Fix slowdown when using periodic_save_freq in MCMC - #436
Draft
sblunt wants to merge 2 commits into
Draft
Conversation
periodic_save_freq previously re-derived the entire chain history (including an unvectorized per-sample prior recomputation) and rewrote the whole HDF5 output file on every periodic save, making cost grow with total run length instead of just the new chunk. _update_chains_from_sampler now only processes steps since the last save, and Results.save_results appends to resizable HDF5 datasets instead of rewriting the file each time. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
… saves Results.add_samples() used np.vstack/np.append, which reallocate and copy the entire accumulated post/lnlike arrays on every call. This meant that even after the earlier fix to periodic_save_freq (only recomputing/writing new chunks), the in-memory accumulation itself was still O(total rows so far) per periodic save, making the MCMC fit progressively slower as it ran. add_samples now writes into a capacity-doubling internal buffer, so appends are amortized O(1) per row instead of O(n) per call. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
periodic_save_freq previously re-derived the entire chain history (including an unvectorized per-sample prior recomputation) and rewrote the whole HDF5 output file on every periodic save, making cost grow with total run length instead of just the new chunk. _update_chains_from_sampler now only processes steps since the last save, and Results.save_results appends to resizable HDF5 datasets instead of rewriting the file each time.
(experimenting with claude)