[perf] Reuse CanonicalVarKinds if no new variables were added by Randl · Pull Request #164304 · rust-lang/rust

Randl · 2026-10-14T08:12:18Z

Have seen a perf improvement locally.
The idea is to reuse the output of mk_canonical_var_kinds if self.state.var_kinds is unchanged

rustbot · 2026-10-14T08:12:24Z

r? @mati865

rustbot has assigned @mati865.
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

Why was this reviewer chosen?

The reviewer was selected based on:

Owners of files modified in this PR: compiler
compiler expanded to 77 candidates
Random selection from 18 candidates

mati865 · 2026-10-14T08:13:53Z

@bors try @rust-timer queue

[perf] Reuse CanonicalVarKinds if no new variables were added

rust-bors · 2026-10-14T09:13:55Z

☀️ Try build successful (CI)
Build commit: e2f1d32 (e2f1d325b097337a5b5f4eec8db11898130cf82f)
Base parent: 2d37576 (2d37576c3d7134aa2d5c2f2396712a8b4a38ef6f)

rust-timer · 2026-10-14T09:59:36Z

Finished benchmarking commit (e2f1d32): comparison URL.

Overall result: ❌✅ regressions and improvements - please read:

Benchmarking means the PR may be perf-sensitive. Consider adding rollup=never if this change is not fit for rolling up.

Next, please: If you can, justify the regressions found in this try perf run in writing along with @rustbot label: +perf-regression-triaged. If not, fix the regressions and do another perf run. Neutral or positive results will clear the label automatically.

@bors rollup=iffy rustc-perf
@rustbot label: -S-waiting-on-perf +perf-regression

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)	0.2%	[0.2%, 0.3%]	2
Regressions ❌ (secondary)	0.1%	[0.1%, 0.1%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-0.8%	[-1.6%, -0.3%]	9
All ❌✅ (primary)	0.2%	[0.2%, 0.3%]	2

Max RSS (memory usage)

Results (secondary -1.3%)

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)	2.2%	[2.2%, 2.2%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-4.9%	[-4.9%, -4.9%]	1
All ❌✅ (primary)	-	-	0

Cycles

Results (primary 2.9%)

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

	mean	range	count
Regressions ❌ (primary)	2.9%	[2.9%, 2.9%]	1
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	2.9%	[2.9%, 2.9%]	1

Binary size

This perf run didn't have relevant results for this metric.

Bootstrap: 487.613s -> 494.35s (1.38%)
Artifact size: 406.81 MiB -> 406.85 MiB (0.01%)

Randl · 2026-10-15T07:48:53Z

Are regressions noise? I believe it should have only influence the new solver

mati865 · 2026-10-15T10:36:50Z

Yes, those are noise.
I'm too unfamiliar with the code to tell whether this caching won't cause bugs, r? project-trait-system-refactor

ShoyuVanilla

IIUC, the idea behind this PR is that we reuse the param_env's var_kind if the rest of the canonicalized value doesn't have any other var.

That sounds plausible, as we may try to intern things less, but perhaps not in some pathological cases that we have to intern things twice for the param_env's var_kinds and that of the value. So I'm not entirely sure whether is this desirable but the perf looks good anyway 🤔

I'd like to try out an extra perf run after some changes:

View changes since this review

rustbot · 2026-10-15T12:06:28Z

Reminder, once the PR becomes ready for a review, use @rustbot ready.

ShoyuVanilla · 2026-10-18T06:54:23Z

@bors try @rust-timer queue

[perf] Reuse CanonicalVarKinds if no new variables were added

rust-bors · 2026-10-18T07:58:54Z

☀️ Try build successful (CI)
Build commit: f6e6dd9 (f6e6dd9401b66428a346b0a4fc02d47ae3796a26)
Base parent: 02fb0a1 (02fb0a16311c3411fa373627706da085cd8fbd3c)