Skip to content

Latest commit

 

History

History
15 lines (13 loc) · 801 Bytes

Readme.md

File metadata and controls

15 lines (13 loc) · 801 Bytes
Raw

Speed comparison between naive implementation and OpenBLAS

This is a simple speedtest with matrix matrix multiplication and matrix vector multiplication with OpenBLAS and a naive implementation. You can change the value rounds to any desired number to alter the time used for the calculation.

Usage

If you installed OpenBLAS with the standard directory, you will need to use:

  • export LD_LIBRARY_PATH=/opt/OpenBLAS/lib/ (or to your directory with OpenBLAS)
  • g++ -o benchmark main.cpp -I /opt/OpenBLAS/include/ -L/opt/OpenBLAS/lib -lopenblas for compiling

Notes

The time is used from the CPU clock and may differ from realtime but since most of the time is running on the CPU, I prefer this way. If you have a different opinion, feel free to use time().