first-kernel

A custom kernel for PyTorch.

Installation

pip install drbh/first-kernel

Usage

import torch
from first_kernel import first_kernel

# Create input tensor
x = torch.randn(1024, 1024, device="cuda")

# Run kernel
result = first_kernel(x)

Development

Building

nix develop
nix run .#build-and-copy

Testing

nix develop .#test
pytest tests/

Test as a kernels user

uv run example.py

License

Apache 2.0

Downloads last month
-
Supported hardwares new
Metal
Apple Silicon
Apple MacBook Neo
8GB
Apple Silicon
Apple M1
8GB
Apple Silicon Pro
Apple M1 Pro
16GB
Apple Silicon Max
Apple M1 Max
16GB
Apple Silicon Ultra
Apple M1 Ultra
16GB
Apple Silicon
Apple M2
8GB
Apple Silicon Pro
Apple M2 Pro
16GB
Apple Silicon Max
Apple M2 Max
32GB
Apple Silicon Ultra
Apple M2 Ultra
64GB
Apple Silicon
Apple M3
8GB
Apple Silicon Pro
Apple M3 Pro
18GB
Apple Silicon Max
Apple M3 Max
36GB
Apple Silicon Ultra
Apple M3 Ultra
96GB
Apple Silicon
Apple M4
16GB
Apple Silicon Pro
Apple M4 Pro
24GB
Apple Silicon Max
Apple M4 Max
36GB
Apple Silicon
Apple M5
16GB
Apple Silicon Pro
Apple M5 Pro
24GB
Apple Silicon Max
Apple M5 Max
36GB
Torch
2.10
OS
macos
Arch
aarch64