b9786

ggml-org/llama.cppb9786Jun 25, 2026by github-actions[bot]same day as b9785

What changed since b9785

low risk

Adds OpenCL normalization support while maintaining chat security hardening.

New features

  • OpenCL non-contiguous row support in normalization operations

Improvements

  • Chat mode input handling hardened for improved security

Upgrade notes

No configuration changes are required. Simply replace the binary with the new release.

Analyzed Oct 18, 2026

Upgrading?

Release notes summary

Enhances the OpenCL backend to support non-contiguous memory rows in normalization operations, improving compatibility with various memory layouts.

Key Highlights

  • OpenCL non-contiguous row support
  • Normalization operation improvements

Release Notes (excerpt)

<details open> opencl: support non-contig rows in norm (#24965) </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