match: Use an aggregate equality comparison for constant array/slice patterns by jakubadamw · Pull Request #155216 · rust-lang/rust

jakubadamw · 2026-04-12T21:59:41Z

When every element in an array or slice pattern is a constant and there is no .. subpattern, the match builder will now emit a single call to PartialEq::eq instead of comparing each element from the value one by one against the respective constant in the pattern.

This drastically reduces the number of MIR basic blocks for large constant-array matches – e.g. a 64-element [u8; 64] match previously generated 64 separate comparison blocks and now generates just one PartialEq::eq call that LLVM can lower to a memcmp(). The optimisation is gated on having at least two constant elements, meaning single-element arrays will still use a plain scalar comparison.

Example:

const FOO: [u8; 64] = *b"0123456789abcdef0123456789abcdef0123456789abcdef0123456789abcdef";

pub fn foo(x: &[u8; 64]) -> bool {
    // Before: 64 basic blocks, one per byte.
    // After:  a single `PartialEq::eq()` call.
    matches!(x, &FOO)
}

Closes #103073.
Closes #110870.

rustbot · 2026-04-12T21:59:50Z

Some changes occurred in match lowering

cc @Nadrieril

rustbot · 2026-04-12T21:59:52Z

r? @JonathanBrouwer

rustbot has assigned @JonathanBrouwer.
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, mir
compiler, mir expanded to 69 candidates
Random selection from 12 candidates

JonathanBrouwer · 2026-04-13T18:15:32Z

@bors try @rust-timer queue

match: Use an aggregate equality comparison for constant array/slice patterns when possible

rust-bors · 2026-04-13T20:24:30Z

☀️ Try build successful (CI)
Build commit: 5a6dba6 (5a6dba60b1150a8b57fc739b0829fa4a65c5b8b3, parent: 14196dbfa3eb7c30195251eac092b1b86c8a2d84)

rust-timer · 2026-04-13T21:13:34Z

Finished benchmarking commit (5a6dba6): comparison URL.

Overall result: ❌✅ regressions and improvements - no action needed

Benchmarking means the PR may be perf-sensitive. It's automatically marked not fit for rolling up. Overriding is possible but disadvised: it risks changing compiler perf.

@bors rollup=never
@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
Regressions ❌ (secondary)	0.2%	[0.2%, 0.2%]	1
Improvements ✅ (primary)	-0.4%	[-0.4%, -0.4%]	1
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-0.4%	[-0.4%, -0.4%]	1

Max RSS (memory usage)

Results (primary -0.0%, secondary 0.5%)

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

	mean	range	count
Regressions ❌ (primary)	1.9%	[1.0%, 2.8%]	2
Regressions ❌ (secondary)	2.7%	[1.2%, 5.0%]	3
Improvements ✅ (primary)	-3.9%	[-3.9%, -3.9%]	1
Improvements ✅ (secondary)	-1.8%	[-2.2%, -1.6%]	3
All ❌✅ (primary)	-0.0%	[-3.9%, 2.8%]	3

Cycles

Results (primary -2.3%, secondary 14.4%)

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)	14.4%	[2.4%, 26.4%]	2
Improvements ✅ (primary)	-2.3%	[-3.1%, -1.7%]	4
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-2.3%	[-3.1%, -1.7%]	4

Binary size

Results (primary -0.1%, secondary 0.1%)

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

	mean	range	count
Regressions ❌ (primary)	0.1%	[0.1%, 0.1%]	1
Regressions ❌ (secondary)	0.1%	[0.1%, 0.1%]	2
Improvements ✅ (primary)	-0.2%	[-0.2%, -0.1%]	4
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-0.1%	[-0.2%, 0.1%]	5

Bootstrap: 491.114s -> 490.988s (-0.03%)
Artifact size: 394.23 MiB -> 394.33 MiB (0.03%)

JonathanBrouwer · 2026-04-15T19:44:15Z

@rustbot reroll
Not familiar enough with this code to comfortably review this, but from a quick glance this looks great, thanks <3

dianne · 2026-04-17T21:38:39Z

+            if let PatKind::Constant { value } = pat.kind {
+                Some(ty::Const::new_value(tcx, value.valtree, value.ty))
+            } else {
+                None
+            }


It might also be worth reconstructing aggregate constants for arrays/slices of arrays of constants, etc.? I'm not a specialization expert, but it looks like arrays of bytewise-comparable things are also bytewise-comparable, at least for common array lengths¹. Since array and slice equality are specialized based on their element types' bytewise-comparability, we should be able to get better codegen for nested array patterns too (as long as the inner arrays are of one of those common lengths), I think?

