b10226

ggml-org/llama.cppb10226Aug 2, 2026by github-actions[bot]same day as b10224

What changed since b10224

low risk

b10226 fixes SYCL iGPU classification on top of b10224's WebGPU F16 repeat support

Improvements

  • SYCL hardware detection now distinguishes integrated GPUs from discrete ones, improving backend stability on iGPU-equipped systems.

Bug fixes

  • Corrected misclassification of integrated GPUs in the SYCL backend, which could previously lead to wrong device handling.

Upgrade notes

No configuration or API changes are involved; this is a drop-in binary swap. SYCL users running on integrated GPUs are the primary beneficiaries and should upgrade, while others gain nothing beyond the prior b10224 WebGPU F16 repeat support. Note that the intermediate b10225 build is not described here, so any changes it carried are not reflected in this comparison.

Analyzed Nov 1, 2026

Upgrading?

Release notes summary

Fixes the classification of integrated GPUs (iGPUs) in the SYCL backend to ensure correct hardware detection.

Key Highlights

  • SYCL backend iGPU classification fix
  • Improved hardware detection logic

New Features

  • SYCL backend stability (iGPU fix)

Release Notes (excerpt)

<details open> sycl: fix classification of iGPUs (#26105) </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) - Ubuntu x64 (SYCL FP16) **Android:**…

Published by ggml-org on GitHub.

Read the full release notes

Related Repositories