The sample projects demonstrate the basic usage of ILGPU to help you get started with high performance GPU programming.
One key feature of ILGPU is the ability to execute the code on the CPU in a way that emulates how the GPU works. Therefore, instead of having to resort to Graphics/CUDA debugging facilities, you can directly use all of Visual Studio's CPU debugging features.
Note that this is not possible when your code is executed on the GPU; in order to execute your code on the CPU, you have to create a CPU context instead of a GPU context (e.g. by replacing new CudaAccelerator with new CPUAccelerator).
After cloning the repository, the folder structure should look as follows:
- .git
- Src
- LICENSE.txt
- ...
Just cloning the repository should be sufficient as all dependencies are usually restored by NuGet. Please refer to the ILGPU readme for further dependencies, like the CUDA runtime.
These samples explain the basic and more advanced capabilities of ILGPU, respectively.
These samples show how to leverage ILGPU.Algorithms for even more comfortable GPU programming. Note that some of the algorithms samples might not work on all GPUs due to their hardware capabilities.
There are a few settings that you should remember to change for your own projects:
- In the project properties, set the target framework to .NET Framework 4.7 or .NET Core 2.0.
- Make sure that Prefer 32-bit is disabled in the project settings and/or that the target platform is set to X64.