perf(codegen): reduce memory allocations in generate_line_offset_tables
#13054
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
Contributor
How to use the Graphite Merge QueueAdd either label to this PR to merge it via the merge queue:
You must have a Graphite account in order to use the merge queue. Sign up using this link. An organization admin has enabled the Graphite Merge Queue in this repository. Please do not merge from GitHub as this will restart CI on PRs being processed by the merge queue. |
github-actions
bot
added
A-codegen
CodSpeed Instrumentation Performance ReportMerging #13054 will not alter performanceComparing Summary
Footnotes |
Member
Author
Merge activity
|
graphite-app
bot
force-pushed
the
08-13-perf_codegen_reduce_memory_allocations_in_generate_line_offset_tables_
branch
from
13cb5f5 to
ab685bd
Compare
graphite-app
bot
deleted the
08-13-perf_codegen_reduce_memory_allocations_in_generate_line_offset_tables_
branch
graphite-app bot
pushed a commit
that referenced
this pull request
Aug 13, 2025
…fset_tables` (#13056) #13054 added a nice optimization to `SourcemapBuilder`. During generation of line/offset tables, it reuses a single `Vec` for column indexes for each line, rather than creating a new `Vec` on each turn of the inner loop. This reduces the number of times that `Vec` may have to grow as column indexes get added to it. Take this optimization a step further by re-using the same `Vec` across *all* lines. `columns` `Vec` is not consumed on each line, but each time the contents are copied into a boxed slice - except when reaching EOF, where we can consume `columns`, as its work is done. This memory-copying was likely happening anyway, as `Vec<u32>` -> `Box<[u32]>` conversion has to drop the spare capacity of the `Vec`, which will likely cause a reallocation. Also, avoid using iterators to create the boxed slices. `Vec::clone` followed by `Vec::into_boxed_slice` is a bit more explicit and so may help compiler to see that it only needs to allocate exactly `columns.len()` slots for the `Box<[u32]>`. Note: I also tried `columns.drain(..).collect()` instead of `columns.clone().into_boxed_slice()` + `columns.clear()`. But it looks like the `Drain` abstraction doesn't get completely removed by compiler. https://godbolt.org/z/Trv47j4hP So I *think* `into_boxed_slice` is probably preferable.
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.
No description provided.