speedups on NVIDIA GPUs</a>."</i></htmltext>
<tokenext>bennyboy64 contributes this excerpt from CRN Australia : " The CSIRO will this week launch a new supercomputer which uses a cluster of GPUs [ pictures ] to gain a processing capacity that competes with supercomputers over twice its size .
The supercomputer is one of the world 's first to combine traditional CPUs with the more powerful GPUs .
It features 100 Intel Xeon CPU chips and 50 Tesla GPU chips , connected to an 80 Terabyte Hitachi Data Systems network attached storage unit .
CSIRO science applications have already seen 10-100x speedups on NVIDIA GPUs .
"</tokentext>
<sentencetext>bennyboy64 contributes this excerpt from CRN Australia:  "The CSIRO will this week launch  a new supercomputer which uses a cluster of GPUs [pictures] to gain a processing capacity that competes with supercomputers over twice its size.
The supercomputer is one of the world's first to combine traditional CPUs with the more powerful GPUs.
It features 100 Intel Xeon CPU chips and 50 Tesla GPU chips, connected to an 80 Terabyte Hitachi Data Systems network attached storage unit.
CSIRO science applications have already seen 10-100x speedups on NVIDIA GPUs.
"</sentencetext>
</article>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200436</id>
	<title>Oh look....</title>
	<author>Ozlanthos</author>
	<datestamp>1258975320000</datestamp>
	<modclass>Redundant</modclass>
	<modscore>0</modscore>
	<htmltext>My next year's desktop specs!
<p>
-Oz</p></htmltext>
<tokenext>My next year 's desktop specs !
-Oz</tokentext>
<sentencetext>My next year's desktop specs!
-Oz</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200740</id>
	<title>Re:lollero</title>
	<author>XDirtypunkX</author>
	<datestamp>1258981260000</datestamp>
	<modclass>Informativ</modclass>
	<modscore>3</modscore>
	<htmltext><p>It's only traditional on very particular workloads that are very parallel, use a lot of floating point and has a largely coherent execution pattern/memory access. The CPU is still the king of general computing tasks that have lots of incoherent branches, indirection and that require serialized execution.</p></htmltext>
<tokenext>It 's only traditional on very particular workloads that are very parallel , use a lot of floating point and has a largely coherent execution pattern/memory access .
The CPU is still the king of general computing tasks that have lots of incoherent branches , indirection and that require serialized execution .</tokentext>
<sentencetext>It's only traditional on very particular workloads that are very parallel, use a lot of floating point and has a largely coherent execution pattern/memory access.
The CPU is still the king of general computing tasks that have lots of incoherent branches, indirection and that require serialized execution.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200316</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200612</id>
	<title>FINALY!</title>
	<author>Anonymous</author>
	<datestamp>1258979340000</datestamp>
	<modclass>Funny</modclass>
	<modscore>2</modscore>
	<htmltext>Finaly a machine good enought to run Crysis at full specs on 1680x1050 (well, I hope so)</htmltext>
<tokenext>Finaly a machine good enought to run Crysis at full specs on 1680x1050 ( well , I hope so )</tokentext>
<sentencetext>Finaly a machine good enought to run Crysis at full specs on 1680x1050 (well, I hope so)</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200342</id>
	<title>Cool but...</title>
	<author>Anonymous</author>
	<datestamp>1258973640000</datestamp>
	<modclass>Funny</modclass>
	<modscore>1</modscore>
	<htmltext><p>..can it Run CRySiS?</p></htmltext>
<tokenext>..can it Run CRySiS ?</tokentext>
<sentencetext>..can it Run CRySiS?</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200478</id>
	<title>The World of Tomorrow</title>
	<author>muphin</author>
	<datestamp>1258976040000</datestamp>
	<modclass>Funny</modclass>
	<modscore>1</modscore>
	<htmltext>wow the world of technology is spiking, i remember only a few years ago there was only 1 massive super computer,
now every university will have one, what next, link every supercomputer and have a supercomputer cloud or should i say nebula now?<nobr> <wbr></nobr>:p

the rise of the machine, let me take this time to welcome our new ovelords.</htmltext>
<tokenext>wow the world of technology is spiking , i remember only a few years ago there was only 1 massive super computer , now every university will have one , what next , link every supercomputer and have a supercomputer cloud or should i say nebula now ?
: p the rise of the machine , let me take this time to welcome our new ovelords .</tokentext>
<sentencetext>wow the world of technology is spiking, i remember only a few years ago there was only 1 massive super computer,
now every university will have one, what next, link every supercomputer and have a supercomputer cloud or should i say nebula now?
:p

the rise of the machine, let me take this time to welcome our new ovelords.</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200846</id>
	<title>what to do when the 'tires' are all flat</title>
	<author>Anonymous</author>
	<datestamp>1258983240000</datestamp>
	<modclass>None</modclass>
	<modscore>0</modscore>
	<htmltext><p>give it more gas, that's what we do? when we run out of/blow up from, gas??? ta da?</p><p>some of us are already learning to walk again. it feels pretty good.</p></htmltext>
<tokenext>give it more gas , that 's what we do ?
when we run out of/blow up from , gas ? ? ?
ta da ? some of us are already learning to walk again .
it feels pretty good .</tokentext>
<sentencetext>give it more gas, that's what we do?
when we run out of/blow up from, gas???
ta da?some of us are already learning to walk again.
it feels pretty good.</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200444</id>
	<title>Re:lollero</title>
	<author>JorDan Clock</author>
	<datestamp>1258975500000</datestamp>
	<modclass>Informativ</modclass>
	<modscore>4</modscore>
	<htmltext>You already have the technology in your box. The difference is, in the past couple of years, GPUs that were once used exclusively to speed up rendering have become more and more generalized on the hardware and instruction set level to the point where they are a very attractive method of speeding up things other than rendering. Physics simulations, such as fluid dynamics, are much faster on a GPU than a CPU. I currently run the GPU client of Folding@Home and it outperforms the CPU client by orders of magnitude.<br> <br>The hardware has been around for quite some time, but now we're realizing all the things a GPU can do besides run pretty games faster.</htmltext>
<tokenext>You already have the technology in your box .
The difference is , in the past couple of years , GPUs that were once used exclusively to speed up rendering have become more and more generalized on the hardware and instruction set level to the point where they are a very attractive method of speeding up things other than rendering .
Physics simulations , such as fluid dynamics , are much faster on a GPU than a CPU .
I currently run the GPU client of Folding @ Home and it outperforms the CPU client by orders of magnitude .
The hardware has been around for quite some time , but now we 're realizing all the things a GPU can do besides run pretty games faster .</tokentext>
<sentencetext>You already have the technology in your box.
The difference is, in the past couple of years, GPUs that were once used exclusively to speed up rendering have become more and more generalized on the hardware and instruction set level to the point where they are a very attractive method of speeding up things other than rendering.
Physics simulations, such as fluid dynamics, are much faster on a GPU than a CPU.
I currently run the GPU client of Folding@Home and it outperforms the CPU client by orders of magnitude.
The hardware has been around for quite some time, but now we're realizing all the things a GPU can do besides run pretty games faster.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200316</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200434</id>
	<title>Seems logical to me.</title>
	<author>JorDan Clock</author>
	<datestamp>1258975200000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext>A super computing cluster is already used for highly parallelized problems. Using hardware that handles those kinds of problems at a far greater speed than a typical CPU is a no-brainer. I think the part of the story that would be real interesting to the<nobr> <wbr></nobr>/. crowd is what exactly are the kinds of problems they're using this cluster to speed up. GPUs aren't too keen on problems involving data that is hard to cache and as far as I know, the instruction set is somewhat limited to doing lots of little, parallel calculations, but have a hard time with large, solid problems.<br> <br>I am very interested in seeing what kinds of research this will help the most with and what areas will still be more efficient to run on Xeons/Opterons.</htmltext>
<tokenext>A super computing cluster is already used for highly parallelized problems .
Using hardware that handles those kinds of problems at a far greater speed than a typical CPU is a no-brainer .
I think the part of the story that would be real interesting to the / .
crowd is what exactly are the kinds of problems they 're using this cluster to speed up .
GPUs are n't too keen on problems involving data that is hard to cache and as far as I know , the instruction set is somewhat limited to doing lots of little , parallel calculations , but have a hard time with large , solid problems .
I am very interested in seeing what kinds of research this will help the most with and what areas will still be more efficient to run on Xeons/Opterons .</tokentext>
<sentencetext>A super computing cluster is already used for highly parallelized problems.
Using hardware that handles those kinds of problems at a far greater speed than a typical CPU is a no-brainer.
I think the part of the story that would be real interesting to the /.
crowd is what exactly are the kinds of problems they're using this cluster to speed up.
GPUs aren't too keen on problems involving data that is hard to cache and as far as I know, the instruction set is somewhat limited to doing lots of little, parallel calculations, but have a hard time with large, solid problems.
I am very interested in seeing what kinds of research this will help the most with and what areas will still be more efficient to run on Xeons/Opterons.</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30201098</id>
	<title>Imagine...</title>
	<author>cyborch</author>
	<datestamp>1258986060000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext>... a beowulf cluster of those!<nobr> <wbr></nobr>;)<br><br>(Sorry, it had to be said)</htmltext>
<tokenext>... a beowulf cluster of those !
; ) ( Sorry , it had to be said )</tokentext>
<sentencetext>... a beowulf cluster of those!
;)(Sorry, it had to be said)</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200330</id>
	<title>Can someone explain...</title>
	<author>Anonymous</author>
	<datestamp>1258973460000</datestamp>
	<modclass>Interestin</modclass>
	<modscore>1</modscore>
	<htmltext><p>Can someone explain exactly what the benefits/drawbacks of using GPUs for processing?</p><p>It would also be nice if someone could give a quick run down of what sort of applications GPUs are good at.</p></htmltext>
<tokenext>Can someone explain exactly what the benefits/drawbacks of using GPUs for processing ? It would also be nice if someone could give a quick run down of what sort of applications GPUs are good at .</tokentext>
<sentencetext>Can someone explain exactly what the benefits/drawbacks of using GPUs for processing?It would also be nice if someone could give a quick run down of what sort of applications GPUs are good at.</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200796</id>
	<title>not first, just big</title>
	<author>mattdm</author>
	<datestamp>1258982400000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext><a href="https://ircs.seas.harvard.edu/display/USERDOCS/SEAS+GPGPU+CUDA+cluster+'resonance'" title="harvard.edu">We</a> [harvard.edu] have one of those already; I imagine a lot of schools do. Ours is only an 18-node cluster so the numbers are much smaller, but the story here is that this is relatively big, not that it's some new thing.</htmltext>
<tokenext>We [ harvard.edu ] have one of those already ; I imagine a lot of schools do .
Ours is only an 18-node cluster so the numbers are much smaller , but the story here is that this is relatively big , not that it 's some new thing .</tokentext>
<sentencetext>We [harvard.edu] have one of those already; I imagine a lot of schools do.
Ours is only an 18-node cluster so the numbers are much smaller, but the story here is that this is relatively big, not that it's some new thing.</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30202252</id>
	<title>The #5 Supercomputer is already GPU based</title>
	<author>thatguymike</author>
	<datestamp>1258992780000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext><a href="http://www.top500.org/system/10186" title="top500.org" rel="nofollow">http://www.top500.org/system/10186</a> [top500.org]

The machine quoted in TFA is quoting single precision.  Currently the ATI boards trounce the Nvidia boards in double precision.  The next GPU cluster down the list is Nvidia based at #56 <a href="http://www.top500.org/site/690" title="top500.org" rel="nofollow">http://www.top500.org/site/690</a> [top500.org]</htmltext>
<tokenext>http : //www.top500.org/system/10186 [ top500.org ] The machine quoted in TFA is quoting single precision .
Currently the ATI boards trounce the Nvidia boards in double precision .
