ci: deserialize cargo bench result

[AvivYossef-starkware]

No description provided.

[reviewable-StarkWare]

This change is 

[AvivYossef-starkware]

• ci: add benchmark config #9625 : 2 dependent PRs (#9626 , #9627 )
• ci: deserialize cargo bench result #9624 👈 (View in Graphite)
• ci: add bench_tools CLI scaffold #9612
• main-v0.14.1

This stack of pull requests is managed by Graphite. Learn more about stacking.

[github-actions]

Artifacts upload workflows:

• Started at 2025-10-19T10:38:26Z
• Started at 2025-10-19T10:58:28Z
• Started at 2025-10-19T11:31:47Z
• Started at 2025-10-19T13:02:16Z
• Started at 2025-10-22T06:16:10Z
• Started at 2025-10-22T06:32:29Z
• Started at 2025-10-22T12:13:08Z
• Started at 2025-10-22T12:23:20Z
• Started at 2025-10-22T12:54:58Z
• Started at 2025-10-23T08:25:22Z

[github-actions]

Benchmark movements: No major performance changes detected.

[meship-starkware]

@meship-starkware reviewed 10 of 10 files at r1, all commit messages.
Reviewable status: all files reviewed, 1 unresolved discussion (waiting on @avi-starkware)

---

bench_tools/src/benches/dummy_bench.rs line 14 at r1 (raw file):

    c.bench_function("dummy_sum_1000", |b| b.iter(|| dummy_function(black_box(1000))));
}

Is this the function that creates the benchmark files? Where are we going to use it?
Can you add a small documentation on the purpose of this function?

Code quote:

#[allow(dead_code)]
fn dummy_benchmark(c: &mut Criterion) {
    c.bench_function("dummy_sum_100", |b| b.iter(|| dummy_function(black_box(100))));

    c.bench_function("dummy_sum_1000", |b| b.iter(|| dummy_function(black_box(1000))));
}

[avi-starkware]

bench_tools/src/benches/dummy_bench.rs line 11 at r1 (raw file):

#[allow(dead_code)]
fn dummy_benchmark(c: &mut Criterion) {
    c.bench_function("dummy_sum_100", |b| b.iter(|| dummy_function(black_box(100))));

Why is black_box needed here?
And shouldn't it be black_box(dummy_function(100))?

Additionally, criterion::black_box is deprecated (see docs) you should use std::docs::black_box instead

Code quote:

    c.bench_function("dummy_sum_100", |b| b.iter(|| dummy_function(black_box(100))));

[avi-starkware]

bench_tools/src/benches/dummy_bench.rs line 11 at r1 (raw file):

Previously, avi-starkware (Avi Cohen) wrote…

Why is black_box needed here?
And shouldn't it be black_box(dummy_function(100))?

Additionally, criterion::black_box is deprecated (see docs) you should use std::docs::black_box instead

correction: it is std::hint::black_box

[avi-starkware]

@avi-starkware reviewed 10 of 10 files at r1, all commit messages.
Reviewable status: all files reviewed, 3 unresolved discussions (waiting on @AvivYossef-starkware and @meship-starkware)

---

bench_tools/src/types/estimates_test.rs line 30 at r1 (raw file):

    // 2) Collect ALL estimates.json files under target/criterion/**/new/
    let patterns =

Why do you collect all estimates.json files and not just the ones related to dummy_benches?

Code quote:

    // 2) Collect ALL estimates.json files under target/criterion/**/new/
    let patterns =

[AvivYossef-starkware]

Reviewable status: 0 of 12 files reviewed, 3 unresolved discussions (waiting on @avi-starkware and @meship-starkware)

---

bench_tools/src/types/estimates_test.rs line 30 at r1 (raw file):

Previously, avi-starkware (Avi Cohen) wrote…

Why do you collect all estimates.json files and not just the ones related to dummy_benches?

Thanks

---

bench_tools/src/benches/dummy_bench.rs line 11 at r1 (raw file):

Previously, avi-starkware (Avi Cohen) wrote…

correction: it is std::hint::black_box

According to AI, the compiler might delete the dummy_function calculation because we don't use the result.

