Build in a separate folder, perhaps fast-wc
In the parent directory, put the linux-5.8-rc7 source files (about 75,000 files).  To make this a "pure" word count, I have the actual file names in fast-wc.hpp, and it assumes the tree is ../linux-5.8-rc7/... 

For gprof, compile this way:

c++ -pg -O3 -o fast-wc fast-wc.cpp -lpthread

To run the program with 8 threads and 8 cores, silently:

taskset 0xFF ./fast-wc -n8 -s

On my laptop, this takes about 14s.  On a server, perhaps 3.5s to 4.5s.  Not 0.0000

Then run gprof ./fast-wc, or try gprof -Awcounter ./fast-wc