The next GPU cluster down the list is Nvidia based at # 56 http : //www.top500.org/site/690 [ top500.org ]</tokentext>
<sentencetext>http://www.top500.org/system/10186 [top500.org]

The machine quoted in TFA is quoting single precision.
Currently the ATI boards trounce the Nvidia boards in double precision.
The next GPU cluster down the list is Nvidia based at #56 http://www.top500.org/site/690 [top500.org]</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200376</id>
	<title>Re:Can someone explain...</title>
	<author>Anonymous</author>
	<datestamp>1258974300000</datestamp>
	<modclass>None</modclass>
	<modscore>0</modscore>
	<htmltext><p>Heavy mathematics - anything that is readily parallelised. GPUs are basically massive vector arithmetic units, capable of crunching through lots and lots of floating point calculations at very high speed. As usual, <a href="http://en.wikipedia.org/wiki/GPGPU" title="wikipedia.org" rel="nofollow">Wikipedia</a> [wikipedia.org] has an informative article.</p></htmltext>
<tokenext>Heavy mathematics - anything that is readily parallelised .
GPUs are basically massive vector arithmetic units , capable of crunching through lots and lots of floating point calculations at very high speed .
As usual , Wikipedia [ wikipedia.org ] has an informative article .</tokentext>
<sentencetext>Heavy mathematics - anything that is readily parallelised.
GPUs are basically massive vector arithmetic units, capable of crunching through lots and lots of floating point calculations at very high speed.
As usual, Wikipedia [wikipedia.org] has an informative article.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200330</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200498</id>
	<title>GPUs are good if</title>
	<author>Sycraft-fu</author>
	<datestamp>1258976580000</datestamp>
	<modclass>Informativ</modclass>
	<modscore>4</modscore>
	<htmltext><p>1) Your problem is one that is more or less infinitely parallel in nature. Their method of operation is a whole bunch of parallel pathways, as such your problem needs to be one that can be broken down in to very small parts that can execute in parallel. A single GPU these days can have hundreds of parallel shaders (the GTX 285 has 240 for example).</p><p>2) Your problem needs to be fairly linear, not a whole lot of branching. Modern GPUs can handle branching, but they take a heavy penalty doing it. They are designed for processing data streams where you just crunch numbers, not a lot of if-then kind of logic. So if your problem should be fairly linear to run well.</p><p>3) Your problem needs to be solvable using single precision floating point math. This is changing, new GPUs are getting double precision capability and better integer handling, but almost all of the ones on the market now are only fast with 32-bit FP. So your problem needs to use that kind of math.</p><p>4) Your problem needs to be able to be broken down in to pieces that can fit in the memory on a GPU board. This varies, it is typically 512MB-1GB for consumer boards and as much as 4GB for Teslas.  Regardless, your problem needs to fit in there for the most part. The memory on a GPU is very fast, 100GB/sec or more of bandwidth for high end ones. The communication back to the system via PCIe is an order of magnitude slower usually. So while you certainly can move data to main memory and to disk, it needs to be done sparingly. For the most part, you need to be cranking on stuff that is in the GPU's memory.</p><p>Now, the more your problem meets those criteria, the better a candidate it is for acceleration by GPUs. If your problem is fairly small, very parallel, very linear and all single precision, well you will see absolutely massive gains over a CPU. It can be 100x or so. These are indeed the kind of gains you see in computer graphics, which is not surprising given that's what GPUs are made for. If your problem is very single threaded, has tons of branching, requires hundreds of gigs of data and such, well then you might find offloading to a GPU slower than trying it on a CPU. The system might spend more time just getting the data moved around than doing any real work.</p><p>The good news is, there's an awful lot of problems that nicely meet the criteria for running on GPUs. They may not be perfectly ideal, but they still run plenty fast. After all, if a GPU is ideally 100x a CPU, and your code can only use it to 10\% efficiency, well hell you are still doing 10x what you did on a CPU.</p><p>So what kind of things are like this? Well graphics would be the most obvious one. That's where the design comes from. You do math on lots of matrices of 32-bit numbers. This doesn't just apply to consumer game graphics though, material shaders in professional 3D programs work the same way. Indeed, you'll find those can be accelerated with GPUs. Audio is another area that is a real good candidate. Most audio processing is the same kind of thing. You have large streams of numbers representing amplitude samples. You need to do various simple math functions on them to add reverb or compress the dynamics or whatever. I don't know of any audio processing that uses GPUs, but they'd do well for it. Protein folding is another great candidate. Folding@Home runs WAY faster on GPUs than CPUs.</p><p>At this point, GPGPU stuff is still really in its infancy. We should start to see more and more of it as more people these days have GPUs that are useful for GPGPU apps (pretty much DX10 or better hardware, nVidia 8000 or higher and ATi 3000 or higher). Also there is starting to be better APIs out for it. nVidia's CUDA is popular, but proprietary to their cards. MS has introduced GPGPU support in DirectX, and OpenCL has come out and is being supported. As such, you should see more apps slowly start to be developed.</p><p>GPUs certainly aren't good at everything, I mean if they were, well then we'd just make CPUs like GPUs and call it good. However there is a large set of problems they are better than the CPU at solving.</p></htmltext>
<tokenext>1 ) Your problem is one that is more or less infinitely parallel in nature .
Their method of operation is a whole bunch of parallel pathways , as such your problem needs to be one that can be broken down in to very small parts that can execute in parallel .
A single GPU these days can have hundreds of parallel shaders ( the GTX 285 has 240 for example ) .2 ) Your problem needs to be fairly linear , not a whole lot of branching .
Modern GPUs can handle branching , but they take a heavy penalty doing it .
They are designed for processing data streams where you just crunch numbers , not a lot of if-then kind of logic .
So if your problem should be fairly linear to run well.3 ) Your problem needs to be solvable using single precision floating point math .
This is changing , new GPUs are getting double precision capability and better integer handling , but almost all of the ones on the market now are only fast with 32-bit FP .
So your problem needs to use that kind of math.4 ) Your problem needs to be able to be broken down in to pieces that can fit in the memory on a GPU board .
This varies , it is typically 512MB-1GB for consumer boards and as much as 4GB for Teslas .
Regardless , your problem needs to fit in there for the most part .
The memory on a GPU is very fast , 100GB/sec or more of bandwidth for high end ones .
The communication back to the system via PCIe is an order of magnitude slower usually .
So while you certainly can move data to main memory and to disk , it needs to be done sparingly .
For the most part , you need to be cranking on stuff that is in the GPU 's memory.Now , the more your problem meets those criteria , the better a candidate it is for acceleration by GPUs .
If your problem is fairly small , very parallel , very linear and all single precision , well you will see absolutely massive gains over a CPU .
It can be 100x or so .
These are indeed the kind of gains you see in computer graphics , which is not surprising given that 's what GPUs are made for .
If your problem is very single threaded , has tons of branching , requires hundreds of gigs of data and such , well then you might find offloading to a GPU slower than trying it on a CPU .
The system might spend more time just getting the data moved around than doing any real work.The good news is , there 's an awful lot of problems that nicely meet the criteria for running on GPUs .
They may not be perfectly ideal , but they still run plenty fast .
After all , if a GPU is ideally 100x a CPU , and your code can only use it to 10 \ % efficiency , well hell you are still doing 10x what you did on a CPU.So what kind of things are like this ?
Well graphics would be the most obvious one .
That 's where the design comes from .
You do math on lots of matrices of 32-bit numbers .
This does n't just apply to consumer game graphics though , material shaders in professional 3D programs work the same way .
Indeed , you 'll find those can be accelerated with GPUs .
Audio is another area that is a real good candidate .
Most audio processing is the same kind of thing .
You have large streams of numbers representing amplitude samples .
You need to do various simple math functions on them to add reverb or compress the dynamics or whatever .
I do n't know of any audio processing that uses GPUs , but they 'd do well for it .
Protein folding is another great candidate .
Folding @ Home runs WAY faster on GPUs than CPUs.At this point , GPGPU stuff is still really in its infancy .
We should start to see more and more of it as more people these days have GPUs that are useful for GPGPU apps ( pretty much DX10 or better hardware , nVidia 8000 or higher and ATi 3000 or higher ) .
Also there is starting to be better APIs out for it .
nVidia 's CUDA is popular , but proprietary to their cards .
MS has introduced GPGPU support in DirectX , and OpenCL has come out and is being supported .
As such , you should see more apps slowly start to be developed.GPUs certainly are n't good at everything , I mean if they were , well then we 'd just make CPUs like GPUs and call it good .
However there is a large set of problems they are better than the CPU at solving .</tokentext>
<sentencetext>1) Your problem is one that is more or less infinitely parallel in nature.
Their method of operation is a whole bunch of parallel pathways, as such your problem needs to be one that can be broken down in to very small parts that can execute in parallel.
A single GPU these days can have hundreds of parallel shaders (the GTX 285 has 240 for example).2) Your problem needs to be fairly linear, not a whole lot of branching.
Modern GPUs can handle branching, but they take a heavy penalty doing it.
They are designed for processing data streams where you just crunch numbers, not a lot of if-then kind of logic.
So if your problem should be fairly linear to run well.3) Your problem needs to be solvable using single precision floating point math.
This is changing, new GPUs are getting double precision capability and better integer handling, but almost all of the ones on the market now are only fast with 32-bit FP.
So your problem needs to use that kind of math.4) Your problem needs to be able to be broken down in to pieces that can fit in the memory on a GPU board.
This varies, it is typically 512MB-1GB for consumer boards and as much as 4GB for Teslas.
Regardless, your problem needs to fit in there for the most part.
The memory on a GPU is very fast, 100GB/sec or more of bandwidth for high end ones.
The communication back to the system via PCIe is an order of magnitude slower usually.
So while you certainly can move data to main memory and to disk, it needs to be done sparingly.
For the most part, you need to be cranking on stuff that is in the GPU's memory.Now, the more your problem meets those criteria, the better a candidate it is for acceleration by GPUs.
If your problem is fairly small, very parallel, very linear and all single precision, well you will see absolutely massive gains over a CPU.
It can be 100x or so.
These are indeed the kind of gains you see in computer graphics, which is not surprising given that's what GPUs are made for.
If your problem is very single threaded, has tons of branching, requires hundreds of gigs of data and such, well then you might find offloading to a GPU slower than trying it on a CPU.
The system might spend more time just getting the data moved around than doing any real work.The good news is, there's an awful lot of problems that nicely meet the criteria for running on GPUs.
They may not be perfectly ideal, but they still run plenty fast.
After all, if a GPU is ideally 100x a CPU, and your code can only use it to 10\% efficiency, well hell you are still doing 10x what you did on a CPU.So what kind of things are like this?
Well graphics would be the most obvious one.
That's where the design comes from.
You do math on lots of matrices of 32-bit numbers.
This doesn't just apply to consumer game graphics though, material shaders in professional 3D programs work the same way.
Indeed, you'll find those can be accelerated with GPUs.
Audio is another area that is a real good candidate.
Most audio processing is the same kind of thing.
You have large streams of numbers representing amplitude samples.
You need to do various simple math functions on them to add reverb or compress the dynamics or whatever.
I don't know of any audio processing that uses GPUs, but they'd do well for it.
Protein folding is another great candidate.
Folding@Home runs WAY faster on GPUs than CPUs.At this point, GPGPU stuff is still really in its infancy.
We should start to see more and more of it as more people these days have GPUs that are useful for GPGPU apps (pretty much DX10 or better hardware, nVidia 8000 or higher and ATi 3000 or higher).
Also there is starting to be better APIs out for it.
nVidia's CUDA is popular, but proprietary to their cards.
MS has introduced GPGPU support in DirectX, and OpenCL has come out and is being supported.
As such, you should see more apps slowly start to be developed.GPUs certainly aren't good at everything, I mean if they were, well then we'd just make CPUs like GPUs and call it good.
However there is a large set of problems they are better than the CPU at solving.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200330</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200426</id>
	<title>YOU FAIL IT</title>
	<author>Anonymous</author>
	<datestamp>1258975140000</datestamp>
	<modclass>Offtopic</modclass>
	<modscore>-1</modscore>
	<htmltext><A HREF="http://goat.cx/" title="goat.cx" rel="nofollow">Every chance I got states that there has brought upon Discussion I'm where it was when downward sp1ral. In centralized models Series of internal</a> [goat.cx]</htmltext>
<tokenext>Every chance I got states that there has brought upon Discussion I 'm where it was when downward sp1ral .
In centralized models Series of internal [ goat.cx ]</tokentext>
<sentencetext>Every chance I got states that there has brought upon Discussion I'm where it was when downward sp1ral.
In centralized models Series of internal [goat.cx]</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200316</id>
	<title>lollero</title>
	<author>Anonymous</author>
	<datestamp>1258973280000</datestamp>
	<modclass>None</modclass>
	<modscore>0</modscore>
	<htmltext><p>Why is it "more powerful" than "traditional" CPUs?</p><p>And why is it not under my hood already if it is superior technology?</p></htmltext>
<tokenext>Why is it " more powerful " than " traditional " CPUs ? And why is it not under my hood already if it is superior technology ?</tokentext>
<sentencetext>Why is it "more powerful" than "traditional" CPUs?And why is it not under my hood already if it is superior technology?</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200552</id>
	<title>In related news...</title>
	<author>sonamchauhan</author>
	<datestamp>1258977780000</datestamp>
	<modclass>Interestin</modclass>
	<modscore>2</modscore>
	<htmltext><p>Hmmm.... is this setup a realisation of this release from Nvidia in March</p><p>Nvidia Touts New GPU Supercomputer<br><a href="http://gigaom.com/2009/05/04/nvidia-touts-new-gpu-supercomputer/" title="gigaom.com">http://gigaom.com/2009/05/04/nvidia-touts-new-gpu-supercomputer/</a> [gigaom.com]</p><p>Another 'standalone' GPGPU supercomputer, without the Infiniband switch<br>University of Antwerp makes 4000EUR NVIDIA supercomputer<br><a href="http://www.dvhardware.net/article27538.html" title="dvhardware.net">http://www.dvhardware.net/article27538.html</a> [dvhardware.net]</p></htmltext>
<tokenext>Hmmm.... is this setup a realisation of this release from Nvidia in MarchNvidia Touts New GPU Supercomputerhttp : //gigaom.com/2009/05/04/nvidia-touts-new-gpu-supercomputer/ [ gigaom.com ] Another 'standalone ' GPGPU supercomputer , without the Infiniband switchUniversity of Antwerp makes 4000EUR NVIDIA supercomputerhttp : //www.dvhardware.net/article27538.html [ dvhardware.net ]</tokentext>
<sentencetext>Hmmm.... is this setup a realisation of this release from Nvidia in MarchNvidia Touts New GPU Supercomputerhttp://gigaom.com/2009/05/04/nvidia-touts-new-gpu-supercomputer/ [gigaom.com]Another 'standalone' GPGPU supercomputer, without the Infiniband switchUniversity of Antwerp makes 4000EUR NVIDIA supercomputerhttp://www.dvhardware.net/article27538.html [dvhardware.net]</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200806</id>
	<title>mistake for open source</title>
	<author>GNUPublicLicense</author>
	<datestamp>1258982640000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext>From an open source point of view... this is a mistake since we (as open source people) must favor AMD GPUs. Moreover, it has been 2 years the AMD GPUs seem faster than nvidia ones.
So from such bad news, open source people must keep the bearing: favor AMD GPUs whatever.</htmltext>
<tokenext>From an open source point of view... this is a mistake since we ( as open source people ) must favor AMD GPUs .
Moreover , it has been 2 years the AMD GPUs seem faster than nvidia ones .
So from such bad news , open source people must keep the bearing : favor AMD GPUs whatever .</tokentext>
<sentencetext>From an open source point of view... this is a mistake since we (as open source people) must favor AMD GPUs.
Moreover, it has been 2 years the AMD GPUs seem faster than nvidia ones.
So from such bad news, open source people must keep the bearing: favor AMD GPUs whatever.</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200640</id>
	<title>NVIDIA, huh?</title>
	<author>Anonymous</author>
	<datestamp>1258979820000</datestamp>
	<modclass>Funny</modclass>
	<modscore>1</modscore>
	<htmltext><p>Does it use wood screws?</p></htmltext>
<tokenext>Does it use wood screws ?</tokentext>
<sentencetext>Does it use wood screws?</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200420</id>
	<title>Re:Can someone explain...</title>
	<author>Anonymous</author>
	<datestamp>1258975080000</datestamp>
	<modclass>Interestin</modclass>
	<modscore>3</modscore>
	<htmltext><p>I can take a stab;  GPUs traditionally render graphics, good at processing vectors and mathsy things.  Now think of a simulation of a bunch of atoms, the forces between the atoms are often approximated to Newtonian laws of motion for computational efficiency reasons, this is especially important when dealing with tens of thousands of atoms - called Molecular Dynamics (MD).  So the same maths used for graphic intensive computer games is the same as classical MD.  The problem hither to is that MD software has never really been compiled for GPU architecture, just Athlons and Pentiums.</p><p>I should mention that I use the CSIRO CPU cluster, it's quite good already, but I'm still waiting weeks to simulate a microsecond of 10,000 atoms using 32 processors.  My new side project will be trying it out on the GPUs.  100x faster they reckon, that'll be a game changer for me</p></htmltext>
<tokenext>I can take a stab ; GPUs traditionally render graphics , good at processing vectors and mathsy things .
Now think of a simulation of a bunch of atoms , the forces between the atoms are often approximated to Newtonian laws of motion for computational efficiency reasons , this is especially important when dealing with tens of thousands of atoms - called Molecular Dynamics ( MD ) .
