Advancing GPU Programming with the CUDA Tile IR Backend for OpenAI Triton

NVIDIA CUDA Tile is a GPU-based programming model that targets portability for NVIDIA Tensor Cores, unlocking peak GPU performance. One of the great things…

NVIDIA CUDA Tile is a GPU-based programming model that targets portability for NVIDIA Tensor Cores, unlocking peak GPU performance. One of the great things about CUDA Tile is that you can build your own DSL on top of it. This post shares the work NVIDIA is doing to integrate CUDA Tile as a backend for OpenAI Triton, an open source Python DSL designed to write DL kernels for GPUs.

Source

Leave a Reply

Your email address will not be published.

Previous post Beyond Good and Evil 2 somehow survives the bloodbath at Ubisoft: creative director says he’s ‘saddened’ by the layoffs and cancellations, but the 19-year project is ‘unaffected’
Next post Top football manager reminds football media that real football ‘is not Football Manager, unfortunately’