Package Details: llama.cpp-vulkan-git b10068-1

Git Clone URL: https://aur.archlinux.org/llama.cpp-vulkan-git.git (read-only, click to copy)
Package Base: llama.cpp-vulkan-git
Description: Port of Facebook's LLaMA model in C/C++ (with Vulkan optimizations)
Upstream URL: https://github.com/ggerganov/llama.cpp
Licenses: MIT
Conflicts: ggml, libggml, llama.cpp, llama.cpp-cuda, llama.cpp-cuda-git, llama.cpp-git, llama.cpp-openvino, llama.cpp-openvino-git, llama.cpp-sycl, llama.cpp-sycl-git, llama.cpp-vulkan
Provides: llama.cpp, llama.cpp-git, llama.cpp-vulkan, llama.cpp-vulkan-git
Submitter: robertfoster
Maintainer: rubin55
Last Packager: rubin55
Votes: 3
Popularity: 0.71
First Submitted: 2024-11-15 20:39 (UTC)
Last Updated: 2026-07-19 20:26 (UTC)

Required by (13)

Sources (3)

Latest Comments

Internauta commented on 2026-07-14 23:21 (UTC)

I was wondering why it takes so much time to compile, and I found out that multicore compiling was missing. It can be enabled with cmake --build build -- -j $(nproc)

zkrx commented on 2026-03-17 14:40 (UTC)

ggml-vulkan-git package no longer exists

Bink commented on 2026-01-08 05:04 (UTC)

I noticed this is missing the -DGGML_VULKAN=1 flag to build for Vulkan. https://github.com/ggml-org/llama.cpp/blob/master/docs/build.md#common-steps

teleprint-me commented on 2025-12-19 22:15 (UTC)

I think this package name might be misleading. It should be labeled kompute, not vulkan. The backends have separate API's and Vulkan has more support.