Rework `collect_and_apply` to not rely on size hint for optimization #141652

compiler-errors · 2025-05-27T14:09:37Z

I saw that we have quite a few collect_and_apply calls for N=3-7 (N=7 corresponding to cumulative 99% of nalgebra's calls). Didn't perf locally, but also this is super low-pri, so let's see what rust-timer says.

compiler-errors · 2025-05-27T14:09:43Z

@bors try @rust-timer queue

compiler-errors · 2025-05-27T14:09:56Z

compiler/rustc_type_ir/src/interner.rs

@@ -402,6 +402,51 @@ impl<T, R> CollectAndApply<T, R> for T {
                assert!(iter.next().is_none());
                f(&[t0, t1])
            }
+            (3, Some(3)) => {


this could be a macro

[PERF] More collect and apply I saw that we have quite a few `collect_and_apply` calls for N=3-7 (N=7 corresponding to cumulative 99% of nalgebra's calls). Didn't perf locally, but also this is super low-pri, so let's see what rust-timer says.

bors · 2025-05-27T14:10:53Z

⌛ Trying commit 618a3b0 with merge 23b0844...

bors · 2025-05-27T16:19:56Z

☀️ Try build successful - checks-actions
Build commit: 23b0844 (23b0844997f3003800cd6bf24365e11d353a5709)

rust-timer · 2025-05-28T02:22:43Z

Finished benchmarking commit (23b0844): comparison URL.

Overall result: ❌✅ regressions and improvements - please read the text below

Benchmarking this pull request likely means that it is perf-sensitive, so we're automatically marking it as not fit for rolling up. While you can manually mark this PR as fit for rollup, we strongly recommend not doing so since this PR may lead to changes in compiler perf.

Next Steps: If you can justify the regressions found in this try perf run, please indicate this with @rustbot label: +perf-regression-triaged along with sufficient written justification. If you cannot justify the regressions please fix the regressions and do another perf run. If the next run shows neutral or positive results, the label will be automatically removed.

@bors rollup=never
@rustbot label: -S-waiting-on-perf +perf-regression

Instruction count

This is the most reliable metric that we have; it was used to determine the overall result at the top of this comment. However, even this metric can sometimes exhibit noise.

	mean	range	count
Regressions ❌ (primary)	3.0%	[3.0%, 3.0%]	1
Regressions ❌ (secondary)	0.6%	[0.1%, 1.3%]	27
Improvements ✅ (primary)	-0.2%	[-0.2%, -0.1%]	6
Improvements ✅ (secondary)	-0.3%	[-0.5%, -0.1%]	19
All ❌✅ (primary)	0.3%	[-0.2%, 3.0%]	7

Max RSS (memory usage)

Results (primary 1.7%, secondary 4.2%)

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	1.7%	[1.7%, 1.7%]	1
Regressions ❌ (secondary)	6.1%	[1.2%, 9.8%]	10
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-1.9%	[-2.5%, -1.2%]	3
All ❌✅ (primary)	1.7%	[1.7%, 1.7%]	1

Cycles

Results (primary 2.7%, secondary 2.7%)

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	2.7%	[2.7%, 2.7%]	1
Regressions ❌ (secondary)	2.7%	[1.3%, 3.1%]	6
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	2.7%	[2.7%, 2.7%]	1

Binary size

Results (primary 1.1%)

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	1.1%	[1.1%, 1.1%]	1
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	1.1%	[1.1%, 1.1%]	1

Bootstrap: 780.158s -> 784.918s (0.61%)
Artifact size: 366.33 MiB -> 368.72 MiB (0.65%)

compiler-errors · 2025-05-31T10:08:55Z

@bors try @rust-timer queue

bors · 2025-05-31T10:10:05Z

⌛ Trying commit 8180ed0 with merge 4997545...

[PERF] More collect and apply I saw that we have quite a few `collect_and_apply` calls for N=3-7 (N=7 corresponding to cumulative 99% of nalgebra's calls). Didn't perf locally, but also this is super low-pri, so let's see what rust-timer says.

compiler-errors · 2025-05-31T10:32:58Z

compiler/rustc_type_ir/src/interner.rs

+        };
+
+        let Some(t2) = iter.next() else {
+            return f(&[t1]);


bors · 2025-05-31T10:33:05Z

💔 Test failed - checks-actions

compiler-errors · 2025-06-01T17:20:38Z

@bors2 try @rust-timer queue

rust-bors · 2025-06-01T17:20:42Z

⌛ Trying commit e311c75 with merge e582a3d…

[PERF] More collect and apply I saw that we have quite a few `collect_and_apply` calls for N=3-7 (N=7 corresponding to cumulative 99% of nalgebra's calls). Didn't perf locally, but also this is super low-pri, so let's see what rust-timer says.

rust-bors · 2025-06-01T19:45:32Z

☀️ Try build successful (CI)
Build commit: e582a3d (e582a3daad87838574e3805e1708c269b5684437)

rustbot · 2025-06-03T14:31:06Z

r? @jieyouxu

rustbot has assigned @jieyouxu.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

compiler-errors · 2025-06-03T14:32:08Z

r? lqd (or other perf ppl)

Seems like a slight improvement, I assume this helps for iterators like .filter() and .flat_map() where they end up knocking out the size hint.

Maybe there's more work to be done here? idk.

lqd · 2025-06-03T16:33:43Z

Yeah, LGTM: a slight improvement, not too complicated, and the gains outweigh the tiny regressions on the secondary stress test. The code may look strange to people stumbling upon it (but it's not unexpected ofc) so I'd like to retain some of the prior comment that we're micro-optimizing for the most common lengths in super hot code.

r=me with that comment (and the commit message fixed :3)

compiler-errors · 2025-06-03T17:59:02Z

@bors r=lqd

bors · 2025-06-03T17:59:05Z

📌 Commit 0479e74 has been approved by lqd

It is now in the queue for this repository.

bors · 2025-06-04T04:53:01Z

⌛ Testing commit 0479e74 with merge d9a7393...

bors · 2025-06-04T08:15:56Z

☀️ Test successful - checks-actions
Approved by: lqd
Pushing d9a7393 to master...

github-actions · 2025-06-04T08:19:00Z

What is this?

This is an experimental post-merge analysis report that shows differences in test outcomes between the merged PR and its parent PR.

Comparing 792fc2b (parent) -> d9a7393 (this PR)

Test differences

No test diffs found

Test dashboard

Run

cargo run --manifest-path src/ci/citool/Cargo.toml -- \
    test-dashboard d9a739363878a06633eb912e44f02a3c5d6ba08c --output-dir test-dashboard

And then open test-dashboard/index.html in your browser to see an overview of all executed tests.

Job duration changes

x86_64-apple-1: 6807.6s -> 8404.7s (23.5%)
dist-x86_64-apple: 9323.0s -> 7523.3s (-19.3%)
aarch64-apple: 4704.8s -> 5402.0s (14.8%)
dist-aarch64-apple: 6738.7s -> 6165.7s (-8.5%)
mingw-check-2: 2019.3s -> 1863.3s (-7.7%)
dist-armhf-linux: 4613.8s -> 4970.0s (7.7%)
dist-apple-various: 6012.1s -> 6434.9s (7.0%)
x86_64-gnu-llvm-19-3: 6398.1s -> 6815.8s (6.5%)
mingw-check-1: 1899.7s -> 1779.8s (-6.3%)
x86_64-gnu-llvm-19-2: 5520.3s -> 5817.3s (5.4%)

How to interpret the job duration changes?

Job durations can vary a lot, based on the actual runner instance
that executed the job, system noise, invalidated caches, etc. The table above is provided
mostly for t-infra members, for simpler debugging of potential CI slow-downs.

rust-timer · 2025-06-04T13:54:31Z

Finished benchmarking commit (d9a7393): comparison URL.

Overall result: ❌✅ regressions and improvements - please read the text below

Our benchmarks found a performance regression caused by this PR.
This might be an actual regression, but it can also be just noise.

Next Steps:

If the regression was expected or you think it can be justified,
please write a comment with sufficient written justification, and add
@rustbot label: +perf-regression-triaged to it, to mark the regression as triaged.
If you think that you know of a way to resolve the regression, try to create
a new PR with a fix for the regression.
If you do not understand the regression or you think that it is just noise,
you can ask the @rust-lang/wg-compiler-performance working group for help (members of this group
were already notified of this PR).

@rustbot label: +perf-regression
cc @rust-lang/wg-compiler-performance

Instruction count

This is the most reliable metric that we have; it was used to determine the overall result at the top of this comment. However, even this metric can sometimes exhibit noise.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	0.3%	[0.3%, 0.4%]	6
Improvements ✅ (primary)	-0.2%	[-0.4%, -0.1%]	9
Improvements ✅ (secondary)	-0.5%	[-1.0%, -0.2%]	49
All ❌✅ (primary)	-0.2%	[-0.4%, -0.1%]	9

Max RSS (memory usage)

Results (secondary 3.3%)

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	5.6%	[4.9%, 6.2%]	2
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-1.2%	[-1.2%, -1.2%]	1
All ❌✅ (primary)	-	-	0

Cycles

Results (secondary 1.5%)

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	1.5%	[1.5%, 1.5%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

Binary size

This benchmark run did not return any relevant results for this metric.

Bootstrap: 745.447s -> 748.459s (0.40%)
Artifact size: 372.35 MiB -> 371.85 MiB (-0.13%)

panstromek · 2025-06-09T11:59:59Z

perf triage:

Improvements outweigh regressions. Regressions are limited to wg-grammar. Also justified by #141652 (comment)

@rustbot label: +perf-regression-triaged

rustbot added the T-compiler label May 27, 2025