b10229

ggml-org/llama.cppb10229Aug 2, 2026by github-actions[bot]same day as b10228

What changed since b10228

low risk

b10229 fixes OpenCL backend reference counting bug preventing profiling data flush

Bug fixes

  • Fixed incorrect reference counting in the OpenCL backend that could lead to profiling data loss during cleanup.

Upgrade notes

No configuration changes are required. This is a patch release addressing a backend stability issue.

Analyzed Oct 16, 2026

Upgrading?

Release notes summary

Fixed a critical bug in the OpenCL backend where reference counting was not incremented correctly, which could prevent profiling data from being flushed and written during cleanup.

Key Highlights

  • Fixed OpenCL backend reference counting in `ggml_backend_opencl_init()`.
  • Ensures profiling data is properly flushed and written.
  • Prevents potential profiling data loss.

New Features

  • OpenCL profiling bugfix

Release Notes (excerpt)

<details open> opencl: bugfix increment ref_count in ggml_backend_opencl_init() (#26162) Incrementing `ref_count` at the beginning is important later in the `free()` method of the `ggml_backend_opencl_context` at program end. If we do not increment the `ref_count`, the result would be -1 here, and consequently, the profiling data would not be flushed and written. ( #ifdef GGML_OPENCL_PROFILING ) </details> **Website:** - <https://llama.app> **macOS/iOS:** - macOS Apple Silicon (arm64)…

Published by ggml-org on GitHub.

Read the full release notes

Related Repositories