Skip to main content

CUDA→Metal universal backend shim (MetaXuda)

Project description

MetaXuda

MetaXuda is a CUDA runtime shim for Apple Silicon, written in Rust, that allows Numba CUDA kernels to run unmodified by mapping CUDA calls to Metal.


✨ Features

  • Drop-in replacement for libcudart.dylib / libcuda.dylib
  • Run Numba CUDA kernels (@cuda.jit) directly on Apple Metal
  • Includes precompiled Metal .metallib shaders for fused math ops
  • Ships with a stubbed libdevice.bc so no CUDA Toolkit is required

⚙️ Installation

Requirements

  • macOS 13+ with Apple Silicon (M1/M2/M3)
  • Python >=3.10
  • NumPy (>=1.23)
  • Numba (>=0.59)

Steps

# Clone the repo
git clone https://github.com/perinban/MetaXuda.git
cd MetaXuda-

# Back to project root
cd ..
pip install -e .

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

metaxuda-0.1.0.tar.gz (822.9 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

metaxuda-0.1.0-py3-none-any.whl (822.8 kB view details)

Uploaded Python 3

File details

Details for the file metaxuda-0.1.0.tar.gz.

File metadata

  • Download URL: metaxuda-0.1.0.tar.gz
  • Upload date:
  • Size: 822.9 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.13.3

File hashes

Hashes for metaxuda-0.1.0.tar.gz
Algorithm Hash digest
SHA256 a990340702e8f4783116860055e0b85c85c4985e7f6b452df4c047f77cc94945
MD5 f15b29b13c801aa633f6f8a2a9549e6b
BLAKE2b-256 0e1a5f0d54a44e7e0883145569668ba4e5046177cd4677991e52bbe0d568bf70

See more details on using hashes here.

File details

Details for the file metaxuda-0.1.0-py3-none-any.whl.

File metadata

  • Download URL: metaxuda-0.1.0-py3-none-any.whl
  • Upload date:
  • Size: 822.8 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.13.3

File hashes

Hashes for metaxuda-0.1.0-py3-none-any.whl
Algorithm Hash digest
SHA256 209164a67552c9af4a1fcb8004b629b5b228897bf04826c61fa1fdfdf1de77f9
MD5 f92241cf61052c4a92d2bfd86148d22d
BLAKE2b-256 7ded4b1ed6551da772cfc19324ce9998d58f68b1ad7cd9d2116584ed5691b68f

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page