MoFtZ

hi @LostBeard. ILGPU is primarily a .NET/MSIL to Cuda PTX compiler. And it can also handle .NET/MSIL to OpenCL C. ILGPU is able to perform this transformation with the help...

There is a major refactor underway, to lay the groundwork for the next generation of ILGPU. e.g. attempting to support AOT compilation, removing the runtime code generation, etc. It will...

Please note that "zero-copy" just means that the GPU has to read from Host Memory rather than Device Memory. However, the cost of reading from Host Memory is considerably more...

hi @alessandroparma. From the GPU perspective, the allocated memory is a contiguous block of 1D bytes. Concepts such as 2D and 3D view are just convenient ways to index into...

hi @delverOne25. Local arrays within a kernel are not well supported - particularly if you are trying to create 1_000_000 elements. This would apply to ILGPU, and also native Cuda....

hi @aka-nse, as discussed on Discord, the reason why your generic specialization is not working is because you are attempting to perform runtime type detection, to decide which function to...

@aka-nse here is a more concrete example code: ```CSharp interface IKernelFunction where T : struct { T ComputeValue(T value); } readonly struct CustomSinFloat : IKernelFunction { public float ComputeValue(float value)...

@ilehtoranta the `Accelerator` class has a `DefaultStream` property. For Cuda, the Accelerator is an instance of `CudaAccelerator`, and the Stream is an instance of `CudaStream`. The CudaStream class has a...