NIXL Python API (CUDA 13)

1.0.1 verified Fri May 01 auth: no python

Python bindings for the NIXL library, providing GPU-accelerated operations on CUDA 13. Current version: 1.0.1. Release cadence: initial release.

pip install nixl-cu13

Common errors

error ModuleNotFoundError: No module named 'nixl' ↓

cause Package not installed or wrong import path.

fix

Run 'pip install nixl-cu13' and use 'from nixl import NIXL'.

error ImportError: libcuda.so.1: cannot open shared object file ↓

cause CUDA drivers not installed or mismatched.

fix

Install NVIDIA driver 545+ and ensure LD_LIBRARY_PATH includes CUDA libs.

breaking CUDA 13.x required; older CUDA versions will cause import errors. ↓

fix Install matching CuPy version: pip install cupy-cuda13x

gotcha Import path is 'from nixl import NIXL', not 'from nixl-cu13 import ...'. The PyPI package name and import module differ. ↓

fix Use 'from nixl import NIXL'.

NIXL
wrong
```
from nixl_cu13 import NIXL
```
correct
```
from nixl import NIXL
```
Module name is 'nixl', not 'nixl_cu13'

Initialize NIXL with default CUDA device and print version.

from nixl import NIXL
import cupy as cp

device = cp.cuda.Device(0)
with device:
    nixl = NIXL()
    print(nixl.version())