eagerly compute sub_unification_table again

[lcnr]

Previously called sub_relations. We still only using them for diagnostics right now. This mostly reverts #119989. Necessary for type inference guidance due to not-yet defined opaque types, cc rust-lang/trait-system-refactor-initiative#182.

We could use them for cycle detection in generalization and it seems desirable to do so in the future. However, this is unsound with the old trait solver as its cache does not track these sub_unification_table in any way.

We now properly track the sub_unification_table when canonicalizing so using them in the new solver is totally sound and the performance impact is far more manageable than I thought back in #119989.

r? @compiler-errors

[lcnr]

@bors try @rust-timer queue

[bors]

⌛ Trying commit 54a32a1 with merge 3c1d077...

[compiler-errors]

@bors try @rust-timer queue

[bors]

⌛ Trying commit 26f4ee0 with merge e184142...

[bors]

☀️ Try build successful - checks-actions
Build commit: e184142 (e1841428183d264d2f48328b3de5c0f25654315a)

[rust-timer]

Finished benchmarking commit (e184142): comparison URL.

Overall result: ❌ regressions - 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)	0.4%	[0.2%, 1.7%]	75
Regressions ❌ (secondary)	0.4%	[0.2%, 0.7%]	48
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.4%	[0.2%, 1.7%]	75

Max RSS (memory usage)

Results (primary -0.6%, secondary -4.5%)

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)	0.8%	[0.8%, 0.8%]	1
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-0.7%	[-1.1%, -0.4%]	12
Improvements ✅ (secondary)	-4.5%	[-4.5%, -4.5%]	1
All ❌✅ (primary)	-0.6%	[-1.1%, 0.8%]	13

Cycles

Results (primary 0.0%, secondary -1.6%)

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)	0.7%	[0.5%, 0.9%]	3
Regressions ❌ (secondary)	2.2%	[2.2%, 2.2%]	1
Improvements ✅ (primary)	-0.6%	[-0.7%, -0.5%]	3
Improvements ✅ (secondary)	-2.6%	[-2.8%, -2.1%]	4
All ❌✅ (primary)	0.0%	[-0.7%, 0.9%]	6

Binary size

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

Bootstrap: 765.701s -> 765.069s (-0.08%)
Artifact size: 365.36 MiB -> 365.11 MiB (-0.07%)

[bors]

☔ The latest upstream changes (presumably #140474) made this pull request unmergeable. Please resolve the merge conflicts.

[lcnr]

see perf run in #140480 (comment)

[lcnr]

comparing these two perf runs, tracking sub_relations in canonical inputs does slightly worsen performance: https://perf.rust-lang.org/compare.html?start=e1841428183d264d2f48328b3de5c0f25654315a&end=9613271ee15ea18ed41f396fc7aaf4c1f760410a&stat=instructions%3Au

[lcnr]

@rustbot ready

[rustbot]

This PR was rebased onto a different master commit. Here's a range-diff highlighting what actually changed.

Rebasing is a normal part of keeping PRs up to date, so no action is needed—this note is just to help reviewers.

[lcnr]

@bors r=BoxyUwU rollup=never

The small perf regression is expected as we're now tracking more information. Part of it is just losing the perf gains from #119989.

[bors]

📌 Commit b51a3a5 has been approved by BoxyUwU

It is now in the queue for this repository.

[bors]

⌛ Testing commit b51a3a5 with merge 9c27f27...

[bors]

☀️ Test successful - checks-actions
Approved by: BoxyUwU
Pushing 9c27f27 to master...

[github-actions]

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 a78f9aa (parent) -> 9c27f27 (this PR)

Test differences

Show 20 test diffs

Stage 1

• [ui] tests/ui/type-alias-impl-trait/fallback.rs: pass -> [missing] (J0)

Stage 2

• [ui] tests/ui/type-alias-impl-trait/fallback.rs: pass -> [missing] (J1)

Additionally, 18 doctest diffs were found. These are ignored, as they are noisy.

Job group index

• J0: x86_64-gnu-llvm-19-3, x86_64-gnu-llvm-20-3
• J1: aarch64-apple, aarch64-gnu, aarch64-gnu-llvm-19-1, aarch64-msvc-1, arm-android, armhf-gnu, dist-i586-gnu-i586-i686-musl, i686-gnu-1, i686-gnu-nopt-1, i686-msvc-1, test-various, x86_64-gnu, x86_64-gnu-debug, x86_64-gnu-llvm-19, x86_64-gnu-llvm-19-2, x86_64-gnu-llvm-20-2, x86_64-gnu-nopt, x86_64-gnu-stable, x86_64-mingw-1, x86_64-msvc-1

Test dashboard

Run

cargo run --manifest-path src/ci/citool/Cargo.toml -- \
    test-dashboard 9c27f27ea3bab79a2fec827ef3ae0009959d60f4 --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

1. dist-aarch64-linux: 6256.2s -> 8830.2s (41.1%)
2. aarch64-apple: 6631.7s -> 5565.9s (-16.1%)
3. pr-check-1: 1408.2s -> 1633.9s (16.0%)
4. x86_64-gnu-llvm-19-2: 6726.2s -> 5969.6s (-11.2%)
5. armhf-gnu: 4891.9s -> 5399.5s (10.4%)
6. dist-ohos-x86_64: 4112.2s -> 4431.7s (7.8%)
7. dist-aarch64-msvc: 5421.6s -> 5807.4s (7.1%)
8. dist-arm-linux-gnueabi: 4814.8s -> 5129.5s (6.5%)
9. i686-gnu-2: 5667.7s -> 6037.6s (6.5%)
10. dist-ohos-aarch64: 4173.3s -> 4434.1s (6.2%)

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]

Finished benchmarking commit (9c27f27): 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

Our most reliable metric. Used to determine the overall result above. However, even this metric can be noisy.

	mean	range	count
Regressions ❌ (primary)	0.3%	[0.1%, 1.4%]	61
Regressions ❌ (secondary)	0.6%	[0.1%, 2.7%]	71
Improvements ✅ (primary)	-0.4%	[-0.8%, -0.1%]	3
Improvements ✅ (secondary)	-0.9%	[-2.9%, -0.2%]	21
All ❌✅ (primary)	0.3%	[-0.8%, 1.4%]	64

Max RSS (memory usage)

Results (primary -1.3%, secondary -0.9%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	1.2%	[0.8%, 1.6%]	2
Improvements ✅ (primary)	-1.3%	[-1.3%, -1.3%]	1
Improvements ✅ (secondary)	-1.6%	[-2.6%, -1.1%]	6
All ❌✅ (primary)	-1.3%	[-1.3%, -1.3%]	1

Cycles

Results (primary -3.0%, secondary -1.5%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	2.7%	[1.9%, 3.9%]	6
Improvements ✅ (primary)	-3.0%	[-3.8%, -2.3%]	7
Improvements ✅ (secondary)	-4.7%	[-8.3%, -2.5%]	8
All ❌✅ (primary)	-3.0%	[-3.8%, -2.3%]	7

Binary size

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

Bootstrap: 466.425s -> 467.433s (0.22%)
Artifact size: 387.38 MiB -> 387.54 MiB (0.04%)

[lcnr]

Unfortunately an expected perf regression, spent a while trying to reduce the impact. Necessary for the new solver

[Kobzol]

Marking as triaged.

@rustbot label: +perf-regression-triaged