monte-carlo statistics program that tests the scalablilty of certain operations
Of note, these programs won't be terribly accurate, as they rely on good random numbers, but because I didn't feel like putting a good source of random data or a good PRNG, it might not line up quite right. But that isn't the point of this benchmark.
make -j whatever
Next, cease any intensive work on the machine
make graph.png -j 1