While most hardware vendors are relying on LLVM when it comes to offloading compute work to GPUs, AI accelerators, FPGAs, and similar heterogeneous compute environments, right now each vendor is basically creating their own LLVM offloading run-time among a lot of other duplicated -- and often downstream only -- code. The new "llvm/offload" project hopes to lead to better collaboration in this area...
Following an FCV optimization for the latest Intel graphics hardware, ASTC LDR emulation, some still-pending Vulkan sparse support for ANV atop the existing i915 driver, and other recent Intel open-source "ANV" Vulkan driver optimizations, another optimization was just merged into Mesa 23.3...