Add PrecompileTools workload to improve TTFX #134
Merged
Conversation
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
Add precompilation for commonly used code paths to significantly reduce time-to-first-execution (TTFX) for key operations: - POD construction (matrix and vector-of-vectors input) - reduce! with SVD, TSVD, and RSVD algorithms - deim_interpolation_indices (core DEIM algorithm) Performance improvements (measured TTFX): - reduce!(pod, SVD()): 0.51s → 0.09s (82% reduction) - reduce!(pod, TSVD()): 2.02s → 0.03s (99% reduction) - reduce!(pod, RSVD()): 0.23s → 0.00s (99% reduction) - deim_interpolation_indices: 1.63s → 0.04s (98% reduction) - Total key TTFX: 4.4s → 0.16s (96% reduction) No invalidations were introduced by ModelOrderReduction itself. The existing invalidations come from upstream dependencies in the SciML ecosystem. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <[email protected]>
This was referenced Jan 2, 2026
Closed
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
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Outdated suggestions cannot be applied.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.
Summary
Performance Improvements
reduce!(pod, SVD())reduce!(pod, TSVD())reduce!(pod, RSVD())deim_interpolation_indicesPackage load time remains similar (~8-9s), as expected since it's dominated by dependencies (ModelingToolkit, Symbolics, etc.).
What was added to precompilation workload
The precompile.jl file covers:
reduce!with all three SVD algorithm variants (SVD, TSVD, RSVD)deim_interpolation_indices(core DEIM algorithm)Invalidation Analysis
No invalidations were introduced by ModelOrderReduction itself. The 348 invalidation trees observed come from upstream dependencies in the SciML ecosystem (SymbolicUtils, StaticArrays, MultivariatePolynomials, etc.), which is a known ecosystem-wide issue.
Test plan
cc @ChrisRackauckas
🤖 Generated with Claude Code