# llama.cpp update fixes CUDA version check regression

> The Oct 8, 2026 release corrects a version‑guard bug that slowed CUDA arg‑sort on newer CCCL releases.

Oossa · 2026-10-08 · https://oossa.com/en/llama-cpp-update-fixes-cuda-version-check-regression

The llama.cpp project released version b11509 on Oct 8, 2026. It fixes a CUDA guard that mis‑identified CCCL version 4.x as older, causing a silent performance drop in the argsort operation. The fix replaces the component‑wise check with a single packed integer comparison, restoring the fast strided‑iterator path. The change was tested on an RTX 4070 with CUDA 13.4 and passed all backend tests.

## The facts

- Release tag b11509 published on 2026-10-08
- Tested on RTX 4070 (sm_89) with CUDA 13.4

## Why it matters

Developers using llama.cpp on modern CUDA setups will see faster sorting performance without changing their code.

## Sources & references

1. [ggml-org/llama.cpp b11509](https://github.com/ggml-org/llama.cpp/releases/tag/b11509) – llama.cpp, 2026-10-08

Last updated: 2026-10-08