Performance gains increasingly come from low-level kernel and driver work—unified GEMM tuning, jagged attention kernels, and open GPU/compute stacks—rather than model scaling alone. The focus is moving toward efficient execution on Hopper, Blackwell, and alternative GPU architectures.
AMD's latest DRM-Next pull for Linux 7.4 is mostly bug fixes, but it also enables IP blocks for future GPUs — a quiet signal for anyone planning self-hosted LLM infrastructure.
Helion's linear backend in vLLM shows a single GEMM kernel with an ahead-of-time autotuner can cover specialized variants and improve inference by up to 10% on Qwen3. For self-hosted LLM teams, the signal is twofold: less dependence on deep CUDA skil...
Red Hat and Meta integrated Helion into vLLM's linear backend. A single GEMM implementation covers Standard, Split-K, and Swap-AB variants, with per-shape autotuning on NVIDIA Hopper GPUs. On H100 tests, quantized kernels outperform CUTLASS, FlashInf...
At XDC 2026 Toronto, Imagination detailed the state of the Mesa Vulkan driver for PowerVR and the upstream PVR DRM kernel driver, with plans for the new Volcanic GPU architecture. Less announcement, more infrastructure: an open-source driver reduces ...
The open-source Intel Iris Gallium3D OpenGL driver has received support for an efficient 64-bit addressing mode. The change follows the announcement of a new 64-bit GPU mode for Nova Lake P graphics. For those using Intel GPUs in local scenarios, the...
Meta's GEM attention kernel, rewritten in TLX, beats FlashAttention-4 on jagged Blackwell workloads: +13% forward and +50% backward on average, with about 3,200 lines versus 10,000. The result shows explicit hardware control no longer requires hand-w...
Edison Design Group has open-sourced its C/C++ front-end, a commercial component known for its extensive dialect support and used by Intel C++ Compiler classic, NVIDIA CUDA NVCC, and Microsoft Visual Studio IntelliSense. The move makes a core part of...
Mesa is developing NOCL, a Gallium3D implementation on top of NVIDIA's CUDA driver API. Paired with Rusticl, it allows OpenCL workloads to run using NVIDIA's proprietary driver. A compute option that shifts control from the CUDA ecosystem to an open ...
