# DTolm/VkFFT

Vulkan/CUDA/HIP/OpenCL/Level Zero/Metal Fast Fourier Transform library

Repository: https://github.com/DTolm/VkFFT
Canonical: https://ross.abutalabs.com/products/vkfft
Language: C++
License: MIT
License Family: permissive
Topics: vulkan, fft, convolution, r2c, c2r, hpc, cuda, hip, opencl, r2r, dct, levelzero, metal
Last push: 2026-04-04T19:41:08+00:00

## Health v2 (maintenance only)
Score: 57/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 75, release rhythm 8, longevity 100
- inputs: {"age_days": 2222, "days_push": 151, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1769, forks 142 (observed 2026-08-28T04:05:33.742479+00:00)

## What it is
VkFFT is an open-source, GPU-accelerated multidimensional Fast Fourier Transform library supporting Vulkan, CUDA, HIP, OpenCL, Level Zero, and Metal backends. It aims to match or exceed cuFFT performance while remaining vendor-neutral and MIT-licensed.

## Use cases
- gpu-accelerated fft library as cuFFT alternative
- compute ffts on vulkan compute shaders
- fast convolution on gpu for image processing
- real-to-complex and dct transforms on gpu
- cross-vendor gpu fft for hip and opencl
- large multidimensional fft for hpc simulations

## When to choose
- you need FFTs across multiple GPU vendors, not just NVIDIA
- you want an open-source cuFFT replacement with competitive performance
- you need R2C/C2R/R2R transforms, convolutions, or zero padding fused into the FFT
- you work in Vulkan, HIP, OpenCL, Level Zero, or Metal compute

## When to avoid
- you only need CPU-based FFTs (use FFTW or similar)
- you want a high-level turnkey application rather than a C/C++ library to integrate
- your project is locked into a vendor library ecosystem and migration cost outweighs gains

## Facets
- artifact type: library
- maturity: active
- function: math, gpu-computing, image-processing, benchmarking
- domain: gpu-computing, performance, graphics, simulation
- platform: cpp, cross-platform, windows
- tags: fft, vulkan, cuda, hip, opencl, metal, level-zero, convolution, dct, hpc, signal-processing, algorithms, gpu, linux, macos

## Member repositories
- DTolm/VkFFT (main) score 57

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:33.742479+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T03:26:04.905842+00:00, confidence not recorded.
  - readme: https://github.com/DTolm/VkFFT (fetched 2026-08-28T04:05:33.742479+00:00, sha 3f11799890fa)
- Data as of 2026-08-30T08:39:29.467469+00:00.