View changes since the review

Footnotes

https://github.com/rust-lang/rust/blob/f29256dd1420dc681bf4956e3012ffe9eccdc7e7/library/core/src/cmp/bytewise.rs#L74-L85 ↩

@dianne, interesting. I’ll look into this next! 🙂

dianne · 2026-04-17T21:49:51Z

+                // When there is no `..`, all elements are constants, and
+                // there are at least two of them, collapse the individual
+                // element subpairs into a single aggregate comparison that
+                // is performed after the length check.
+                if slice.is_none()


An additional possibility: even if there is a .., the comparisons for the sub-slices before and after the .. could be done via aggregate equality when applicable. Credit to #121540, which I think did this?

Edit: assuming prefixes and suffixes are typically small and hand-written, it's probably not worth the trouble to use aggregate equality for them.

Even if only handling the case with no .., it might be worth moving the special-casing into prefix_slice_suffix to share it between PatKind::Slice and PatKind::Array, since that's where the commonalities live.

Edit: after prefix_slice_suffix's cleanup in #154943, I don't think it makes much sense to put this in there. I still think the logic for deciding whether to use aggregate equality is complex enough that it could be worth factoring out, but that's probably not the way to do it.

View changes since the review

jackh726 · 2026-04-23T18:43:22Z

@dianne do you want to take over review here?

dianne

I don't have much context on the ctfe or const traits side of things to evaluate the approach. cc @oli-obk maybe? It feels like we have to make some sort of compromise here to avoid calling anything const-unstable. Possibly we could block on const_cmp stabilizing, or possibly we could have some workaround until then if we want to land this first.

I think I can handle the technical review, at least. Tentatively, r? me (though feel free to steal the assignment if that'd be easier ^^)

View changes since this review

…n't get captured by this logic

…` features

…ng it Following the review feedback, `const_to_pat()` now records the original constant value on the array and slice pattern nodes it expands, and match lowering reads that value back instead of attempting to reconstruct an aggregate constant from the individual element subpatterns. This has two consequences. First, hand-written array and slice patterns are no longer collapsed into aggregate comparisons; the user's intent to match element by element is respected before the MIR boundary. Only patterns that were expanded from an actual constant (a named constant or a byte-string literal) use the aggregate `PartialEq::eq` comparison, and for those the semantics of matching against a constant and comparing with `PartialEq::eq` coincide. Second, nested constant arrays now benefit from the aggregate comparison as well, since the recorded value covers the whole constant, whereas the reconstruction required every immediate element of the pattern to be a leaf constant.

Give the `PartialEq::eq` call emitted for constant array/slice patterns `UnwindAction::Unreachable` instead of an unwind edge. The built-in `PartialEq` implementations for arrays and slices can be trusted not to panic, and the unwind edge would not be harmless: since the aggregate comparison replaces a series of `SwitchInt` tests that could never unwind, the extra edge would make borrow-checking stricter about the drop order in unwinding code, turning previously accepted programs into errors. The string equality tests keep their unwind edge, as they have always had one.

…regates too

rustbot · 2026-08-23T19:09:50Z

This PR was rebased onto a different main 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.

jakubadamw · 2026-08-23T19:22:53Z

@dianne, thank you for your review, indeed. Apologies, it’s been a few busy weeks, but this is now ready for a re-review. I wonder if it’d make sense to do another perf run, seeing the LLVM 23 upgrade has taken place in the mean time.

dianne · 2026-08-24T14:13:10Z

no worries! I'll queue up another perf run. I'm not expecting much of a difference since (as far as I know) we don't have a stress test for big constant array/slice patterns as a benchmark, but best not to be surprised. the last perf run was on a version of the PR that used aggregate equality in more places too, so there's a small chance of the old results being off if a benchmark has big array/slice constant patterns of non-bytewise-comparable types.

@bors try @rust-timer queue

match: Use an aggregate equality comparison for constant array/slice patterns

rust-bors · 2026-08-24T15:11:59Z

☀️ Try build successful (CI)
Build commit: 950abd5 (950abd5388fe918da67b410a597cb425e519dc45)
Base parent: ac62df9 (ac62df9b49f9b9036af2a4957db70bf3850785e1)

rust-timer · 2026-08-24T16:19:23Z

Finished benchmarking commit (950abd5): comparison URL.

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

Benchmarking means the PR may be perf-sensitive. It's automatically marked not fit for rolling up. Overriding is possible but disadvised: it risks changing compiler perf.

