b10354

ggml-org/llama.cppb10354Aug 10, 2026by github-actions[bot]same day as b10353

What changed since b10353

low risk

b10354 is a single Android-specific fix for CPU affinity masks being ignored in ggml-cpu.

Improvements

  • CPU affinity settings are now honored on Android, restoring intended thread pinning and scheduling behavior.
  • Thread placement on Android devices should be more predictable, avoiding performance regressions from unpinned worker threads.

Bug fixes

  • Fixed ggml-cpu ignoring the configured CPU affinity mask on Android.

Upgrade notes

No configuration or API changes are required; this is a drop-in replacement for b10353. Android deployments that rely on thread affinity tuning (e.g. via --cpu-mask or thread count settings) should upgrade to regain the intended behavior. Non-Android platforms see no functional change.

Analyzed Oct 31, 2026

Upgrading?

Release notes summary

Fixes a bug where CPU affinity masks were being ignored on Android, which could lead to incorrect thread scheduling or performance issues on Android devices.

Key Highlights

  • Fixed CPU affinity mask being ignored on Android.

Release Notes (excerpt)

<details open> ggml-cpu : fix CPU affinity mask being ignored on Android (#26838) </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)…

Published by ggml-org on GitHub.

Read the full release notes

Related Repositories