Darshana Sanjeewan Adikari

Works and tests running happily (using the GPU) for rocBLAS (after adding archs to it too) :D, using Tensile & rocBLAS branch gfx10

Hi, I also had the same questions, did some digging to see what's needed. Usecases: 1. Inference with libtorch (C++), where the runtime performance for inference is even better (lesser...

Small update, I managed to port the current master branch for jit compilation. It is still a bit hacky and I'm planning to open a PR once 2.1 sources are...

Hi, My apologies for the late reply. I managed to get the conversion to work by following the recipe I shared above. However, I'm unable to share that implementation due...

I did manage to build rocBLAS and Tensile using their WIP branch `gfx10`. The unit tests for rocBLAS also went well with the GPU being actually used. Special thanks for...

Aha I was wondering if this is the exact issue. So this error means the kernels I got when building rocBLAS (and specifying gfx1010) aren't enough, I need to build...

That's sad to hear. I'm very happy with this card for my gaming uses and much cheaper than closest Nvidia card. I got mine in last December but now only...