# llama.cpp adds correct CLAMP handling for strided tensors

> The Oct 15, 2026 update fixes a bug in the CLAMP operation when inputs are non‑contiguous, improving accuracy on many platforms.

Oossa · 2026-10-15 · https://oossa.com/en/llama-cpp-adds-correct-clamp-handling-for-strided-tensors

The open‑source llama.cpp library released version b11639 on Oct 15, 2026. It adds a new kernel that correctly applies the CLAMP operation to non‑contiguous (strided) tensors. The previous kernel read the wrong elements when the source array wasn’t stored contiguously. The fix introduces a strided version called kernel_clamp_nc, used automatically whenever needed.

## The facts

- Release date: Oct 15, 2026
- Fix: adds kernel_clamp_nc for non‑contiguous CLAMP

## Why it matters

Developers can now rely on accurate CLAMP results when running models on hardware that uses strided memory layouts.

## Sources & references

1. [ggml-org/llama.cpp b11639](https://github.com/ggml-org/llama.cpp/releases/tag/b11639) – llama.cpp, 2026-10-15

Last updated: 2026-10-15