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=never 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)	0.3%	[0.3%, 0.3%]	2
Regressions ❌ (secondary)	0.1%	[0.1%, 0.1%]	2
Improvements ✅ (primary)	-0.2%	[-0.3%, -0.1%]	3
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-0.0%	[-0.3%, 0.3%]	5

Max RSS (memory usage)

Results (primary -0.8%)

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

	mean	range	count
Regressions ❌ (primary)	3.0%	[3.0%, 3.0%]	1
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-2.7%	[-4.1%, -1.2%]	2
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-0.8%	[-4.1%, 3.0%]	3

Cycles

Results (primary -0.0%, secondary -3.0%)

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

	mean	range	count
Regressions ❌ (primary)	2.3%	[2.3%, 2.3%]	1
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-2.4%	[-2.4%, -2.4%]	1
Improvements ✅ (secondary)	-3.0%	[-3.0%, -3.0%]	1
All ❌✅ (primary)	-0.0%	[-2.4%, 2.3%]	2

Binary size

Results (primary -0.1%)

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

	mean	range	count
Regressions ❌ (primary)	1.0%	[1.0%, 1.0%]	1
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-0.3%	[-0.8%, -0.1%]	7
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-0.1%	[-0.8%, 1.0%]	8

Bootstrap: 475.143s -> 469.292s (-1.23%)
Artifact size: 400.28 MiB -> 400.36 MiB (0.02%)

rust-bors · 2026-09-01T23:56:03Z

☔ The latest upstream changes (presumably #162148) made this pull request unmergeable. Please resolve the merge conflicts by rebasing.

dianne

Finally getting back around to this. Thanks for your patience!

Perf seems fine probably? I haven't tried profiling locally to see where the extra instructions are coming from, but at a glance I'm not seeing anything too relevant in the perf report. Considering nothing looks extreme and the prior run came back clean, I'm not too worried.

@rustbot author

View changes since this review

dianne · 2026-09-09T18:03:21Z

+
+impl<'a, 'tcx> Builder<'a, 'tcx> {
+    /// Check if we can use aggregate `PartialEq::eq` comparisons for constant array/slice patterns.
+    /// This is not possible in const contexts, because `PartialEq` is not const-stable yet.


Suggested change

/// This is not possible in const contexts, because `PartialEq` is not const-stable yet.

/// This is not possible in const contexts, because `PartialEq` is not const-stable yet.

// FIXME(const_cmp): remove this restriction once `const_cmp` stabilizes

Not sure whether this will help make it easier to notice, but it can't hurt!

dianne · 2026-09-09T18:18:56Z

+/// Below this length, an array or slice pattern is compared element by element
+/// rather than as a single aggregate, since the per-element comparisons are
+/// unlikely to be more expensive than a `PartialEq::eq` call.
+const AGGREGATE_EQ_MIN_LEN: usize = 4;


I'm not too picky here, but do you have numbers or codegen comparisons for this? Could be a nice simplification to get rid of the cutoff number if it's not impactful. I expect on the types we're doing this for, eq calls are pretty cheap.

dianne · 2026-09-09T18:33:53Z

+        let (ty::Array(element_ty, _) | ty::Slice(element_ty)) = *pattern.ty.kind() else {
+            return None;
+        };


Since the expanded_const_value field is currently only added for arrays and slices, maybe this should be an assertion? Alternatively, we could always put the expanded const value in the THIR to keep all the logic for this in this function.

dianne · 2026-09-09T18:41:35Z

+                // If this pattern was expanded from a constant, compare the
+                // whole slice against that constant at once via
+                // `PartialEq::eq` after the length check, rather than
+                // element by element.


PartialEq::eq on slices should take length into account. Could we just do a TestableCase::Constant without an additional length check?

dianne · 2026-09-09T18:58:13Z

+                // That is sound because they are only used for element types whose
+                // `PartialEq` impl compares the aggregates directly with the
+                // `compare_bytes` and `raw_eq` intrinsics, which cannot panic.


I wonder if there's a good way to assert that we get the monomorphization we expect, or at least test for it. Maybe with a codegen-units/item-collection/ test we could look for the mono-item we want? Not sure if there's a better approach. Unfortunately not my area of expertise.

Maybe that'd be overkill for how much it'd give us. Not sure. I'm kind of wary about asserting that standard library functions can't unwind, so having some certainty that things are working as expected would be good. Maybe we should at least document in the standard library that we require those impls not to panic?

rustbot · 2026-09-09T19:13:06Z

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

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. T-compiler Relevant to the compiler team, which will review and decide on the PR/issue. labels Apr 12, 2026

rustbot assigned JonathanBrouwer Apr 12, 2026

This comment has been minimized.

Sign in to view

jakubadamw marked this pull request as draft April 12, 2026 23:12

rustbot added S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Apr 12, 2026

jakubadamw marked this pull request as ready for review April 12, 2026 23:50

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Apr 12, 2026

This comment has been minimized.

