/float-performance-test

test your CPU's float performance with C++ and openmp

Primary LanguageC++GNU General Public License v3.0GPL-3.0

float-performance-test

test your CPU's float performance with C++ and openmp.

functions

This program can not only calculate your CPU's FP32, FP64 performance, but automatically compute the speedup rate, parallel efficiency and parallelism rate.

complie command

  • for macOS
g++ -Ofast -Xpreprocessor -fopenmp -I/opt/local/include -L/opt/local/lib -lomp FPtest.cpp -o FPtest
  • for windows/Linux
g++ -Ofast -fopenmp FPtest.cpp -o FPtest
  • if support you may add AVX2 or AVX512 intrinsics
#for AVX2
g++ -Ofast -fopenmp -mavx2 FPtest.cpp -o FPtest

run command

./FPtest [test size] [threads] [test times]

for example

./FPtest 100000000 4 10

Copyrights

Copyright (c) SCUT. 2022. All rights reserved.

This program is a benchmark for floating point performance. It is designed to be run on both single and double precision systems. In multi-core systems, it is designed to be run on omp.