b9733

ggml-org/llama.cppb9733Jun 20, 2026by github-actions[bot]same day as b9732

What changed since b9732

low risk

Adds WebGPU adapter toggles enabling F16 precision on Vulkan and NVIDIA GPUs; otherwise unchanged from b9732.

New features

  • WebGPU backend gains configuration toggles to enable F16 precision specifically on Vulkan and NVIDIA adapters.

Improvements

  • Expands WebGPU backend configuration options, letting users opt into F16 on adapters where it was previously unavailable.

Upgrade notes

This is a single-commit increment over b9732 focused on the WebGPU backend; no config migration or API changes are required. Users on Vulkan or NVIDIA hardware via WebGPU can enable the new F16 toggles to potentially improve throughput, but should validate output quality since F16 may affect numerics.

Analyzed Oct 24, 2026

Upgrading?

Release notes summary

Added configuration toggles to enable F16 precision on Vulkan and NVIDIA adapters in the WebGPU backend.

Key Highlights

  • Added adapter toggles for F16 on Vulkan and NVIDIA
  • Enhanced WebGPU backend configuration options

Release Notes (excerpt)

<details open> ggml-webgpu: add adapter toggles for F16 on Vulkan + NVIDIA </details> **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) - Ubuntu x64 (SYCL FP16) **Android:** - Android arm64 (CPU) **Windows:**…

Published by ggml-org on GitHub.

Read the full release notes

Related Repositories