So the same maths used for graphic intensive computer games is the same as classical MD .
The problem hither to is that MD software has never really been compiled for GPU architecture , just Athlons and Pentiums.I should mention that I use the CSIRO CPU cluster , it 's quite good already , but I 'm still waiting weeks to simulate a microsecond of 10,000 atoms using 32 processors .
My new side project will be trying it out on the GPUs .
100x faster they reckon , that 'll be a game changer for me</tokentext>
<sentencetext>I can take a stab;  GPUs traditionally render graphics, good at processing vectors and mathsy things.
Now think of a simulation of a bunch of atoms, the forces between the atoms are often approximated to Newtonian laws of motion for computational efficiency reasons, this is especially important when dealing with tens of thousands of atoms - called Molecular Dynamics (MD).
So the same maths used for graphic intensive computer games is the same as classical MD.
The problem hither to is that MD software has never really been compiled for GPU architecture, just Athlons and Pentiums.I should mention that I use the CSIRO CPU cluster, it's quite good already, but I'm still waiting weeks to simulate a microsecond of 10,000 atoms using 32 processors.
My new side project will be trying it out on the GPUs.
100x faster they reckon, that'll be a game changer for me</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200330</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30202722</id>
	<title>Re:not first, just big</title>
	<author>dlapine</author>
	<datestamp>1258995540000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext><p> We've had the <a href="http://www.ncsa.illinois.edu/News/08/0908NCSAto.html" title="illinois.edu">Lincoln cluster</a> [illinois.edu] online and offering processing time since February of 2009. 196 computing nodes (dual quad cores) and 96 Tesla units. That being said, congrats to the Aussie's for bringing a powerful new system online.</p><p>Someone later in thread asked if these GPU units would actually be useful for scientific computing. We think so. Our users and researchers here have developed implementations of both <a href="http://www.ks.uiuc.edu/Research/namd/" title="uiuc.edu">NAMD, a parallel molecular dynamics simulator</a> [uiuc.edu] and <a href="http://www.physics.indiana.edu/~sg/milc.html" title="indiana.edu">MIMD Lattice Computation (MILC) Collaboration</a> [indiana.edu] that use the power of the GPU's. Both of these codes are freely available and widely used in the HPC community. We've had no lack of requests for time on the Lincoln cluster.</p><p>Are these GPUS for everyone? Nope. To disappoint all you gamers out there, the Tesla units have <b>no</b> graphics out ports. All the communication is done over the the PCIe bus. But for all of you budding scientists out there, these cards use the same freely available CUDA language that runs on all modern (8xxx and above) Nvidia hardware, so you may already have compatible GPU in your desktop now, even if it's just a single unit and slower. </p><p>One last note, while these units run really fast with single precision, they are capable of running in double precision, albeit much slower. For some problems, multiple initial runs can be done at the lower precision to localize the solution set, before doing a slower high precision run to find the final solution. This is similar to what Hollywood does when rendering animated movies- they first render a quick lo res version to see if the timing and characters are correct, then they run a hi-res version which takes longer to get a finished product. (Yes, I know, there's a lot more steps to it, but hey, this is just an analogy)</p></htmltext>
<tokenext>We 've had the Lincoln cluster [ illinois.edu ] online and offering processing time since February of 2009 .
196 computing nodes ( dual quad cores ) and 96 Tesla units .
That being said , congrats to the Aussie 's for bringing a powerful new system online.Someone later in thread asked if these GPU units would actually be useful for scientific computing .
We think so .
Our users and researchers here have developed implementations of both NAMD , a parallel molecular dynamics simulator [ uiuc.edu ] and MIMD Lattice Computation ( MILC ) Collaboration [ indiana.edu ] that use the power of the GPU 's .
Both of these codes are freely available and widely used in the HPC community .
We 've had no lack of requests for time on the Lincoln cluster.Are these GPUS for everyone ?
Nope. To disappoint all you gamers out there , the Tesla units have no graphics out ports .
All the communication is done over the the PCIe bus .
But for all of you budding scientists out there , these cards use the same freely available CUDA language that runs on all modern ( 8xxx and above ) Nvidia hardware , so you may already have compatible GPU in your desktop now , even if it 's just a single unit and slower .
One last note , while these units run really fast with single precision , they are capable of running in double precision , albeit much slower .
For some problems , multiple initial runs can be done at the lower precision to localize the solution set , before doing a slower high precision run to find the final solution .
This is similar to what Hollywood does when rendering animated movies- they first render a quick lo res version to see if the timing and characters are correct , then they run a hi-res version which takes longer to get a finished product .
( Yes , I know , there 's a lot more steps to it , but hey , this is just an analogy )</tokentext>
<sentencetext> We've had the Lincoln cluster [illinois.edu] online and offering processing time since February of 2009.
196 computing nodes (dual quad cores) and 96 Tesla units.
That being said, congrats to the Aussie's for bringing a powerful new system online.Someone later in thread asked if these GPU units would actually be useful for scientific computing.
We think so.
Our users and researchers here have developed implementations of both NAMD, a parallel molecular dynamics simulator [uiuc.edu] and MIMD Lattice Computation (MILC) Collaboration [indiana.edu] that use the power of the GPU's.
Both of these codes are freely available and widely used in the HPC community.
We've had no lack of requests for time on the Lincoln cluster.Are these GPUS for everyone?
Nope. To disappoint all you gamers out there, the Tesla units have no graphics out ports.
All the communication is done over the the PCIe bus.
But for all of you budding scientists out there, these cards use the same freely available CUDA language that runs on all modern (8xxx and above) Nvidia hardware, so you may already have compatible GPU in your desktop now, even if it's just a single unit and slower.
One last note, while these units run really fast with single precision, they are capable of running in double precision, albeit much slower.
For some problems, multiple initial runs can be done at the lower precision to localize the solution set, before doing a slower high precision run to find the final solution.
This is similar to what Hollywood does when rendering animated movies- they first render a quick lo res version to see if the timing and characters are correct, then they run a hi-res version which takes longer to get a finished product.
