David Airlie Shares His Thoughts On Current Challenges With Linux GPU Compute Stacks

Sriram Ramkrishna at Intel, who serves as the community manager and developer relations for oneAPI, held a virtual oneAPI meetup this week with Red Hat's David Airlie. Airlie should not need any introduction for longtime Phoronix readers given his longtime contributions to the Linux kernel graphics drivers, Mesa, and related open-source graphics work at Red Hat. Airlie shared some interesting remarks around the current Linux GPU compute stacks from the different vendors and associated challenges.

The oneAPI meetup with David Airlie mostly focused on his thoughts around the different graphics compute stacks, the areas that can be improved upon, and more. While mostly focused on the open-source angle, Airlie does acknowledge the success of NVIDIA's CUDA thanks to it working across all NVIDIA GPUs regardless of consumer or workstation/data-center focus. He also calls out vendors sometimes being too HPC/data-center focused and their corporate organizational structure where they can sometimes be too segmented between consumer and professional products.

Airlie also touched on the challenges from the Linux distribution perspective from vendors not understanding open-source to the all too frequent LLVM/Clang forks maintained by different driver stacks. And of course the most open-source friendly approach for GPU computing being around open standards from The Khronos Group like SYCL, OpenCL, and Vulkan compute.

Those interested in Airlie's thoughts/rants about these topics can find the oneAPI meetup presentation embedded below for some interesting weekend content.