avoid contending on single process-wide precompiled regexes #37
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
precompiling the `Regex` instances is good! but being behind `lazy_static!` means all threads are contending an implicit futex somewhere in the lazy_static implementation. on a 32-core system with 32 logs to process in parallel, this ends up with the top symbol in `perf stat` to be something in `futex` and an awful lot of total time waiting on _system stuff_ for no apparent reason. having set of Regex bundles per-thread keeps contention down at the cost of a bit more init on the multi-processor case, but overall at least on my system brings a 31-file benchmark down from 4.6 seconds to about 2.2 seconds.
no reason to limit to 128k really, probably not a big deal either way
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Outdated suggestions cannot be applied.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.
precompiling the
Regexinstances is good! but being behindlazy_static!means all threads are contending an implicit futexsomewhere in the lazy_static implementation. on a 32-core system with 32
logs to process in parallel, this ends up with the top symbol in
perf statto be something infutexand an awful lot of total timewaiting on system stuff for no apparent reason.
having set of Regex bundles per-thread keeps contention down at the
cost of a bit more init on the multi-processor case, but overall at
least on my system brings a 31-file benchmark down from 4.6 seconds to
about 2.2 seconds.
i included a bonus second change to make the read buffers a bit larger but that's not nearly as useful