(Yes, I know, there's a lot more steps to it, but hey, this is just an analogy)</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200796</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200338</id>
	<title>twice the size but what cost?</title>
	<author>Anonymous</author>
	<datestamp>1258973580000</datestamp>
	<modclass>Interestin</modclass>
	<modscore>0</modscore>
	<htmltext><p>The article didn't seem to mention cost, power usage, heat, or anything remotely relevant. Just a nice happy fluff piece for NVIDIA who I do adore but really these articles on slashdot do not have as much tech sustenance as it used to.</p></htmltext>
<tokenext>The article did n't seem to mention cost , power usage , heat , or anything remotely relevant .
Just a nice happy fluff piece for NVIDIA who I do adore but really these articles on slashdot do not have as much tech sustenance as it used to .</tokentext>
<sentencetext>The article didn't seem to mention cost, power usage, heat, or anything remotely relevant.
Just a nice happy fluff piece for NVIDIA who I do adore but really these articles on slashdot do not have as much tech sustenance as it used to.</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30202408</id>
	<title>frIs-t stop</title>
	<author>Anonymous</author>
	<datestamp>1258993680000</datestamp>
	<modclass>Offtopic</modclass>
	<modscore>-1</modscore>
	<htmltext>to be a3o0t doing is mired in an</htmltext>
<tokenext>to be a3o0t doing is mired in an</tokentext>
<sentencetext>to be a3o0t doing is mired in an</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30203354</id>
	<title>article is incorrect</title>
	<author>Anonymous</author>
	<datestamp>1258998900000</datestamp>
	<modclass>None</modclass>
	<modscore>0</modscore>
	<htmltext><p>the cluster actually has 50 Tesla C1070 boards, each of which contains 4 GPUs<br>so its 200 GPUs, and that is just the initial rollout with additional nodes to be delivered pretty quickly (perhaps waiting for Fermi)</p></htmltext>
<tokenext>the cluster actually has 50 Tesla C1070 boards , each of which contains 4 GPUsso its 200 GPUs , and that is just the initial rollout with additional nodes to be delivered pretty quickly ( perhaps waiting for Fermi )</tokentext>
<sentencetext>the cluster actually has 50 Tesla C1070 boards, each of which contains 4 GPUsso its 200 GPUs, and that is just the initial rollout with additional nodes to be delivered pretty quickly (perhaps waiting for Fermi)</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200374</id>
	<title>Re:Can someone explain...</title>
	<author>EdZ</author>
	<datestamp>1258974240000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext>Benefits: blindingly fast when running massively parallel computations (think several hundred thousand threads).<br>
Drawbacks: trying to program something to take advantage of all that power requires you to scale up to several thousand threads. Not always that easy.</htmltext>
<tokenext>Benefits : blindingly fast when running massively parallel computations ( think several hundred thousand threads ) .
Drawbacks : trying to program something to take advantage of all that power requires you to scale up to several thousand threads .
Not always that easy .</tokentext>
<sentencetext>Benefits: blindingly fast when running massively parallel computations (think several hundred thousand threads).
Drawbacks: trying to program something to take advantage of all that power requires you to scale up to several thousand threads.
Not always that easy.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200330</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30201422</id>
	<title>Floating point operations</title>
	<author>Anonymous</author>
	<datestamp>1258988400000</datestamp>
	<modclass>None</modclass>
	<modscore>0</modscore>
	<htmltext><p>The biggest benefit to GPU processing is that they are much more adept at floating-point math...that is, 2.5436*23.561234 instead of 1829*2304.  The distributed computing efforts (Folding, Boinc projects) have started writing clients for users' gpus as well, and have seen great success so far.</p><p>Floating point operations are tedious on normal processors, but the shader units on GPU's, designed to handle complex calculations for graphical effects, process non-integers much faster.</p></htmltext>
<tokenext>The biggest benefit to GPU processing is that they are much more adept at floating-point math...that is , 2.5436 * 23.561234 instead of 1829 * 2304 .
The distributed computing efforts ( Folding , Boinc projects ) have started writing clients for users ' gpus as well , and have seen great success so far.Floating point operations are tedious on normal processors , but the shader units on GPU 's , designed to handle complex calculations for graphical effects , process non-integers much faster .</tokentext>
<sentencetext>The biggest benefit to GPU processing is that they are much more adept at floating-point math...that is, 2.5436*23.561234 instead of 1829*2304.
The distributed computing efforts (Folding, Boinc projects) have started writing clients for users' gpus as well, and have seen great success so far.Floating point operations are tedious on normal processors, but the shader units on GPU's, designed to handle complex calculations for graphical effects, process non-integers much faster.</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200882</id>
	<title>Re:GPUs are good if</title>
	<author>Rockoon</author>
	<datestamp>1258984200000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext>You touched on but I think you missed the #1 biggest winner for high end GPU's as it pertains to most GPGPU stuff.. convolution.<br>
<br>
It is not an exaggeration to call these things super-convolvers, excelling at doing large-scale pairwise multiply-and-add's on arrays of floats, which can be leveraged to do more specific things like large matrix multiplication in what amounts to (in practice) sub-linear time. A great many different problem sets can be expressed as a series of convolutions, including neural networks, navier stokes, fourier, signal filtering, cross-correlation, and so on and on..</htmltext>
<tokenext>You touched on but I think you missed the # 1 biggest winner for high end GPU 's as it pertains to most GPGPU stuff.. convolution . It is not an exaggeration to call these things super-convolvers , excelling at doing large-scale pairwise multiply-and-add 's on arrays of floats , which can be leveraged to do more specific things like large matrix multiplication in what amounts to ( in practice ) sub-linear time .
A great many different problem sets can be expressed as a series of convolutions , including neural networks , navier stokes , fourier , signal filtering , cross-correlation , and so on and on. .</tokentext>
<sentencetext>You touched on but I think you missed the #1 biggest winner for high end GPU's as it pertains to most GPGPU stuff.. convolution.

It is not an exaggeration to call these things super-convolvers, excelling at doing large-scale pairwise multiply-and-add's on arrays of floats, which can be leveraged to do more specific things like large matrix multiplication in what amounts to (in practice) sub-linear time.
A great many different problem sets can be expressed as a series of convolutions, including neural networks, navier stokes, fourier, signal filtering, cross-correlation, and so on and on..</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200498</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200382</id>
	<title>Re:Can someone explain...</title>
	<author>Sockatume</author>
	<datestamp>1258974360000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext><p>GPUs are fast but limited to very specific kinds of instructions. If you can write your code using those instructions, it will run much quicker than it would on a general-purpose processor. They're also ahead of the curve on things like parallelisation, compared to desktop chips: the idea of writing graphics code for a 12-pipe GPU was mundane half a decade ago while there's still scant support for multiple cores in CPUs.</p></htmltext>
<tokenext>GPUs are fast but limited to very specific kinds of instructions .
If you can write your code using those instructions , it will run much quicker than it would on a general-purpose processor .
They 're also ahead of the curve on things like parallelisation , compared to desktop chips : the idea of writing graphics code for a 12-pipe GPU was mundane half a decade ago while there 's still scant support for multiple cores in CPUs .</tokentext>
<sentencetext>GPUs are fast but limited to very specific kinds of instructions.
If you can write your code using those instructions, it will run much quicker than it would on a general-purpose processor.
They're also ahead of the curve on things like parallelisation, compared to desktop chips: the idea of writing graphics code for a 12-pipe GPU was mundane half a decade ago while there's still scant support for multiple cores in CPUs.</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200330</parent>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30201028</id>
	<title>CUDA, GPGPU, OpenCL etc.</title>
	<author>Anonymous</author>
	<datestamp>1258985520000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext>What API would be the best approach for writing some future proof GPU code?
<br>
I'm willing to sacrifice some bleeding edge performance now for ease of maintainability.
<br>
<br>
Other GPU possibilities
<br>
    * OpenCL <br>
    * GPGPU<br>
    * CUDA<br>
    * DirectCompute<br>
    * FireStream<br>
    * Larrabee<br>
    * Close to Metal<br>
    * BrookGPU<br>
    * Lib Sh<br>
<br>
Cheers</htmltext>
<tokenext>What API would be the best approach for writing some future proof GPU code ?
I 'm willing to sacrifice some bleeding edge performance now for ease of maintainability .
Other GPU possibilities * OpenCL * GPGPU * CUDA * DirectCompute * FireStream * Larrabee * Close to Metal * BrookGPU * Lib Sh Cheers</tokentext>
<sentencetext>What API would be the best approach for writing some future proof GPU code?
I'm willing to sacrifice some bleeding edge performance now for ease of maintainability.
Other GPU possibilities

    * OpenCL 
    * GPGPU
    * CUDA
    * DirectCompute
    * FireStream
    * Larrabee
    * Close to Metal
    * BrookGPU
    * Lib Sh

Cheers</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30205240</id>
	<title>I seem to recall ...</title>
	<author>PPH</author>
	<datestamp>1259009820000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext>... reading a story some time ago about the use of GPU clusters by organizations on national security watch lists to circumvent ITAR controls.</htmltext>
<tokenext>... reading a story some time ago about the use of GPU clusters by organizations on national security watch lists to circumvent ITAR controls .</tokentext>
<sentencetext>... reading a story some time ago about the use of GPU clusters by organizations on national security watch lists to circumvent ITAR controls.</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200296</id>
	<title>fr1st</title>
	<author>Anonymous</author>
	<datestamp>1258972980000</datestamp>
	<modclass>Offtopic</modclass>
	<modscore>-1</modscore>
	<htmltext>p1zz0t</htmltext>
<tokenext>p1zz0t</tokentext>
<sentencetext>p1zz0t</sentencetext>
</comment>
<comment>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30202766</id>
	<title>Re:FINALY!</title>
	<author>TheKidWho</author>
	<datestamp>1258995780000</datestamp>
	<modclass>None</modclass>
	<modscore>1</modscore>
	<htmltext><p>My current computer already runs Crysis at full specs at 1680x1050 you insensitive clod!</p></htmltext>
<tokenext>My current computer already runs Crysis at full specs at 1680x1050 you insensitive clod !</tokentext>
<sentencetext>My current computer already runs Crysis at full specs at 1680x1050 you insensitive clod!</sentencetext>
	<parent>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200612</parent>
</comment>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_11_23_090213_3</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200420
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200330
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_11_23_090213_0</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200382
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200330
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_11_23_090213_7</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30202722
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200796
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_11_23_090213_6</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30202766
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200612
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_11_23_090213_4</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200376
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200330
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_11_23_090213_8</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200882
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200498
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200330
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_11_23_090213_1</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200740
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200316
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_11_23_090213_5</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200374
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200330
</commentlist>
</thread>
<thread>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#thread_09_11_23_090213_2</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200444
http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200316
</commentlist>
</thread>
<conversation>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#conversation09_11_23_090213.2</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30201028
</commentlist>
</conversation>
<conversation>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#conversation09_11_23_090213.0</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200330
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200382
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200420
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200376
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200374
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200498
--http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200882
</commentlist>
</conversation>
<conversation>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#conversation09_11_23_090213.5</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200612
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30202766
</commentlist>
</conversation>
<conversation>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#conversation09_11_23_090213.3</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200796
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30202722
</commentlist>
</conversation>
<conversation>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#conversation09_11_23_090213.6</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200316
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200740
-http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200444
</commentlist>
</conversation>
<conversation>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#conversation09_11_23_090213.4</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200338
</commentlist>
</conversation>
<conversation>
	<id>http://www.semanticweb.org/ontologies/ConversationInstances.owl#conversation09_11_23_090213.1</id>
	<commentlist>http://www.semanticweb.org/ontologies/ConversationInstances.owl#comment09_11_23_090213.30200478
</commentlist>
</conversation>
