perf(libs/json): Lower heap overhead of JSON encoding (backport #2846… #29
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
…bft#2846) (cometbft#2876) --- Many RPC methods require JSON marshalled responses. We saw this taking a notable amount of heap allocation in query serving full nodes. This PR removes some extra heap allocations that were being done. We avoided using the more efficient encoder.Encode before, because it added a newline. This PR changes the function signature for these private methods to be using *bytes.Buffer, and then uses the in-buffer methods (rather than a second copy). We then just truncate the final byte after each such call, which does not waste any allocations. I added a benchmark for the most complex test case. OLD: ``` BenchmarkJsonMarshalStruct-12 78992 15542 ns/op 4487 B/op 191 allocs/op ``` New: ``` BenchmarkJsonMarshalStruct-12 93346 11132 ns/op 3245 B/op 58 allocs/op ``` Roughly a 3-4x reduction in the number of allocations, and 20% speedup. - [x] Tests written/updated - Existing tests cover this - [x] Changelog entry added in `.changelog` (we use [unclog](https://github.com/informalsystems/unclog) to manage our changelog) - [x] Updated relevant documentation (`docs/` or `spec/`) and code comments - [x] Title follows the [Conventional Commits](https://www.conventionalcommits.org/en/v1.0.0/) spec <hr>This is an automatic backport of pull request cometbft#2846 done by [Mergify](https://mergify.com). --------- Co-authored-by: Dev Ojha <[email protected]> Co-authored-by: Andy Nogueira <[email protected]>
mergify bot
pushed a commit
that referenced
this pull request
Apr 30, 2024
…bft#2846… (#29) * perf(libs/json): Lower heap overhead of JSON encoding (backport cometbft#2846) (cometbft#2876) --- Many RPC methods require JSON marshalled responses. We saw this taking a notable amount of heap allocation in query serving full nodes. This PR removes some extra heap allocations that were being done. We avoided using the more efficient encoder.Encode before, because it added a newline. This PR changes the function signature for these private methods to be using *bytes.Buffer, and then uses the in-buffer methods (rather than a second copy). We then just truncate the final byte after each such call, which does not waste any allocations. I added a benchmark for the most complex test case. OLD: ``` BenchmarkJsonMarshalStruct-12 78992 15542 ns/op 4487 B/op 191 allocs/op ``` New: ``` BenchmarkJsonMarshalStruct-12 93346 11132 ns/op 3245 B/op 58 allocs/op ``` Roughly a 3-4x reduction in the number of allocations, and 20% speedup. - [x] Tests written/updated - Existing tests cover this - [x] Changelog entry added in `.changelog` (we use [unclog](https://github.com/informalsystems/unclog) to manage our changelog) - [x] Updated relevant documentation (`docs/` or `spec/`) and code comments - [x] Title follows the [Conventional Commits](https://www.conventionalcommits.org/en/v1.0.0/) spec <hr>This is an automatic backport of pull request cometbft#2846 done by [Mergify](https://mergify.com). --------- Co-authored-by: Dev Ojha <[email protected]> Co-authored-by: Andy Nogueira <[email protected]> * changelog --------- Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com> Co-authored-by: Dev Ojha <[email protected]> Co-authored-by: Andy Nogueira <[email protected]> (cherry picked from commit ee66963)
7 tasks
czarcas7ic
added a commit
that referenced
this pull request
Apr 30, 2024
…bft#2846… (#29) (#33) * perf(libs/json): Lower heap overhead of JSON encoding (backport cometbft#2846) (cometbft#2876) --- Many RPC methods require JSON marshalled responses. We saw this taking a notable amount of heap allocation in query serving full nodes. This PR removes some extra heap allocations that were being done. We avoided using the more efficient encoder.Encode before, because it added a newline. This PR changes the function signature for these private methods to be using *bytes.Buffer, and then uses the in-buffer methods (rather than a second copy). We then just truncate the final byte after each such call, which does not waste any allocations. I added a benchmark for the most complex test case. OLD: ``` BenchmarkJsonMarshalStruct-12 78992 15542 ns/op 4487 B/op 191 allocs/op ``` New: ``` BenchmarkJsonMarshalStruct-12 93346 11132 ns/op 3245 B/op 58 allocs/op ``` Roughly a 3-4x reduction in the number of allocations, and 20% speedup. - [x] Tests written/updated - Existing tests cover this - [x] Changelog entry added in `.changelog` (we use [unclog](https://github.com/informalsystems/unclog) to manage our changelog) - [x] Updated relevant documentation (`docs/` or `spec/`) and code comments - [x] Title follows the [Conventional Commits](https://www.conventionalcommits.org/en/v1.0.0/) spec <hr>This is an automatic backport of pull request cometbft#2846 done by [Mergify](https://mergify.com). --------- Co-authored-by: Dev Ojha <[email protected]> Co-authored-by: Andy Nogueira <[email protected]> * changelog --------- Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com> Co-authored-by: Dev Ojha <[email protected]> Co-authored-by: Andy Nogueira <[email protected]> (cherry picked from commit ee66963) Co-authored-by: Adam Tucker <[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.
…) (cometbft#2876)
Many RPC methods require JSON marshalled responses. We saw this taking a notable amount of heap allocation in query serving full nodes. This PR removes some extra heap allocations that were being done. We avoided using the more efficient encoder.Encode before, because it added a newline. This PR changes the function signature for these private methods to be using *bytes.Buffer, and then uses the in-buffer methods (rather than a second copy). We then just truncate the final byte after each such call, which does not waste any allocations.
I added a benchmark for the most complex test case.
OLD:
New:
Roughly a 3-4x reduction in the number of allocations, and 20% speedup.
.changelog(we use unclog to manage our changelog)docs/orspec/) and code commentsThis is an automatic backport of pull request perf(libs/json): Lower heap overhead of JSON encoding cometbft/cometbft#2846 done by Mergify.
PR checklist
.changelog(we use unclog to manage our changelog)docs/orspec/) and code comments