Perf: Don't allocate as much in Decoder::string #168
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
26 tasks
Jasper-Bekkers
approved these changes
Oct 7, 2020
| .position(|&c| c == 0) | ||
| .ok_or_else(|| match self.limit { | ||
| Some(_) => Error::LimitReached(self.offset + slice.len()), | ||
| None => Error::StreamExpected(self.offset), |
There was a problem hiding this comment.
Shouldn't this also be self.offset + slice.len() since previously this would only happen scanning one of the last words in the slice?
There was a problem hiding this comment.
No, tests assert this behavior. I don't know why, but, they do.
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Outdated suggestions cannot be applied.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.
On a sample module, time taken just in the Decoder::string method was 15% of the entire parse time, a significant majority of which is in the
extendcall reallocating over and over. This is obviously a hot spot, so it should be optimized a little more.So, instead of consuming words and building up a buffer, do a scan over the bytes, extract the string, and advance the proper length. This does only one allocation, the
.to_stringat the end, as opposed to the log(n) allocations before.This new method is only 5% of the entire parse time. Parse time for the module went from
14.866765msto12.630514ms, a 15% improvement. 1% of that 5% time is scanning for the null byte, another 1% for utf8 validation, and 3% for theto_stringallocation/copy.(Unfortunately there is no std function for null byte scan and utf8 validation at the same time, which would speed this up even more 😢. Also, there's an optimized
memchrfunction in std, but frustratingly, it's not exposed anywhere - it's so close right here, but they discard the position :( https://github.com/rust-lang/rust/blob/efbaa413061c2a6e52f06f00a60ee7830fcf3ea5/library/core/src/slice/cmp.rs#L273)