FlashQuad

A user-friendly Python numerical integration library with GPU acceleration.

FlashQuad provides a single unified API for numerical integration across multiple array backends — NumPy, PyTorch, CuPy, and JAX — so you can move from CPU prototyping to GPU-accelerated computation with one line change.

Key features

  • Multiple quadrature methods — trapezoidal, Simpson’s, Boole’s, Gauss-Legendre, Monte Carlo, and adaptive Monte Carlo

  • GPU acceleration — run on CUDA via PyTorch, CuPy, or JAX with no code changes beyond swapping the backend

  • Batched parameter sweeps — evaluate integrals over a batch of parameters in a single call

  • Arbitrary dimensions — integrate over 1D, 2D, 3D, or higher-dimensional domains

  • Boundary masking — apply custom domain boundaries via mask functions

Quick example

import numpy as np
from flashquad import FlashQuad

fq = FlashQuad(backend="numpy")

result = fq.simpson(
    func=lambda x, y: np.sin(x) * np.cos(y),
    intervals=[[0, np.pi], [0, np.pi]],
    num_points=[101, 101],
)

Contents