Back to mainpage

Benchmarks presented below were gathered by running CMT on every listed machine with same parameters. We ran CMT with 1024 input bytes and did 10000 iterations in 10 tests. The validation process was performed with probability of 80% that gathered results are reliable, and coefficient of variation equal to 8%. Before each actual measurement, we did a one second long CPU warming. Given the benchmarks below, one can easily compare obtained measurement results and analyze the performance of tested machines.

Here one can find the exact commands we ran CMT with to gather standard CMT results:

For GNU/Linux:
./CMTool -l 1024 -i 10000 -t 10 -p 8 -cl 80 -m 0 md4 md5 whirlpool ripemd160 sha hmac aes camellia blowfish seed rc4 rc2 des dsa rsashaed
(CMT command for single core call. For multi core call we used the number of physical cores of the machine as the argument of the -m option)

For Microsoft Windows:
CMTool.exe -l 1024 -i 10000 -t 10 -p 8 -cl 80 -m 0 md4 md5 whirlpool ripemd160 sha hmac aes camellia blowfish seed rc4 rc2 des dsa rsashaed
(CMT command for single core call. For multi core call we used the number of physical cores of the machine as the argument of the -m option)





For single core benchmarks OpenMP support was disabled so only one core of the CPU was used for these benchmarks.




For multi core benchmarks OpenMP support was enabled so the given number of cores of the CPU(s) were used for these benchmarks.
Benchmarks were run on all available CPU(s) cores, we gathered benchmarking results from every thread we ran.