| Git Clone URL: | https://aur.archlinux.org/ggml-core.git (read-only, click to copy) |
|---|---|
| Package Base: | ggml-core |
| Description: | Tensor library for machine learning (core runtime) |
| Upstream URL: | https://github.com/ggml-org/ggml |
| Licenses: | MIT |
| Groups: | ggml |
| Conflicts: | ggml, libggml |
| Provides: | ggml, libggml |
| Submitter: | NewYearPrism |
| Maintainer: | NewYearPrism |
| Last Packager: | NewYearPrism |
| Votes: | 0 |
| Popularity: | 0.000000 |
| First Submitted: | 2026-07-02 19:26 (UTC) |
| Last Updated: | 2026-07-19 05:16 (UTC) |
Note: CUDA and ROCm DO NOT conflicts in ggml. My 4060Ti and RX9070 works fine together on CUDA0 and ROCm0, to run a model that cannot fit into a single one's VRAM, also performace being even better than Vulkan0+1. Just install ggml-cuda-backend and ggml-hip-backend :)
Maintainance in https://github.com/NewYearPrism/ggml-PKGBUILDs
need to remove old package first before building new version
$ sudo pacman -Rscn ggml-core
Pinned Comments
NewYearPrism commented on 2026-07-02 19:32 (UTC) (edited on 2026-07-02 19:33 (UTC) by NewYearPrism)
Note: CUDA and ROCm DO NOT conflicts in ggml. My 4060Ti and RX9070 works fine together on CUDA0 and ROCm0, to run a model that cannot fit into a single one's VRAM, also performace being even better than Vulkan0+1. Just install ggml-cuda-backend and ggml-hip-backend :)
Maintainance in https://github.com/NewYearPrism/ggml-PKGBUILDs
need to remove old package first before building new version
$ sudo pacman -Rscn ggml-core