-
Hong Zhang authored
- The HIP kernels are converted directly from their CUDA version - AMD GPUs and NVIDIA GPUs use different warp sizes. We set the warp size to 64 by default for AMD GPUs to faciliate compile-time code optimization
773bf0f6
As of 2024-05-14 a change has been implemented to enforce access token expiry. If you see higher numbers of HTTP 401 responses, please check tokens that may have expired. For more details, including how to mitigate disruption, please check our docs guide on Expired Access Tokens. If you have a support contract, you may open a support ticket for additional assistance.
- The HIP kernels are converted directly from their CUDA version - AMD GPUs and NVIDIA GPUs use different warp sizes. We set the warp size to 64 by default for AMD GPUs to faciliate compile-time code optimization