Async task parallelization #6
Merged
+86
−62
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
Pull Request Test Coverage Report for Build 19540738427Details
💛 - Coveralls |
aeddins-ibm
reviewed
Nov 19, 2025
Contributor
aeddins-ibm
left a comment
aeddins-ibm
left a comment
There was a problem hiding this comment.
Great to see the speedup in those lightcone plots at the top of the PR!
Commenting even though this is WIP. I'm suspicious of the use of nonlocal seemingly to share write access to an object across processes, but OTOH it's apparently working so far...
Co-authored-by: aeddins-ibm <[email protected]>
Member
Author
|
For reerence purposes, this is what the logging output now looks like: |
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Outdated suggestions cannot be applied.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.
This refactors the parallelization of the unequal time commutator bound computation.
The parallelization is still done using a
multiprocessing.Poolwith a user-configurable number of processes and maximum computation timeout.However, prior to this PR the code was processing the circuit layer-by-layer, where the bound computation of all noise terms appearing in a single layer have to complete before moving on to the next one. This can hinder performance when only a few of these terms take a long time to compute during which time the remanining processes end up idling.
This PR updates the implementation to submit the computation of all noise term bounds across all layers as asynchronous tasks to be processed in parallel. This avoids the bottleneck above as pending tasks can continue to be processed even when long running ones block single resources.
This allowed a tets computation of the forward bounds on a mirrored time evolution using 20 Trotter steps on a 40 qubit kicked Ising Hamiltonian to complete entirely within 55 minutes on 96 parallel processes while the old implementation timed out after processing 54 of the 80 layers.
Before this PR can be merged, I still need to address the following:
(opt.) add some unittesting of the timeout behavior