b9699

ggml-org/llama.cppb9699Jun 18, 2026by github-actions[bot]same day as b9698

What changed since b9698

low risk

Adds SYCL backend support for Q1_0 quantized MUL_MAT and OUT_PROD, plus FP32/FP16 SYCL builds.

New features

  • SYCL backend can now execute MUL_MAT and OUT_PROD operations on Q1_0-quantized tensors.
  • SYCL FP32 and FP16 build variants are now enabled and published as release artifacts.

Improvements

  • Expands GPU-accelerated quantization coverage for Intel/SYCL users working with Q1_0 models.

Upgrade notes

This is a routine build-to-build bump with no config or API changes; existing setups continue to work unchanged. SYCL users running Q1_0 models should see offload support where it previously fell back, and can pick the new FP32/FP16 SYCL artifacts if needed.

Analyzed Nov 1, 2026

Upgrading?

Release notes summary

A build release adding SYCL support for Q1_0 quantization operations.

Key Highlights

  • SYCL support for MUL_MAT and OUT_PROD with Q1_0
  • Enabled SYCL FP32 and FP16 builds

Release Notes (excerpt)

<details open> sycl : support MUL_MAT and OUT_PROD with Q1_0 (#24721) </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