---

bench_tools/src/benches/dummy_bench.rs line 14 at r1 (raw file):

Previously, meship-starkware (Meshi Peled) wrote…

Is this the function that creates the benchmark files? Where are we going to use it?
Can you add a small documentation on the purpose of this function?

It's a test util, we use it in estimates_test

[avi-starkware]

@avi-starkware reviewed 9 of 12 files at r2, 1 of 1 files at r3, all commit messages.
Reviewable status: 10 of 12 files reviewed, 3 unresolved discussions (waiting on @meship-starkware)

---

crates/bench_tools/src/types/estimates_test.rs line 27 at r3 (raw file):

fn workspace_root(bench_tools_crate_dir: PathBuf) -> PathBuf {
    bench_tools_crate_dir.parent().unwrap().parent().unwrap().to_path_buf()
}

Consider using workspace_root() from the crate infra_utils.

Code quote:

fn workspace_root(bench_tools_crate_dir: PathBuf) -> PathBuf {
    bench_tools_crate_dir.parent().unwrap().parent().unwrap().to_path_buf()
}

---

crates/bench_tools/src/types/estimates_test.rs line 87 at r3 (raw file):

fn deserialize_dummy_bench_estimates(dummy_bench_results_dir: PathBuf, dummy_bench_names: &[&str]) {
    assert_deserialize_dummy_bench_estimates(&dummy_bench_results_dir, dummy_bench_names);
}

Why is this test needed?

Code quote:

#[rstest]
/// Test that Estimates can be deserialized from the saved results.
fn deserialize_dummy_bench_estimates(dummy_bench_results_dir: PathBuf, dummy_bench_names: &[&str]) {
    assert_deserialize_dummy_bench_estimates(&dummy_bench_results_dir, dummy_bench_names);
}

[AvivYossef-starkware]

Reviewable status: 7 of 12 files reviewed, 1 unresolved discussion (waiting on @avi-starkware and @meship-starkware)

---

crates/bench_tools/src/types/estimates_test.rs line 27 at r3 (raw file):

Previously, avi-starkware (Avi Cohen) wrote…

Consider using workspace_root() from the crate infra_utils.

Done

---

crates/bench_tools/src/types/estimates_test.rs line 87 at r3 (raw file):

Previously, avi-starkware (Avi Cohen) wrote…

Why is this test needed?

to verify that the struct deserializes the json successfully

[avi-starkware]

@avi-starkware reviewed 2 of 3 files at r4, 1 of 1 files at r5, all commit messages.
Reviewable status: 10 of 12 files reviewed, 1 unresolved discussion (waiting on @AvivYossef-starkware and @meship-starkware)

---

crates/bench_tools/src/types/estimates_test.rs line 87 at r3 (raw file):

Previously, AvivYossef-starkware wrote…

to verify that the struct deserializes the json successfully

But isn't this verified in the test above as well?

[graphite-app]

Merge activity

• Oct 23, 8:25 AM UTC: Graphite rebased this pull request, because this pull request is set to merge when ready.

[AvivYossef-starkware]

Reviewable status: 9 of 12 files reviewed, 1 unresolved discussion (waiting on @avi-starkware and @meship-starkware)

---

crates/bench_tools/src/types/estimates_test.rs line 87 at r3 (raw file):

Previously, avi-starkware (Avi Cohen) wrote…

But isn't this verified in the test above as well?

But the test above is ignored because I dont want to actual run the benechmark in the CI

[avi-starkware]

@avi-starkware reviewed 1 of 1 files at r6, all commit messages.
Reviewable status: 10 of 12 files reviewed, all discussions resolved (waiting on @meship-starkware)

[avi-starkware]

@avi-starkware reviewed 2 of 12 files at r2.
Reviewable status: complete! all files reviewed, all discussions resolved (waiting on @meship-starkware)

[avi-starkware]

Reviewable status: complete! all files reviewed, all discussions resolved (waiting on @meship-starkware)