Repository navigation
test: measure a warm call in the update field allocation tests - #586
Merged
Merged
Conversation
These three measured the first call of the code under test. JIT and type loading now and then land on a runtime table that has to grow, and that counts against the thread: 3240, 49200 and 98352 bytes in failing runs, and the two big ones are a doubling. Every other allocation test warms up first. They now call the action once before measuring. A boxed index still fails all three, as it costs every call. With two suite loops at once, they failed 3 runs of 120 before and none of 200 after. Fixes #584 Co-Authored-By: Claude Opus 5.5 <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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Fixes #584
UpdateFieldsArrayAllocationTestsfailed now and then in full runs, in any of its three tests.They measured the first call of the code under test, and were the only allocation tests that did. Failing runs reported 3240, 49200 and 98352 bytes. The last two are a doubling (49152 + 48, 98304 + 48), which points to a global table growing during JIT or type loading, charged to whichever thread trips it. That table is inferred from those numbers, not observed.
AllocatedBynow calls the action once before measuring, as the other allocation tests do. Test code only.Tested
The second row shows the warm-up doesn't hide what these tests are for: a box costs every call.
Ruled out on the way: forced gen0 collections from another thread didn't move the counter in 7.2M measurements.
🤖 Generated with Claude Code