CUDA code runs on AMD's RX 9060 XT via ZLUDA and ROCm
CUDA for AMD on Windows
A reproducible Windows setup lets CUDA-targeted applications run on AMD GPUs using ZLUDA and the AMD HIP SDK. Validated on the Radeon RX 9060 XT (gfx1200), it passes cuda_check for nvcuda, cuBLAS, cuBLASLt, cuSPARSE, and cuFFT, and completes a 2.2-million-parameter PPO training run. cuDNN is unavailable, so convolution-heavy workloads may fail. Other AMD GPUs are unverified candidates.
The public, upstream-only path has been tested without any private/recovered DLLs.
- linuxhansl
Off-topic and somewhat of a rant, but I'd far prefer us all focusing on open standards like HIP, SYCL, OpenCL, etc.
It's unbearable that most LLM inference happens on closed H/W, closed drivers, and closed SDKs.
- swerner
AI will take down Nvidia’s moat. When it becomes trivial to translate CUDA/PTX to HIP, SYCL or Metal, CUDA is no longer the moat, it becomes the intermediate representation.
- lulzx
I made cuda-metal btw (for mac kek), https://github.com/lulzx/cuda-metal
- Nurysso
man i can't say how much i used to like cuda when i had a nvidia gpu it made ml so much fun and on amd its a war especially on rdna 2 cards which i have. hopefully one day we will be able to properly translate cuda for its amd counter parts
- latchkey
There are also interesting efforts like: