gpu-normal-computation

Repository accompanying the associated Kapernikov blog post.

Build and install

Prerequisites

• OpenCL development files

Instructions

The build uses cmake. Use the usual cmake incantations with an appropriate generator. Make sure to build a Release build when running the benchmarks (add -DCMAKE_BUILD_TYPE=Release to the cmake command line). In addition, install the binaries. A custom install folder can be set by passing -DCMAKE_INSTALL_PREFIX=<custom install dir> to the cmake command line. The default installation directory is /usr/local. See below for some examples on specific platforms.

Note: Check the .gitlab-ci.yml file for an example system installation and configuration.

Examples

Linux

The example uses the make generator.

$ cmake -H. -Bbuild -DCMAKE_INSTALL_PREFIX=install -DCMAKE_EXPORT_COMPILE_COMMANDS=ON -DCMAKE_BUILD_TYPE=Release
$ make -C build --jobs 4
$ make -C build install

The required binaries will then be present in the install directory.

Windows

The example uses the ninja generator.

Prepare the build environment:

1. Open the appropriate build environment shell
2. Make sure that cmake.exe and ninja.exe are in your path

Then execute:

$ cmake -G "Ninja" -H. -Bbuild -DCMAKE_INSTALL_PREFIX=install -DCMAKE_EXPORT_COMPILE_COMMANDS=ON -DCMAKE_BUILD_TYPE=Release
$ ninja -C build --jobs 4
$ ninja -C build install

The required binaries will then be present in the install directory.

Run

Prerequisites

• A properly functioning OpenCL runtime
• For the benchmarks: A tool to fix your CPU to a specific frequency (e.g. cpupower on *NIX systems)
• For the benchmarks: A binary built in release mode

Benchmarks

Linux

$ <installation_dir>/bin/benchmark_normal_computation

Windows

$ <installation_dir>\bin\benchmark_normal_computation.exe

Unittests

Linux

$ <installation_dir>/bin/test_normal_computation

Windows

$ <installation_dir>\bin\test_normal_computation.exe