Worksharing-For Optimizations #94
Merged
Conversation
This file contains 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
computablee
added
bug
Codecov ReportAttention:
Additional details and impacted files@@ Coverage Diff @@
## main #94 +/- ##
==========================================
- Coverage 99.54% 99.44% -0.10%
==========================================
Files 12 12
Lines 1092 1080 -12
Branches 101 103 +2
==========================================
- Hits 1087 1074 -13
- Partials 5 6 +1
☔ View full report in Codecov by Sentry. |
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Outdated suggestions cannot be applied.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.
Which issue are you addressing?
Closes #14 and #46. Makes progress towards #16. Fixes an erroneous
Ordered_workstest. Fixed an issue with chunk size inference onstaticscheduling.How have you addressed the issue?
I have implemented a lock-free dynamic scheduler which drastically outperforms the prior scheduler when the chunk size is set to be very small. Additionally, I have applied AggressiveInlining attributes around the core scheduler code, making small improvements to performance here and there. These inlining attributes improve performance of all 3 primary schedulers, as well as some of the code for calculating collapse indices. A lock-free guided scheduler was not implemented after testing, because the amount of lock contentions was decided to be minuscule compared to dynamic, and a lock free scheduler was not likely to make much of a difference, if at all.
How have you tested your patch?
The HeatTransfer benchmark was used on my personal system with an i5-8400T CPU. Proper benchmarking protocols were used, and the results were analyzed constantly throughout optimization. For the non-optimization stuff, unit tests were altered and tested. All unit tests currently pass.