FICHA · MANJARO

ggml-cublas-git

Tensor library for machine learning (with NVIDIA CUDA optimizations)

  • Library
  • Dev
  • HARDWARE
official+codex · reviewed · May 30, 2026 description in en

Description

Machine-learning tensor operations can use the Git version of ggml with NVIDIA CUDA acceleration. It is useful for developers and local AI tools that need GPU-backed inference or experimentation on supported NVIDIA hardware.

GPU-accelerated ML libraries depend on drivers, CUDA compatibility, and model workload. Users should expect higher power use and should test stability before relying on it for long-running jobs.

Permissions

Permissions not analysed for this source yet.