[Memtrie] Optimize NibbleSlice during the memtrie construction code path #11342
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
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## master #11342 +/- ##
==========================================
+ Coverage 71.08% 71.09% +0.01%
==========================================
Files 783 784 +1
Lines 156875 156945 +70
Branches 156875 156945 +70
==========================================
+ Hits 111508 111580 +72
- Misses 40528 40539 +11
+ Partials 4839 4826 -13
Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Sentry. |
tayfunelmas
reviewed
May 18, 2024
tayfunelmas
approved these changes
May 20, 2024
nagisa
reviewed
May 21, 2024
| use super::node::MemTrieNodeId; | ||
| use crate::trie::mem::node::InputMemTrieNode; | ||
| use crate::NibbleSlice; | ||
| use elastic_array::ElasticArray4; |
There was a problem hiding this comment.
This looks like an iffy crate to use compared to the much better established smallvec: the documentation is pretty light, the maintenance status is unclear and its just plain a less popular crate.
Though I guess this is a preexisting dependency, so I'll file an issue to address this instead...
marcelo-gonzalez
pushed a commit
to marcelo-gonzalez/nearcore
that referenced
this pull request
May 23, 2024
…ath (near#11342) This contains two top optimizations that speed up the memtrie construction code: * Rather than allocating a new Vec<u8> every time we encode nibbles, we use a freelist so that previously allocated Vec<u8> can be reused. This reduces allocations from O(millions) to O(tens). * As a side effect, InputMemTrieNode has to be changed to contain borrowed slices, as we can no longer give owned Box<[u8]> to it. This is totally fine though because anyway the InputMemTrieNode's data needs to be copied into the constructed memtrie nodes in the arena. * Optimize the nibble slice encoding function so that rather than encoding nibble by nibble, we encode the first nibble and then just copy the remaining bytes directly. Also optimize the at(i) function, removing branching entirely (this code now gets compiled to a bunch of arithmetic operations). On shard 5 these optimizations reduce load time from 146s to 122s.
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 contains two top optimizations that speed up the memtrie construction code:
On shard 5 these optimizations reduce load time from 146s to 122s.