Package Details: llama.cpp-hip-git b9881.r1.g4871961-1

Git Clone URL: https://aur.archlinux.org/llama.cpp-hip-git.git (read-only, click to copy)
Package Base: llama.cpp-hip-git
Description: Port of Facebook's LLaMA model in C/C++ (with AMD ROCm optimizations, git version)
Upstream URL: https://github.com/ggml-org/llama.cpp
Licenses: MIT
Conflicts: ggml, ggml-rocm, libggml, llama-cpp, llama.cpp, llama.cpp-hip, stable-diffusion.cpp
Provides: ggml, ggml-rocm, libggml, libggml-hip.so, llama-cpp, llama.cpp, llama.cpp-hip
Submitter: Orion-zhen
Maintainer: Orion-zhen
Last Packager: Orion-zhen
Votes: 0
Popularity: 0.000000
First Submitted: 2026-07-06 08:26 (UTC)
Last Updated: 2026-07-06 08:26 (UTC)

Dependencies (19)

Required by (18)

Sources (3)

Pinned Comments

Orion-zhen commented on 2026-07-06 08:34 (UTC)

Currently, package maintainers have added the two packages llama-cpp-rocm and llama-cpp-vulkan to the extra repository, resulting in the deletion of llama.cpp-rocm and llama.cpp-vulkan, which had been actively updated and maintained on AUR for nearly two years. Considering that the package maintainer responsible for those two packages has not actively kept up with upstream version updates as of the time I write this comment, I have recreated this -git version of the package so that users can experience the latest code repository.

See also: llama.cpp-vulkan-git maintained by rubin55.

Latest Comments

s-k commented on 2026-07-09 08:21 (UTC)

why did you move from the tarball?

Orion-zhen commented on 2026-07-06 08:34 (UTC)

Currently, package maintainers have added the two packages llama-cpp-rocm and llama-cpp-vulkan to the extra repository, resulting in the deletion of llama.cpp-rocm and llama.cpp-vulkan, which had been actively updated and maintained on AUR for nearly two years. Considering that the package maintainer responsible for those two packages has not actively kept up with upstream version updates as of the time I write this comment, I have recreated this -git version of the package so that users can experience the latest code repository.

See also: llama.cpp-vulkan-git maintained by rubin55.

EndlessEden commented on 2026-06-02 06:02 (UTC)

please add "sed -i 's/split_state.nr/split_state.ne/g' src/llama-model.cpp" to the prepare line. Upstream is lagging behind ggml by quite a bit. NR is now NE to make it generic.

edtoml commented on 2025-02-15 13:57 (UTC) (edited on 2025-02-15 13:58 (UTC) by edtoml)

It would seem most people are using llama.cpp-hip-git now. However that package gives me less than one token per second. This one works as expected if the build section of its PKGBUILD is updated to:

build() {
  local _cmake_args=(
    -B build
    -S .
    -DCMAKE_INSTALL_PREFIX=/usr
    -DCMAKE_BUILD_TYPE=Release
    -DGGML_STATIC=OFF
    -DGGML_HIP=ON
  )