Skip to content

Perf triage for this week #2100

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 3 commits into from
May 5, 2025
Merged

Perf triage for this week #2100

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
28d3fba
Add triage for this week
Mark-Simulacrum May 5, 2025
90ada33
Update root cause for 140633
Mark-Simulacrum May 5, 2025
11a8ba7
Update triage/2025-05-04.md
Mark-Simulacrum May 5, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
153 changes: 153 additions & 0 deletions triage/2025-05-04.md
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,153 @@
# 2025-05-04 Triage Log

A relatively noisy week due to addition of new benchmarks as part of our [2025
benchmark update], and a number of large regressions in a rollup landing late
in the week, with investigation [here](https://github.com/rust-lang/rust/pull/140652).

[2025 benchmark update]: https://github.com/rust-lang/rustc-perf/issues/2024

Triage done by **@simulacrum**.
Revision range: [25cdf1f6..62c5f58f](https://perf.rust-lang.org/?start=25cdf1f67463c9365d8d83778c933ec7480e940b&end=62c5f58f57670ce65e7fec34f8c4ba00c27da2d9&absolute=false&stat=instructions%3Au)

**Summary**:

| (instructions:u) | mean | range | count |
|:----------------------------------:|:-----:|:--------------:|:-----:|
| Regressions ❌ <br /> (primary) | 1.9% | [0.2%, 14.4%] | 47 |
| Regressions ❌ <br /> (secondary) | 2.0% | [0.1%, 11.8%] | 156 |
| Improvements ✅ <br /> (primary) | -0.6% | [-3.4%, -0.2%] | 34 |
| Improvements ✅ <br /> (secondary) | -1.6% | [-3.7%, -0.3%] | 20 |
| All ❌✅ (primary) | 0.9% | [-3.4%, 14.4%] | 81 |

2 Regressions, 2 Improvements, 6 Mixed; 3 of them in rollups
31 artifact comparisons made in total

#### Regressions

Introduce Arena::try_alloc_from_iter. [#140023](https://github.com/rust-lang/rust/pull/140023) [(Comparison Link)](https://perf.rust-lang.org/compare.html?start=74509131e85a97353c67c503ea32e148a56cf4bd&end=0fbb922e530399599aab8296ea975cb9e7ed67bf&stat=instructions:u)

| (instructions:u) | mean | range | count |
|:----------------------------------:|:----:|:------------:|:-----:|
| Regressions ❌ <br /> (primary) | 0.3% | [0.3%, 0.4%] | 3 |
| Regressions ❌ <br /> (secondary) | 0.3% | [0.1%, 0.4%] | 15 |
| Improvements ✅ <br /> (primary) | - | - | 0 |
| Improvements ✅ <br /> (secondary) | - | - | 0 |
| All ❌✅ (primary) | 0.3% | [0.3%, 0.4%] | 3 |

Regressions look plausibly real (they've held up for a bit) but given they're
small and limited to just doc builds I'm inclined to accept. It seems plausible
this is just some slight differences in the ABI of the functions here or
similar. Marking as triaged.

Set groundwork for proper const normalization [#140549](https://github.com/rust-lang/rust/pull/140549) [(Comparison Link)](https://perf.rust-lang.org/compare.html?start=3559e0ab0e370b98bc3f9e6b688152c47eb9e6c8&end=1bea580f364c65bd5f7380a1056e150df7b8a1a6&stat=instructions:u)

| (instructions:u) | mean | range | count |
|:----------------------------------:|:----:|:------------:|:-----:|
| Regressions ❌ <br /> (primary) | 0.7% | [0.3%, 1.4%] | 3 |
| Regressions ❌ <br /> (secondary) | 0.5% | [0.1%, 0.6%] | 7 |
| Improvements ✅ <br /> (primary) | - | - | 0 |
| Improvements ✅ <br /> (secondary) | - | - | 0 |
| All ❌✅ (primary) | 0.7% | [0.3%, 1.4%] | 3 |

Mostly what looks like inlining noise. See [comment](https://github.com/rust-lang/rust/pull/140549#issuecomment-2849664075)
for more details.

#### Improvements

perf: delay checking of `#[rustc_no_implicit_autorefs]` in autoref lint [#140406](https://github.com/rust-lang/rust/pull/140406) [(Comparison Link)](https://perf.rust-lang.org/compare.html?start=7c96085b64580af5c5f619384f0e3f082138ff13&end=4824c2bb7445cb2478aab0190c268c939d77a0f6&stat=instructions:u)

| (instructions:u) | mean | range | count |
|:----------------------------------:|:-----:|:--------------:|:-----:|
| Regressions ❌ <br /> (primary) | - | - | 0 |
| Regressions ❌ <br /> (secondary) | - | - | 0 |
| Improvements ✅ <br /> (primary) | -1.2% | [-3.4%, -0.2%] | 13 |
| Improvements ✅ <br /> (secondary) | -0.4% | [-0.4%, -0.4%] | 1 |
| All ❌✅ (primary) | -1.2% | [-3.4%, -0.2%] | 13 |

Use a closure instead of three chained iterators [#140464](https://github.com/rust-lang/rust/pull/140464) [(Comparison Link)](https://perf.rust-lang.org/compare.html?start=5fe04cbebd593c6530a21cd7bd0e8e9503b1ffe4&end=d7df5bdf2986e596aeaeec38e732711c69ebbce1&stat=instructions:u)

| (instructions:u) | mean | range | count |
|:----------------------------------:|:-----:|:--------------:|:-----:|
| Regressions ❌ <br /> (primary) | - | - | 0 |
| Regressions ❌ <br /> (secondary) | - | - | 0 |
| Improvements ✅ <br /> (primary) | -0.3% | [-0.6%, -0.2%] | 13 |
| Improvements ✅ <br /> (secondary) | -0.1% | [-0.1%, -0.1%] | 2 |
| All ❌✅ (primary) | -0.3% | [-0.6%, -0.2%] | 13 |

#### Mixed

Rollup of 7 pull requests [#140474](https://github.com/rust-lang/rust/pull/140474) [(Comparison Link)](https://perf.rust-lang.org/compare.html?start=efcbb94dccbe0fcc1076f031f88332db4a6a2a52&end=74509131e85a97353c67c503ea32e148a56cf4bd&stat=instructions:u)

| (instructions:u) | mean | range | count |
|:----------------------------------:|:-----:|:--------------:|:-----:|
| Regressions ❌ <br /> (primary) | 0.3% | [0.3%, 0.3%] | 1 |
| Regressions ❌ <br /> (secondary) | - | - | 0 |
| Improvements ✅ <br /> (primary) | -2.9% | [-2.9%, -2.9%] | 1 |
| Improvements ✅ <br /> (secondary) | -0.2% | [-0.2%, -0.2%] | 2 |
| All ❌✅ (primary) | -1.3% | [-2.9%, 0.3%] | 2 |

Regression is limited to unicode-normalization-0.1.19 and appears within noise
limit for that benchmark. Marking as triaged.

Simplify `LazyAttrTokenStream` [#127516](https://github.com/rust-lang/rust/pull/127516) [(Comparison Link)](https://perf.rust-lang.org/compare.html?start=0fbb922e530399599aab8296ea975cb9e7ed67bf&end=f242d6c26cc6fc187257bd1be9590b4b39632425&stat=instructions:u)

| (instructions:u) | mean | range | count |
|:----------------------------------:|:-----:|:--------------:|:-----:|
| Regressions ❌ <br /> (primary) | - | - | 0 |
| Regressions ❌ <br /> (secondary) | 0.6% | [0.3%, 0.8%] | 7 |
| Improvements ✅ <br /> (primary) | - | - | 0 |
| Improvements ✅ <br /> (secondary) | -2.1% | [-4.0%, -1.4%] | 16 |
| All ❌✅ (primary) | - | - | 0 |

> Improvements outweigh regressions, and all the movement is in secondary benchmarks.

