🦆 QuACK: A Quirky Assortment of CuTe Kernels 🦆 ❤️ PaddlePaddle

Note

This repo is a fork of the original QuACK project, with modifications to enhance compatibility and integration with PaddlePaddle. Currently branch is align with 3d0ab3ec2164749caac8f269f771e66a40efd2de

Installation

git clone https://github.com/PFCCLab/quack.git
cd quack
pip install .

Usage

import paddle
paddle.enable_compat(scope={"quack"})  # Enable torch proxy before importing quack
import quack
# use quack

The original README.md content is as follows:

---

Kernels are written in the CuTe-DSL.

Installation

pip install quack-kernels

Requirements

• H100 or B200 GPU
• CUDA toolkit 12.9+
• Python 3.12

Kernels 🐥

• 🦆 RMSNorm forward + backward
• 🦆 Softmax forward + backward
• 🦆 Cross entropy forward + backward
• 🦆 Layernorm forward
• 🦆 Hopper gemm + epilogue
• 🦆 Blackwell gemm + epilogue

Usage

from quack import rmsnorm, softmax, cross_entropy

Documentations

[2025-07-10] We have a comprehensive blogpost on how to get memory-bound kernels to speed-of-light, right in the comfort of Python thanks to the CuTe-DSL.

Performance

See our blogpost for the details.

Development

To set up the development environment:

pip install -e '.[dev]'
pre-commit install