Skip to content
fft2m
Switch branches/tags
Code
This branch is 1405 commits behind master.
Contribute

Latest commit

 

Git stats

Files

Permalink
Failed to load latest commit information.
Type
Name
Latest commit message
Commit time
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

gpuOwL

GPU Lucas-Lehmer primality test.

Build

Use make or build.sh.

Prerequisites:

  • a C++ compiler (e.g. gcc, clang)
  • an OpenCL implementation (which provides the libOpenCL library), e.g. AMDGPU-Pro

The build is just a C++ compiler invocation on gpuowl.cpp, specifying the OpenCL library in the -L to the compiler, e.g.:

g++ gpuowl.cpp -ogpuowl -L/opt/amdgpu-pro/lib/x86_64-linux-gnu -lOpenCL

Usage

  • Get exponents for testing from GIPMS Manual Testing ( http://mersenne.org/ ). gpuOwL best handles exponents is the vicinity of 77M, less than 78M.
  • Copy the lines from GIMPS to a file named 'worktodo.txt'
  • Run gpuowl. It prints progress report on stdout and in gpuowl.log, and writes result lines to results.txt
  • Submit the result lines from results.txt to http://mersenne.org/ manual testing.

Command line arguments

  • It is possible to pass arguments to the OpenCL compiler by passing "-cl" to gpuowl. For example, to dump the compiled ISA use gpuowl -cl -save-temps ; to put the ISA into an existing folder 'isa' use gpuowl -cl -save-temps=isa/