scuda
scuda copied to clipboard
kevmo314
SCUDA is a GPU over IP bridge allowing GPUs on remote machines to be attached to CPU-only machines.
If the `cudaLaunchKernel()`'d kernel includes a call to `printf()`, it will print on the server. This is correct behavior as far as CUDA is concerned but makes using SCUDA a...
You can see an example of a failing build for cudnn8 here: [https://github.com/kevmo314/scuda/actions/runs/12663027033/job/35288904058](url) IMO, the easiest approach to fixing this would be to create a new docker image using `docker.io/nvidia/cuda:12.1.0-cudnn8-devel-ubuntu22.04`....
Add build files (Docker), as stated in https://github.com/kevmo314/scuda/issues/31, i would like to use earthly to build binaries (docker does not seem to be the correct tool to do this). We...
Hi. Thx for the great project :) There are some issues with compatibility with cudnn version < 9.5.0: cudnnBackendUpdateCudaGraph and cudnnBackendPopulateCudaGraph should be wrapped to #if (in gen_server.cpp) when cudnn...
I am trying to get juice working but no luck. They did have windows server. The use case is for homelab environment where I pass through my GPU to a...
Trying to compile the Dockerfile from source and stuck on /opt/cuda/bin/nvcc -shared -o libscuda.so client.c Where is client.c?
