Improve performance of hfe listpack #13279
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
sundb
reviewed
May 20, 2024
tezc
commented
May 20, 2024
|
@sundb I addressed your comments and moved lpBatchInsert() as a separate function. Let's see if we are more comfortable with this one. Maybe we should add a comment that this function might be merged with lpInsert() later in future. |
|
@tezc yeah, it's clear now. |
sundb
reviewed
May 21, 2024
Co-authored-by: debing.sun <[email protected]>
sundb
previously approved these changes
May 21, 2024
tezc
commented
May 21, 2024
sundb
previously approved these changes
May 21, 2024
sundb
approved these changes
May 22, 2024
sundb
added a commit
that referenced
this pull request
Sep 3, 2024
) In #13279 (found by @filipecosta90), for custom lookups, we introduce a comparison function for `lpFind()` to compare entry, but it also introduces some overhead. To avoid the overhead of function pointer calls: 1. Extract the lpFindCb() method into a lpFindCbInternal() method that is easier to inline. 2. Use unlikely to annotate the comparison method, as can only success once. --------- Co-authored-by: Ozan Tezcan <[email protected]>
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Outdated suggestions cannot be applied.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.
This PR contains a few optimizations for hfe listpack.
Hfe fields are ordered by TTL in the listpack. There are two cases that we want to search listpack according to TTLs:
Iterating with lpNext() to compare TTLs has a performance cost as lpNext() calls lpValidateIntegrity() for each entry. Instead, this PR adds
lpFindCb()to the listpack which accepts a comparator callback. It preserves same validation logic of lpFind() which is faster than search with lpNext().We have field name, value, ttl for a single hfe field. Inserting these items one by one to listpack is costly. Especially, as we place fields according to TTL, most additions will end up in the middle of the listpack. Each insert causes realloc + memmove. This PR introduces
lpBatchInsert()to add multiple items in one go.For hsetf, if we are going to update value and TTL at the same time, currently, we update the value first and later update the TTL (two distinct listpack operation). This PR improves it by doing it with a single update operation.