[EN Performance] Optimize MTrie reading single path by adding Ledger.GetSingleValue() to boost speed and reduce memory use #2473
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
- Added Ledger.GetSingleValue() to improve speed, alloc/op, and allocs/op - Modified delta.View.readFunc to use Ledger.GetSingleValue() - Optimized existing mtrie batch read for single path Bench comparision between current Ledger.Get() vs Ledger.GetSingleValue() Benchstat results: name old time/op new time/op delta LedgerGetOneValue/batch_get-4 6.54µs ± 0% 5.24µs ± 0% -19.92% (p=0.000 n=9+10) name old alloc/op new alloc/op delta LedgerGetOneValue/batch_get-4 1.74kB ± 0% 1.62kB ± 0% -6.85% (p=0.000 n=10+10) name old allocs/op new allocs/op delta LedgerGetOneValue/batch_get-4 21.0 ± 0% 15.0 ± 0% -28.57% (p=0.000 n=10+10) Bench comparision between optimized new Ledger.Get() vs Ledger.GetSingleValue() Benchstat results: name old time/op new time/op delta LedgerGetOneValue/batch_get-4 5.70µs ± 0% 5.23µs ± 0% -8.24% (p=0.000 n=9+10) name old alloc/op new alloc/op delta LedgerGetOneValue/batch_get-4 1.69kB ± 0% 1.62kB ± 0% -3.79% (p=0.000 n=10+10) name old allocs/op new allocs/op delta LedgerGetOneValue/batch_get-4 18.0 ± 0% 15.0 ± 0% -16.67% (p=0.000 n=10+10)
fxamacker
requested review from
ramtinms,
m4ksio and
AlexHentschel
as code owners
fxamacker
changed the title
Codecov Report
@@ Coverage Diff @@
## master #2473 +/- ##
==========================================
- Coverage 56.94% 56.93% -0.02%
==========================================
Files 653 651 -2
Lines 60019 59979 -40
==========================================
- Hits 34180 34149 -31
+ Misses 22913 22897 -16
- Partials 2926 2933 +7
Flags with carried forward coverage won't be shown. Click here to find out more.
Continue to review full report at Codecov.
|
ramtinms
reviewed
May 24, 2022
ramtinms
approved these changes
May 24, 2022
|
we might actually add a single empty line to the FVM code to trigger the FVM benchmarking. |
Change Forest.Read to return []ledger.Value without deep copying payload keys. This avoids 4 heap allocation per key. This change doesn't affect the caller (Ledger.Get) because it discards the payload keys. name old time/op new time/op delta TrieRead-4 524µs ± 1% 420µs ± 1% -19.77% (p=0.000 n=10+10) name old alloc/op new alloc/op delta TrieRead-4 190kB ± 0% 95kB ± 0% -50.04% (p=0.000 n=10+10) name old allocs/op new allocs/op delta TrieRead-4 1.52k ± 0% 0.32k ± 0% -79.17% (p=0.000 n=10+10)
Changed Forest.ReadSingleValue to return ledger.Value without deep copying payload keys. This avoid 4 heap allocation per key. This change doesn't affect the caller (Ledger.GetSingleValue) because it discards the payload key.
tarakby
approved these changes
May 25, 2022
ledger/complete/mtrie/trie/trie.go
Outdated
Comment on lines
208
to
216
| for !head.IsLeaf() { | ||
| depth := ledger.NodeMaxHeight - head.Height() // distance to the tree root | ||
| bit := bitutils.ReadBit(pathBytes, depth) | ||
| if bit == 0 { | ||
| head = head.LeftChild() | ||
| } else { | ||
| head = head.RightChild() | ||
| } | ||
| } |
There was a problem hiding this comment.
maybe sliiightly faster :)
Suggested change
| for !head.IsLeaf() { | |
| depth := ledger.NodeMaxHeight - head.Height() // distance to the tree root | |
| bit := bitutils.ReadBit(pathBytes, depth) | |
| if bit == 0 { | |
| head = head.LeftChild() | |
| } else { | |
| head = head.RightChild() | |
| } | |
| } | |
| depth := ledger.NodeMaxHeight - head.Height() // distance to the tree root | |
| for !head.IsLeaf() { | |
| bit := bitutils.ReadBit(pathBytes, depth) | |
| if bit == 0 { | |
| head = head.LeftChild() | |
| } else { | |
| head = head.RightChild() | |
| } | |
| depth++ | |
| } |
This optimization improves speed by 0.49% on Haswell CPU. Thanks for suggestion @tarakby. Co-authored-by: Tarak Ben Youssef <[email protected]>
…llocs Optimize Ledger.Get() by making Forest.Read() use ~5x fewer allocs/op and run ~20% faster
fxamacker
changed the title
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.
Changes
When reading a payload from a single path, avoid needlessly path partitioning and recursive calls. Also, optimize MTrie batch read.
Closes #1876
Benchmark comparisions
Current Ledger.Get() vs Ledger.GetSingleValue()
Optimized new Ledger.Get() vs Ledger.GetSingleValue()