b10092

ggml-org/llama.cppb10092Jul 23, 2026by github-actions[bot]same day as b10091

What changed since b10091

low risk

Adds PowerPC backend support for AIX and fixes a PowerPC Q0 matmul build error; no impact on other platforms.

New features

  • PowerPC CPU backend variants can now be built on AIX via an extended CMake platform check, reusing existing PowerPC implementations unchanged.

Improvements

  • AIX joins the set of platforms with first-class PowerPC backend build support, with no behavioral change to existing backends.
  • Standard multi-platform binary set (macOS, Linux, Windows, Android, openEuler) continues to be published for this build.

Bug fixes

  • Corrected a missing semicolon in the PowerPC Q0 matmul implementation in sgemm.cpp that prevented compilation.

Upgrade notes

This is a build-system and compile-fix release; only AIX/PowerPC builders see any difference, and no runtime flags or configuration need to change. Users on other architectures can upgrade as a drop-in replacement with the same binaries and CLI behavior.

Analyzed Nov 1, 2026

Upgrading?

Release notes summary

Enables PowerPC backend variants on AIX and fixes a missing semicolon in the PowerPC Q0 matmul implementation.

Key Highlights

  • Enable PowerPC backend variants on AIX
  • Fix missing semicolon in PowerPC Q0 matmul implementation
  • Multi-platform binaries maintained

Release Notes (excerpt)

<details open> ggml: enable PowerPC backend variants on AIX (#25983) * ggml: enable PowerPC backend variants on AIX Allow the PowerPC CPU backend variants to be built on AIX by extending the platform check in the CMake configuration. This reuses the existing PowerPC backend implementations without changing their behavior. Also fix a missing semicolon in the PowerPC Q0 matmul implementation. * Fix missing semicolon in sgemm.cpp </details> **Website:** - <https://llama.app> **macOS/iOS:**…

Published by ggml-org on GitHub.

Read the full release notes

Related Repositories