General purpose use of the GPU is common now. And most importantly, Matrix Multiplication is the first in OpenCL tutorials. Instead of writing codes and kernel codes for a particular gpu. Is it possible to call them from a library, for example, MKL. Arrayfire is, but not for free. I am trying to port my simulation software to a GPU that has Matrix Mul, LU decomposition, FFT, etc. I am looking for reliable codes for them, instead of writing them from scratch. I already have a mul opencl working program with me, but I just don’t have enough knowledge about integrating with my code. In the next step, trying with LU, is there any code for LU that could help me.
source share