b10336

ggml-org/llama.cppb10336Aug 10, 2026by github-actions[bot]1 days after b10333

What changed since b10333

low risk

WebGPU WGSL shaders refactored and Flash Attention code simplified, with no functional or API changes noted.

Improvements

  • Refactored multiple WebGPU WGSL shader files for cleaner structure and easier maintenance.
  • Simplified the Flash Attention WGSL implementation, reducing shader complexity.
  • Shader refactor may improve WebGPU compilation efficiency.

Upgrade notes

This is an internal WebGPU shader refactor with no reported CLI, config, or API changes, so a drop-in binary swap should suffice. Users running the WebGPU backend should rebuild or pull fresh binaries to pick up the reworked shaders. The SpaceMiT Q5_0 dispatch fix from the previous build remains included.

Analyzed Oct 29, 2026

Upgrading?

Release notes summary

Refactored WebGPU (WGSL) files to simplify the Flash Attention implementation, improving code maintainability and compilation efficiency.

Key Highlights

  • Refactored several WGSL files for WebGPU.
  • Simplified Flash Attention WGSL code.

Release Notes (excerpt)

<details open> ggml-webgpu : refactor several wgsl files and simplify flash_attn wgsl. (#26134) </details> **Website:** - <https://llama.app> **macOS/iOS:** - macOS Apple Silicon (arm64) - macOS Apple Silicon (arm64, KleidiAI enabled) DISABLED - macOS Intel (x64) - iOS XCFramework **Linux:** - Ubuntu x64 (CPU) - Ubuntu arm64 (CPU) - Ubuntu s390x (CPU) - Ubuntu x64 (Vulkan) - Ubuntu arm64 (Vulkan) - Ubuntu x64 (ROCm 7.2) - Ubuntu x64 (OpenVINO) - Ubuntu x64 (SYCL FP32)…

Published by ggml-org on GitHub.

Read the full release notes

Related Repositories