Conversation
Jammy2211
changed the title
Contributor
There was a problem hiding this comment.
Pull request overview
This PR exposes a new utility function from the PyAutoArray dependency to enable JAX/GPU-accelerated W-Tilde curvature preload computation for interferometry data. The change enables significant performance improvements (100x+ speedup) for high-resolution ALMA datasets by allowing the expensive curvature matrix precomputation to run on GPU instead of CPU.
Changes:
- Added import for
load_curvature_preload_if_compatiblefunction from autoarray's interferometer w_tilde module
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The fast interferometry JAX GPU implementation (Jammy2211/PyAutoArray#201) uses a preload of the NUFFT in order to compute the
curvature_matrix.This
curvature_preloadmatrix is computed once, before lens modeling, and reused throughout lens modeling. For high number of visibilities and resolution real space mask, this calculation can take minutes or hours on a CPU.The previous PR did not convert this calculation to JAX or run on GPU.
This pull request makes the W-Tilde curvature preload computation support JAX and GPU, with profiling suggeting at least x100 speed up for high resolution datasets, with the calculation taking under a minute for the highest resolution / visibilities ALMA datasets tested.
It also includes utilities for safely saving and loading precomputed data, which check metadata to ensure the loaded data matches the data being analysed.