Update the performance benchmark results
hungptit opened this issue · comments
From my latest benchmark results, grep is faster than ucg even though it only uses a single thread. Beside that can you add fgrep into the list?
Below are the performance benchmark results obtained using boost source code and Mark Twain's book in my MacBook pro.
./all_tests
Celero
Timer resolution: 0.001000 us
-----------------------------------------------------------------------------------------------------------------------------------------------
Group | Experiment | Prob. Space | Samples | Iterations | Baseline | us/Iteration | Iterations/sec |
-----------------------------------------------------------------------------------------------------------------------------------------------
mark_twain | grep | 0 | 5 | 1 | 1.00000 | 1299385.00000 | 0.77 |
boost_source | grep | 0 | 5 | 1 | 1.00000 | 1545215.00000 | 0.65 |
mark_twain | ag | 0 | 5 | 1 | 1.73053 | 2248620.00000 | 0.44 |
mark_twain | ripgrep | 0 | 5 | 1 | 0.68130 | 885275.00000 | 1.13 |
mark_twain | ucg | 0 | 5 | 1 | 0.96383 | 1252389.00000 | 0.80 |
mark_twain | fgrep_mmap | 0 | 5 | 1 | 0.54665 | 710308.00000 | 1.41 |
mark_twain | fgrep_default | 0 | 5 | 1 | 0.50417 | 655106.00000 | 1.53 |
boost_source | ag | 0 | 5 | 1 | 1.16845 | 1805503.00000 | 0.55 |
boost_source | ripgrep | 0 | 5 | 1 | 1.05553 | 1631014.00000 | 0.61 |
boost_source | ucg | 0 | 5 | 1 | 1.56576 | 2419435.00000 | 0.41 |
boost_source | fgrep | 0 | 5 | 1 | 1.11264 | 1719263.00000 | 0.58 |
Complete.
Could you please give enough details such that someone else can reliably reproduce your benchmark?
I forgot to include link to my benchmark. You can find a more information about how I create my benchmark here https://github.com/hungptit/fastgrep and all binaries can be downloaded this repo https://github.com/hungptit/tools. Let me know if you need more information.