util/interval/generic: avoid new allocations in btree benchmarks#44617
Merged
craig[bot] merged 1 commit intocockroachdb:masterfrom Feb 3, 2020
Merged
Conversation
Member
|
This change is |
petermattis
approved these changes
Feb 1, 2020
Collaborator
petermattis
left a comment
There was a problem hiding this comment.
Reviewable status:
complete! 1 of 0 LGTMs obtained (waiting on @sumeerbhola)
sumeerbhola
approved these changes
Feb 2, 2020
Collaborator
sumeerbhola
left a comment
There was a problem hiding this comment.
Reviewable status:
The generic template for interval btree makes an attempt to work with both value and pointer parameterized types. To facilitate this option in tests, it used reflection to create new instances of the item type. This reflection was preventing allocations of the item type from being inlined and avoided by escape analysis, which was skewing benchmarks. This commit fixes this by avoiding the reflection and instead adding a constructor to the contract of the parameterized type. This is cleaner and allows the allocations to be avoided in these benchmarks.
fb6c093 to
f80ed60
Compare
Contributor
Author
|
bors r+ |
craig bot
pushed a commit
that referenced
this pull request
Feb 3, 2020
44617: util/interval/generic: avoid new allocations in btree benchmarks r=nvanbenschoten a=nvanbenschoten The generic template for interval btree makes an attempt to work with both value and pointer parameterized types. To facilitate this option in tests, it used reflection to create new instances of the item type. This reflection was preventing allocations of the item type from being inlined and avoided by escape analysis, which was skewing benchmarks. This commit fixes this by avoiding the reflection and instead adding a constructor to the contract of the parameterized type. This is cleaner and allows the allocations to be avoided in these benchmarks. Here's the new `benchdiff` output of the latch manager's btree from before #43850 to after this change: [benchdiff sheet](https://docs.google.com/spreadsheets/d/1ZxtPPSWV0z76msCiwIyqL0xnUQF530ZwnbnfrvEzNAE/edit#gid=4). Co-authored-by: Nathan VanBenschoten <nvanbenschoten@gmail.com>
Contributor
Build succeeded |
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
4 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.
The generic template for interval btree makes an attempt to work with
both value and pointer parameterized types. To facilitate this option
in tests, it used reflection to create new instances of the item type.
This reflection was preventing allocations of the item type from being
inlined and avoided by escape analysis, which was skewing benchmarks.
This commit fixes this by avoiding the reflection and instead adding
a constructor to the contract of the parameterized type. This is cleaner
and allows the allocations to be avoided in these benchmarks.
Here's the new
benchdiffoutput of the latch manager's btree frombefore #43850 to after this change: benchdiff sheet.