CudaKernel

Cuda kernels are load from cubin or ptx files. You can load a kernel using the LoadKernel…() methods of a CudaContext using a byte array representation of the kernel file (e.g. an embedded resource) or by specifying the file name where the kernel is stored. Further you need the kernel name as defined in the source *.cu file. The LoadKernel methods return a CudaKernel object bound to the given context. CudaKernel does not implement IDisposable, as the kernels are automatically destroyed as soon as the corresponding context is destroyed.

Example code:

//One kernel per cu file:
CudaKernel kernel = ctx.LoadKernel(@"path\to\kernel.ptx", "kernelname");
kernel.GridDimensions = new dim3(1, 1, 1);
kernel.BlockDimensions = new dim3(16, 16),

//Multiple kernels per cu file:
CUmodule cumodule = ctx.LoadModule(@"path\to\kernel.ptx");
CudaKernel kernel1 = new CudaKernel("kernel1", cumodule, ctx)
{
	GridDimensions = new dim3(1, 1, 1),
	BlockDimensions = new dim3(16, 16),
};
CudaKernel kernel2 = new CudaKernel("kernel2", cumodule, ctx)
{
	GridDimensions = new dim3(1, 1, 1),
	BlockDimensions = new dim3(16, 16),
};