Sign in to view

rust-bors Bot pushed a commit that referenced this pull request Apr 13, 2026

Auto merge of #155216 - jakubadamw:issue-110870-103073, r=<try>

5a6dba6

match: Use an aggregate equality comparison for constant array/slice patterns when possible

rustbot added the S-waiting-on-perf Status: Waiting on a perf run to be completed. label Apr 13, 2026

This comment has been minimized.

Sign in to view

rustbot removed the S-waiting-on-perf Status: Waiting on a perf run to be completed. label Apr 13, 2026

rustbot assigned jackh726 and unassigned JonathanBrouwer Apr 15, 2026

dianne reviewed Apr 17, 2026

View reviewed changes

This comment has been minimized.

Sign in to view

Nadrieril reviewed Apr 22, 2026

View reviewed changes

Comment thread compiler/rustc_mir_build/src/builder/matches/match_pair.rs

dianne reviewed Apr 23, 2026

View reviewed changes

Comment thread compiler/rustc_mir_build/src/builder/matches/test.rs Outdated

Comment thread compiler/rustc_mir_build/src/builder/matches/match_pair.rs

rustbot assigned dianne Apr 23, 2026

jakubadamw added 11 commits August 23, 2026 18:43

Update the tests to cover const contexts as well

940ef6d

Raise the aggregate equality comparison threshold so simple arrays do…

6ec9cc1

…n't get captured by this logic

Merge the AggregateEq and StringEq test arms into a single case

485d99c

Add a missing space

f36ae4a

Remove the special treatment of the const_cmp and `const_trait_impl…

c3ad87b

…` features

Add a MIR test for a constant array with a custom element type

08a103e

Expand on why the aggregate comparison is assumed not to unwind

6b8eeaa

Note in the comment that exhaustiveness relies on PartialEq for agg…

06c2579

…regates too

Restrict the aggregate comparison to bytewise-comparable element types

d2a16fb

jakubadamw force-pushed the issue-110870-103073 branch from 2ea4f27 to d2a16fb Compare August 23, 2026 19:09

This comment has been minimized.

Sign in to view

rustbot added the S-waiting-on-perf Status: Waiting on a perf run to be completed. label Aug 24, 2026

This comment has been minimized.

Sign in to view

rust-bors Bot pushed a commit that referenced this pull request Aug 24, 2026

Auto merge of #155216 - jakubadamw:issue-110870-103073, r=<try>

950abd5

match: Use an aggregate equality comparison for constant array/slice patterns

This comment has been minimized.

Sign in to view

rustbot added perf-regression Performance regression. and removed S-waiting-on-perf Status: Waiting on a perf run to be completed. labels Aug 24, 2026

Nadrieril mentioned this pull request Sep 8, 2026

Don't require unsafe for struct and array patterns against union fields #161771

Open

dianne reviewed Sep 9, 2026

View reviewed changes

rustbot added S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Sep 9, 2026

	/// This is not possible in const contexts, because `PartialEq` is not const-stable yet.
	/// This is not possible in const contexts, because `PartialEq` is not const-stable yet.
	// FIXME(const_cmp): remove this restriction once `const_cmp` stabilizes

Uh oh!

Uh oh!

Conversation

jakubadamw commented Apr 12, 2026 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rustbot commented Apr 12, 2026

Uh oh!

rustbot commented Apr 12, 2026

Uh oh!

This comment has been minimized.

JonathanBrouwer commented Apr 13, 2026

Uh oh!

This comment has been minimized.

This comment has been minimized.

rust-bors Bot commented Apr 13, 2026

Uh oh!

This comment has been minimized.

rust-timer commented Apr 13, 2026

Overall result: ❌✅ regressions and improvements - no action needed

Uh oh!

JonathanBrouwer commented Apr 15, 2026

Uh oh!

Uh oh!

Uh oh!

dianne Apr 17, 2026 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Footnotes

Uh oh!

jakubadamw Apr 26, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

dianne Apr 17, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

This comment has been minimized.

Uh oh!

jackh726 commented Apr 23, 2026

Uh oh!

dianne left a comment • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

rustbot commented Aug 23, 2026

Uh oh!

jakubadamw commented Aug 23, 2026

Uh oh!

dianne commented Aug 24, 2026

Uh oh!

This comment has been minimized.

This comment has been minimized.

rust-bors Bot commented Aug 24, 2026

Uh oh!

This comment has been minimized.

rust-timer commented Aug 24, 2026

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

Uh oh!

rust-bors Bot commented Sep 1, 2026

Uh oh!

dianne left a comment • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

jakubadamw commented Apr 12, 2026 •

edited by rustbot

Loading

dianne Apr 17, 2026 •

edited by rustbot

Loading

jakubadamw Apr 26, 2026 •

edited

Loading

dianne Apr 17, 2026 •

edited

Loading

dianne left a comment •

edited by rustbot

Loading

dianne left a comment •

edited by rustbot

Loading