Search Criteria
Package Details: ggml-cuda-backend 0.17.0-1
Package Actions
| Git Clone URL: | https://aur.archlinux.org/ggml-cuda-backend.git (read-only, click to copy) |
|---|---|
| Package Base: | ggml-cuda-backend |
| Description: | CUDA backend for ggml |
| Upstream URL: | https://github.com/ggml-org/ggml |
| Licenses: | MIT |
| Provides: | ggml-acceleration |
| 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) |
Dependencies (11)
- cuda (cuda11.1AUR, cuda-12.2AUR, cuda12.0AUR, cuda11.4AUR, cuda-12.5AUR, cuda-12.8AUR, cuda-pascalAUR, cuda-12.9AUR)
- ggml-coreAUR (ggml-core-llama.cppAUR)
- glibc (glibc-gitAUR, glibc-git-native-pgoAUR, glibc-eacAUR)
- libgcc (libgcc-fast-optimizedAUR, libgcc-snapshotAUR)
- libstdc++ (libstdc++-fast-optimizedAUR, libstdc++-snapshotAUR)
- nccl (nccl-gitAUR, nccl-cuda12.9AUR)
- nvidia-utils (nvidia-410xx-utilsAUR, nvidia-440xx-utilsAUR, nvidia-430xx-utilsAUR, nvidia-340xx-utilsAUR, nvidia-510xx-utilsAUR, nvidia-utils-teslaAUR, nvidia-575xx-utilsAUR, nvidia-340xx-utils-macbookAUR, nvidia-535xx-utilsAUR, nvidia-470xx-utilsAUR, nvidia-390xx-utilsAUR, nvidia-550xx-utilsAUR, nvidia-525xx-utilsAUR, nvidia-580xx-utilsAUR, nvidia-vulkan-utilsAUR, nvidia-utils-betaAUR)
- cmake (cmake3AUR, cmake-gitAUR) (make)
- cudnn (cudnn9.10-cuda12.9AUR, cudnn-pascalAUR) (make)
- ninja (ninja-gitAUR, ninja-memAUR, ninja-noemacs-gitAUR, ninja-kitwareAUR, ninja-fuchsia-gitAUR, n2-gitAUR) (make)
- patch (patch-gitAUR) (make)
Required by (2)
- llama.cpp-system (requires ggml-acceleration) (optional)
- stable-diffusion.cpp-system (requires ggml-acceleration) (optional)
Latest Comments
NewYearPrism commented on 2026-07-02 19:40 (UTC)
By default, ggml CUDA enables all GPU targets supported by host CUDA toolkit. You can set environment variable CUDA_TARGETS=native to let CUDA decide native GPU, which reduces build time and binary size. Or set CUDA_TARGETS=89 (or any specific architecture).