Streamline the `format` macro. [#140188](https://github.com/rust-lang/rust/pull/140188) [(Comparison Link)](https://perf.rust-lang.org/compare.html?start=f242d6c26cc6fc187257bd1be9590b4b39632425&end=427288b3ce2d574847fdb41cc3184c893750e09a&stat=instructions:u)

| (instructions:u) | mean | range | count |
|:----------------------------------:|:-----:|:--------------:|:-----:|
| Regressions ❌ <br /> (primary) | 0.2% | [0.2%, 0.2%] | 1 |
| Regressions ❌ <br /> (secondary) | 1.8% | [1.7%, 1.9%] | 6 |
| Improvements ✅ <br /> (primary) | -0.3% | [-0.4%, -0.1%] | 10 |
| Improvements ✅ <br /> (secondary) | - | - | 0 |
| All ❌✅ (primary) | -0.2% | [-0.4%, 0.2%] | 11 |

> The improvements on primary benchmarks were deemed to outweigh the regressions on secondary benchmarks.

shared-generics: Do not share instantiations that contain local-only types [#138522](https://github.com/rust-lang/rust/pull/138522) [(Comparison Link)](https://perf.rust-lang.org/compare.html?start=6e23095adf9209614a45f7f75fea36dad7b92afb&end=0e517d38ad0e72f93c734b14fabd4bb9b7441de6&stat=instructions:u)

| (instructions:u) | mean | range | count |
|:----------------------------------:|:-----:|:--------------:|:-----:|
| Regressions ❌ <br /> (primary) | 0.9% | [0.4%, 1.5%] | 3 |
| Regressions ❌ <br /> (secondary) | 0.1% | [0.1%, 0.1%] | 1 |
| Improvements ✅ <br /> (primary) | -0.5% | [-0.7%, -0.3%] | 6 |
| Improvements ✅ <br /> (secondary) | -1.0% | [-1.0%, -1.0%] | 1 |
| All ❌✅ (primary) | -0.0% | [-0.7%, 1.5%] | 9 |

> Marking perf as triaged. There's some improvements, some regressions, but the
> regressions look like potentially bimodality -- and where they're possibly
> real hard to justify further digging in. The change seems overall a slight
> improvement on our benchmarks.

Rollup of 9 pull requests [#140596](https://github.com/rust-lang/rust/pull/140596) [(Comparison Link)](https://perf.rust-lang.org/compare.html?start=4824c2bb7445cb2478aab0190c268c939d77a0f6&end=2d5ffc513f1c56b7bc95bacb2519705096e8cc2b&stat=instructions:u)

| (instructions:u) | mean | range | count |
|:----------------------------------:|:-----:|:--------------:|:-----:|
| Regressions ❌ <br /> (primary) | - | - | 0 |
| Regressions ❌ <br /> (secondary) | 0.4% | [0.2%, 0.5%] | 2 |
| Improvements ✅ <br /> (primary) | -1.2% | [-1.2%, -1.2%] | 1 |
| Improvements ✅ <br /> (secondary) | - | - | 0 |
| All ❌✅ (primary) | -1.2% | [-1.2%, -1.2%] | 1 |

Regression triaged. Regressions look like bimodality/noise, not worth further follow up.

Rollup of 7 pull requests [#140633](https://github.com/rust-lang/rust/pull/140633) [(Comparison Link)](https://perf.rust-lang.org/compare.html?start=1bea580f364c65bd5f7380a1056e150df7b8a1a6&end=622ac043764d5d4ffff8de8cf86a1cc938a8a71b&stat=instructions:u)

| (instructions:u) | mean | range | count |
|:----------------------------------:|:-----:|:--------------:|:-----:|
| Regressions ❌ <br /> (primary) | 1.2% | [0.2%, 13.4%] | 77 |
| Regressions ❌ <br /> (secondary) | 1.9% | [0.2%, 11.0%] | 155 |
| Improvements ✅ <br /> (primary) | -1.5% | [-1.5%, -1.5%] | 1 |
| Improvements ✅ <br /> (secondary) | - | - | 0 |
| All ❌✅ (primary) | 1.2% | [-1.5%, 13.4%] | 78 |

Per investigation on the PR, the root cause is [#139675](https://github.com/rust-lang/rust/pull/139675). Per discussion [here](https://github.com/rust-lang/rust/pull/140652) we expect that this will be mostly resolved in our next LLVM bump, but is likely an acceptable regression in the meantime (primarily to very short compiles) for the feature in question.
Loading