I'm going to test my FDTD simulation code on my old Nvidia GTX 970 card as well, giving its "Maxwell" architecture a new meaning...
Follow
@niconiconi Have you adapted your ROCm code for CUDA or maintained a different branch?
I'm going to test my FDTD simulation code on my old Nvidia GTX 970 card as well, giving its "Maxwell" architecture a new meaning...
@niconiconi Have you adapted your ROCm code for CUDA or maintained a different branch?
@fulkrum@qoto.org It's SYCL, so it compiles to OpenCL SPIR-V target, AMD HIP target and CUDA target automatically.