perf(katana): optimisations + benchmark setup

[remybar]

Description

WIP

Related issue

#2719

Tests

• Yes
• No, because they aren't needed
• No, because I need help

Added to documentation?

• README.md
• Dojo Book
• No documentation needed

Checklist

• I've formatted my code (scripts/prettier.sh, scripts/rust_fmt.sh, scripts/cairo_fmt.sh)
• I've linted my code (scripts/clippy.sh, scripts/docs.sh)
• I've commented my code
• I've requested a review after addressing the comments

[kariy]

This benchmark includes data setup within the measured function (ie Uncommitted::commit). We should move the preparation of the test data (all the arguments for Uncommitted::new) outside the benchmark routine to ensure we're measuring only the commit operation's performance, not including the setup overhead.

We can use the Arbitrary trait to derive random values for the test data.

I'm not sure what's the best block composition but maybe we can start with something like 20 transactions, each with 2 events, and a state update of size 100.

[kariy]

I think what we should do, is have two .commit methods - the serialized and parallelized versions (ie .commit and .commit_parallel) - run two benchmarks against the same test vectors using each method.

[remybar]

Hi @kariy !

So, I've kept 2 versions of optimized functions (original + parallel one) to be able to create commit and commit_parallel benches.

I used Arbitrary trait to generate random values (not sure if I used the idiomatic way to do this but ...) and set vector sizes to 20 for txs and receipts and 100 for state update stuff.

Maybe it would be a good idea to have 2 sets of data for these benches:

• a first one with a tiny block to be sure that parallelizing block processings do not add any overhead,
• a second one with a quite big block, as you proposed, to check that the performance improvement is worth it.

WDYT ?

[codecov]

Codecov Report

Attention: Patch coverage is 1.51515% with 65 lines in your changes missing coverage. Please review.

Project coverage is 55.79%. Comparing base (8c484fd) to head (347b9b0).

Files with missing lines	Patch %	Lines
crates/katana/core/src/backend/mod.rs	1.51%	65 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #2900      +/-   ##
==========================================
- Coverage   55.86%   55.79%   -0.07%     
==========================================
  Files         447      447              
  Lines       57859    57924      +65     
==========================================
- Hits        32323    32319       -4     
- Misses      25536    25605      +69     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[kariy]

Awesome work @remybar! It's looking good. Can we also include benchmark for block with only 1 tx and very small state updates - to mimic a small block (eg on instant mining when there's only 1 tx in a block)

[kariy]

Suggested change

	criterion_group!(benches, commit_benchmark);
	criterion_group! {
	name = benches;
	config = Criterion::default().with_profiler(PProfProfiler::new(100, Output::Flamegraph(None)));
	targets = commit_benchmark
	}

[kariy]

nit:

Suggested change

	c.bench_function("commit", |b| {
	c.bench_function("Commit.Big.Serial", |b| {

[kariy]

nit:

Suggested change

	c.bench_function("commit_parallel", |b| {
	c.bench_function("Commit.Big.Parallel", |b| {