# llama.cpp release adds OpenCL dp4a matrix multiply variants

> The open-source llama.cpp library version b11789 adds new OpenCL kernels for faster matrix math and provides fresh binaries for macOS, iOS and Ubuntu.

Oossa · 2026-10-21 · https://oossa.com/en/llama-cpp-release-adds-opencl-dp4a-matrix-multiply-variants

The ggml‑org team released llama.cpp version b11789. The update adds dp4a GEMM (matrix‑multiply) variants for OpenCL with vector activation loads, and renames related kernels and environment variables. Pre‑built binaries are now available for Apple Silicon, Intel macOS, iOS, and several Ubuntu builds (CPU, Vulkan, CUDA). The changes are aimed at improving performance on GPUs that support the dp4a instruction.

## The facts

- Release version b11789 published on 2026-10-21
- Adds OpenCL dp4a GEMM kernels with vector activation loads

## Why it matters

Developers can run llama.cpp faster on compatible GPUs, reducing inference time for local LLM applications.

## Sources & references

1. [ggml-org/llama.cpp b11789](https://github.com/ggml-org/llama.cpp/releases/tag/b11789) – llama.cpp, 2026-10-21

Last updated: 2026-10-21