{"id":23067,"library":"distributed-ucxx-cu12","title":"distributed-ucxx-cu12","description":"UCX communication module for Dask Distributed, providing high-performance GPU-aware communication using UCX (Unified Communication X). This is the CUDA 12 variant of distributed-ucxx. Version 0.49.0 requires Python >=3.11 and CUDA 12.2+. Released as part of RAPIDS, with monthly release cadence.","status":"active","version":"0.49.0","language":"python","source_language":"en","source_url":"https://github.com/rapidsai/ucxx","tags":["dask","distributed","ucx","high-performance-computing","gpu","rapids","cuda12"],"install":[{"cmd":"pip install distributed-ucxx-cu12","lang":"bash","label":"Install from PyPI"},{"cmd":"conda install -c rapidsai -c conda-forge -c nvidia distributed-ucxx-cu12","lang":"bash","label":"Install via conda"}],"dependencies":[{"reason":"Replaces numba-cuda runtime dependency starting v0.49.0","package":"cuda-core","optional":false},{"reason":"Core UCX Python bindings","package":"ucx-py","optional":false},{"reason":"Dask Distributed integration","package":"distributed","optional":false}],"imports":[{"note":"","wrong":"","symbol":"UCX","correct":"from distributed_ucxx import UCX"},{"note":"","wrong":"","symbol":"ucx_create_endpoint","correct":"from distributed_ucxx import ucx_create_endpoint"}],"quickstart":{"code":"from distributed_ucxx import UCX\nimport asyncio\n\nasync def main():\n    ucx = UCX()\n    await ucx.start()\n    print(\"UCX started successfully\")\n    await ucx.close()\n\nasyncio.run(main())","lang":"python","description":"Minimal example to start UCX communication using distributed-ucxx-cu12."},"warnings":[{"fix":"Update imports and ensure cuda-core is installed. Remove direct numba-cuda dependency if not explicitly needed.","message":"v0.49.0 replaces numba-cuda runtime dependency with cuda-core. Existing code relying on numba types may break.","severity":"breaking","affected_versions":">=0.49.0"},{"fix":"Upgrade to CUDA 12.2 or later, or use an older version of distributed-ucxx-cu12 (e.g., 0.46.0).","message":"v0.47.0 requires CUDA 12.2+. Older CUDA versions are no longer supported.","severity":"breaking","affected_versions":">=0.47.0"},{"fix":"Use `from distributed_ucxx import ...`","message":"The import path is `distributed_ucxx`, not `distributed-ucxx`. Hyphens are not valid in Python module names.","severity":"gotcha","affected_versions":"all"},{"fix":"Use the modern async context manager pattern.","message":"Some older APIs like `ucx_init` are deprecated in favor of `UCX.start()`. Check the documentation for updated patterns.","severity":"deprecated","affected_versions":">=0.46.0"}],"env_vars":null,"last_verified":"2026-05-01T00:00:00.000Z","next_check":"2026-07-30T00:00:00.000Z","problems":[{"fix":"Install with 'pip install distributed-ucxx-cu12' and ensure you are using Python >=3.11.","cause":"Tried to import 'distributed_ucxx' but package is not installed or wrong environment","error":"ModuleNotFoundError: No module named 'distributed_ucxx'"},{"fix":"Install UCX via conda: 'conda install -c conda-forge ucx' or use the RAPIDS conda environment.","cause":"UCX system library not installed or not found in LD_LIBRARY_PATH","error":"ImportError: libucp.so.0: cannot open shared object file: No such file or directory"},{"fix":"Upgrade CUDA toolkit to 12.2+ or use a compatible version like distributed-ucxx-cu12==0.46.0.","cause":"Installed distributed-ucxx-cu12 requires CUDA 12.2+ but system has older CUDA","error":"RuntimeError: CUDA version mismatch: expecting 12.2, got 12.1"}],"ecosystem":"pypi","meta_description":null,"install_score":null,"install_tag":null,"quickstart_score":null,"quickstart_tag":null}