Skip to content

Conversation

@cliffburdick
Copy link
Collaborator

This PR introduces generic JIT compilation with cuFFTDx as the first use case. cuFFTDx is a compile-time device FFT library that can achieve large speedups over cuFFT for certain situations. This PR introduces 3 main changes:

  1. The use of Jitify2 and NVRTC for JIT compilation
  2. A capabilities API to retrieve capabilities of an operator and its nested operators
  3. cuFFTDx and libmathdx libraries

JIT/cuFFTDx support is considered experimental at this point. This PR is considered an early draft until the debug is removed, but it is functional and passes all unit tests.

@copy-pr-bot
Copy link

copy-pr-bot bot commented Jul 25, 2025

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

@cliffburdick
Copy link
Collaborator Author

/build

@cliffburdick
Copy link
Collaborator Author

/build

@cliffburdick cliffburdick deleted the cufftdx branch January 9, 2026 23:06
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant