Do not gather local all together at the beginning of typeck

[compiler-errors]

r? lcnr

[compiler-errors]

@bors try @rust-timer queue

[bors]

⌛ Trying commit fc8f9a4 with merge 154b8c2...

[bors]

☀️ Try build successful - checks-actions
Build commit: 154b8c2 (154b8c2f37d3dea13f53c27255d91094e5044378)

[rust-timer]

Finished benchmarking commit (154b8c2): comparison URL.

Overall result: ✅ improvements - no action needed

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.

@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)	0.2%	[0.2%, 0.2%]	1
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-2.0%	[-22.4%, -0.2%]	60
Improvements ✅ (secondary)	-0.8%	[-1.9%, -0.1%]	12
All ❌✅ (primary)	-2.0%	[-22.4%, 0.2%]	61

Max RSS (memory usage)

Results (primary -0.8%, secondary -0.2%)

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)	0.6%	[0.4%, 0.9%]	6
Regressions ❌ (secondary)	3.4%	[2.0%, 4.8%]	2
Improvements ✅ (primary)	-1.1%	[-3.5%, -0.4%]	25
Improvements ✅ (secondary)	-3.8%	[-3.8%, -3.8%]	2
All ❌✅ (primary)	-0.8%	[-3.5%, 0.9%]	31

Cycles

Results (primary -1.6%, secondary -1.7%)

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)	0.6%	[0.5%, 0.7%]	4
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-1.9%	[-11.4%, -0.4%]	40
Improvements ✅ (secondary)	-1.7%	[-1.7%, -1.7%]	1
All ❌✅ (primary)	-1.6%	[-11.4%, 0.7%]	44

Binary size

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

Bootstrap: 768.358s -> 767.967s (-0.05%)
Artifact size: 365.48 MiB -> 365.38 MiB (-0.03%)

[lcnr]

small nits, otherwise r=me

the perf impact is wild :o

[lcnr]

r=me after nit

[compiler-errors]

@bors r=lcnr rollup=never

[bors]

📌 Commit 45598de has been approved by lcnr

It is now in the queue for this repository.

[bors]

⌛ Testing commit 45598de with merge f5d3fe2...

[bors]

☀️ Test successful - checks-actions
Approved by: lcnr
Pushing f5d3fe2 to master...

[rust-log-analyzer]

A job failed! Check out the build log: (web) (plain)

Click to see the possible cause of the failure (guessed by this bot)

   Compiling diff v0.1.13
   Compiling glob-match v0.2.1
   Compiling citool v0.1.0 (/home/runner/work/rust/rust/src/ci/citool)
    Finished `release` profile [optimized] target(s) in 37.82s
     Running `target/release/citool post-merge-report 651e9cf327358b28db7e37a2ae61727f4a2ef232 f5d3fe273b8b9e7125bf8856d44793b6cc4b6735`
Downloading metrics of job aarch64-gnu
Downloading metrics of job aarch64-gnu-debug
Downloading metrics of job arm-android
Downloading metrics of job armhf-gnu
Downloading metrics of job dist-aarch64-linux
Downloading metrics of job dist-android
Downloading metrics of job dist-arm-linux
Downloading metrics of job dist-armhf-linux
Downloading metrics of job dist-armv7-linux
Downloading metrics of job dist-i586-gnu-i586-i686-musl
Downloading metrics of job dist-i686-linux
Downloading metrics of job dist-loongarch64-linux
Downloading metrics of job dist-loongarch64-musl
Downloading metrics of job dist-ohos
Downloading metrics of job dist-powerpc-linux
Downloading metrics of job dist-powerpc64-linux
Downloading metrics of job dist-powerpc64le-linux
Downloading metrics of job dist-riscv64-linux
Downloading metrics of job dist-s390x-linux
Downloading metrics of job dist-various-1
Downloading metrics of job dist-various-2
Downloading metrics of job dist-x86_64-freebsd
Downloading metrics of job dist-x86_64-illumos
Downloading metrics of job dist-x86_64-linux
Downloading metrics of job dist-x86_64-linux-alt
Downloading metrics of job dist-x86_64-musl
Downloading metrics of job dist-x86_64-netbsd
Downloading metrics of job i686-gnu-1
Downloading metrics of job i686-gnu-2
Downloading metrics of job i686-gnu-nopt-1
Downloading metrics of job i686-gnu-nopt-2
Downloading metrics of job mingw-check
Downloading metrics of job test-various
Downloading metrics of job x86_64-rust-for-linux
Downloading metrics of job x86_64-gnu
Downloading metrics of job x86_64-gnu-stable
Downloading metrics of job x86_64-gnu-aux
Downloading metrics of job x86_64-gnu-debug
Downloading metrics of job x86_64-gnu-distcheck
Downloading metrics of job x86_64-gnu-llvm-20-1
Downloading metrics of job x86_64-gnu-llvm-20-2
Downloading metrics of job x86_64-gnu-llvm-20-3
Downloading metrics of job x86_64-gnu-llvm-19-1
Downloading metrics of job x86_64-gnu-llvm-19-2
Downloading metrics of job x86_64-gnu-llvm-19-3
Downloading metrics of job x86_64-gnu-nopt
Downloading metrics of job x86_64-gnu-tools
Downloading metrics of job dist-x86_64-apple
Downloading metrics of job dist-apple-various
Downloading metrics of job x86_64-apple-1
Downloading metrics of job x86_64-apple-2
Downloading metrics of job dist-aarch64-apple
Downloading metrics of job aarch64-apple
Downloading metrics of job x86_64-msvc-1
Downloading metrics of job x86_64-msvc-2
Downloading metrics of job i686-msvc-1
Downloading metrics of job i686-msvc-2
Downloading metrics of job x86_64-msvc-ext1
Downloading metrics of job x86_64-msvc-ext2
Downloading metrics of job x86_64-msvc-ext3
Downloading metrics of job x86_64-mingw-1
Downloading metrics of job x86_64-mingw-2
Downloading metrics of job dist-x86_64-msvc
Downloading metrics of job dist-i686-msvc
Downloading metrics of job dist-aarch64-msvc
Downloading metrics of job dist-i686-mingw
Downloading metrics of job dist-x86_64-mingw
Downloading metrics of job dist-x86_64-msvc-alt
<details>
<summary>What is this?</summary>
This is an experimental post-merge analysis report that shows differences in test outcomes between the merged PR and its parent PR.
</details>


Comparing 651e9cf327358b28db7e37a2ae61727f4a2ef232 (parent) -> f5d3fe273b8b9e7125bf8856d44793b6cc4b6735 (this PR)

# Test differences
<details>
<summary>Show 7 test diffs
</summary>


7 doctest diffs were found. These are ignored, as they are noisy.
</details>

<details>
<summary>Test dashboard</summary>

Run

```bash
cargo run --manifest-path src/ci/citool/Cargo.toml -- \
    test-dashboard f5d3fe273b8b9e7125bf8856d44793b6cc4b6735 --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. aarch64-apple: 4201.4s -> 3643.9s (-13.3%)
2. dist-x86_64-apple: 7884.2s -> 8887.1s (12.7%)
3. x86_64-apple-1: 8180.6s -> 9156.9s (11.9%)
4. dist-x86_64-freebsd: 4900.8s -> 5478.5s (11.8%)
5. dist-apple-various: 7467.3s -> 6786.8s (-9.1%)
6. x86_64-gnu-llvm-20-3: 7060.5s -> 6744.6s (-4.5%)
7. dist-aarch64-apple: 5424.5s -> 5665.6s (4.4%)
8. x86_64-mingw-2: 7233.5s -> 6919.6s (-4.3%)
9. dist-s390x-linux: 5479.2s -> 5245.3s (-4.3%)
10. x86_64-msvc-ext2: 5939.8s -> 5714.8s (-3.8%)

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.

no pull requests found for branch "master"
##[error]Process completed with exit code 1.
Post job cleanup.

</details>

[lqd]

cc @Kobzol on that post merge job failure

[Kobzol]

Thanks! #140703 should hopefully help.

[rust-timer]

Finished benchmarking commit (f5d3fe2): 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)	0.4%	[0.4%, 0.4%]	1
Regressions ❌ (secondary)	0.3%	[0.2%, 0.4%]	3
Improvements ✅ (primary)	-1.8%	[-22.4%, -0.1%]	73
Improvements ✅ (secondary)	-0.9%	[-2.0%, -0.3%]	12
All ❌✅ (primary)	-1.8%	[-22.4%, 0.4%]	74

Max RSS (memory usage)

Results (primary -0.0%, secondary 4.1%)

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)	1.2%	[0.5%, 2.2%]	12
Regressions ❌ (secondary)	4.1%	[4.1%, 4.1%]	1
Improvements ✅ (primary)	-0.9%	[-2.2%, -0.4%]	16
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-0.0%	[-2.2%, 2.2%]	28

Cycles

Results (primary -1.8%)

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)	1.0%	[0.5%, 2.8%]	5
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-2.3%	[-11.8%, -0.4%]	28
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-1.8%	[-11.8%, 2.8%]	33

Binary size

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

Bootstrap: 771.17s -> 770.498s (-0.09%)
Artifact size: 365.44 MiB -> 365.48 MiB (0.01%)

[nnethercote]

@compiler-errors: some amazing perf results there, well done. How did you find this/think of this?

[compiler-errors]

@nnethercote: Gathering locals at the beginning of typeck ended up leading to a hang in the new solver (-Znext-solver) for tests/ui/issues/issue-29466.rs. Investigating it led to my solution to register locals more closely to the first chance of them being used, which I only really intended to fix the hang in the new solver.

I guess it coincidentally ended up also being positive perf-wise in the old solver, but I didn't look too closely into why. Probably has something to do with holding a ton of pending _: Sized goals in the fulfillment context, which get repeatedly checked if they're unstalled.

[panstromek]

perf triage:

Large improvements outweigh regresions. Regressions are spurious (all regressed crates returned back to previous state in a following commit).

@rustbot label: +perf-regression-triaged

[nnethercote]

Here's the most important part of the Cachegrind before/after diff:

> -5,126,028,204 (54.6%,  54.6%)  <rustc_data_structures::obligation_forest::ObligationForest<rustc_trait_selection::traits::fulfill::PendingPredicateObligation>>::process_obligations::<rustc_trait_selection::traits::fulfill::FulfillProcessor>:
    -659,857,493  (7.0%)            /home/njn/dev/rustN/compiler/rustc_infer/src/infer/mod.rs
    -630,986,724  (6.7%)            /home/njn/dev/rustN/library/core/src/ptr/non_null.rs
    -599,600,922  (6.4%)            /home/njn/dev/rustN/compiler/rustc_trait_selection/src/traits/fulfill.rs
    -577,517,695  (6.2%)            /home/njn/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/ena-0.14.3/src/snapshot_vec.rs
    -528,960,891  (5.6%)            /home/njn/dev/rustN/compiler/rustc_data_structures/src/obligation_forest/mod.rs
    -430,668,099  (4.6%)            /home/njn/dev/rustN/library/alloc/src/raw_vec/mod.rs
    -372,323,242  (4.0%)            /home/njn/dev/rustN/library/core/src/slice/iter/macros.rs
    -296,879,490  (3.2%)            /home/njn/dev/rustN/library/core/src/iter/adapters/take_while.rs
    -290,626,855  (3.1%)            /home/njn/dev/rustN/library/alloc/src/vec/mod.rs
    -170,473,538  (1.8%)            /home/njn/dev/rustN/compiler/rustc_type_ir/src/ty_kind.rs
    -148,436,708  (1.6%)            /home/njn/dev/rustN/library/core/src/iter/traits/iterator.rs
    -110,168,691  (1.2%)            /home/njn/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/ena-0.14.3/src/unify/mod.rs
     -99,717,494  (1.1%)            /home/njn/dev/rustN/library/core/src/mem/mod.rs
     -86,851,469  (0.9%)            /home/njn/dev/rustN/library/core/src/slice/index.rs
     -67,686,216  (0.7%)            /home/njn/dev/rustN/library/core/src/cell.rs
     -55,646,480  (0.6%)            /home/njn/dev/rustN/library/core/src/iter/adapters/enumerate.rs

> -2,127,773,593 (22.7%,  77.3%)  <rustc_data_structures::obligation_forest::ObligationForest<rustc_trait_selection::traits::fulfill::PendingPredicateObligation>>::compress::<<rustc_data_structures::obligation_forest::ObligationForest<rustc_trait_selection::traits::fulfill::PendingPredicateObligation>>::process_obligations<rustc_trait_selection::traits::fulfill::FulfillProcessor>::{closure#1}>:
    -927,200,064  (9.9%)            /home/njn/dev/rustN/library/core/src/intrinsics/mod.rs
    -298,968,202  (3.2%)            /home/njn/dev/rustN/compiler/rustc_data_structures/src/obligation_forest/mod.rs
    -211,105,032  (2.2%)            /home/njn/dev/rustN/library/core/src/slice/index.rs
    -166,939,394  (1.8%)            /home/njn/dev/rustN/library/core/src/cell.rs
    -111,777,757  (1.2%)            /home/njn/dev/rustN/library/core/src/cmp.rs
     -88,304,768  (0.9%)            /home/njn/dev/rustN/library/core/src/slice/mod.rs
     -85,886,927  (0.9%)            /home/njn/dev/rustN/library/core/src/ptr/mod.rs
     -77,725,688  (0.8%)            /home/njn/dev/rustN/library/alloc/src/vec/mod.rs
     -62,446,541  (0.7%)            /home/njn/dev/rustN/library/core/src/iter/range.rs
     -55,646,692  (0.6%)            /home/njn/dev/rustN/library/alloc/src/raw_vec/mod.rs
     -41,768,204  (0.4%)            /home/njn/dev/rustN/library/core/src/num/uint_macros.rs

> -1,146,832,801 (12.2%,  89.5%)  <hashbrown::map::HashMap<rustc_middle::ty::ParamEnvAnd<rustc_middle::ty::predicate::Predicate>, usize, rustc_hash::FxBuildHasher>>::retain::<<rustc_data_structures::obligation_forest::ObligationForest<rustc_trait_selection::traits::fulfill::PendingPredicateObligation>>::apply_rewrites::{closure#0}>:
    -388,784,173  (4.1%)            /home/njn/dev/rustN/compiler/rustc_data_structures/src/obligation_forest/mod.rs
    -337,959,770  (3.6%)            /home/njn/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/hashbrown-0.15.2/src/map.rs
    -158,883,558  (1.7%)            /home/njn/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/hashbrown-0.15.2/src/control/bitmask.rs
    -122,798,105  (1.3%)            /home/njn/dev/rustN/library/core/src/ptr/mut_ptr.rs
     -58,120,140  (0.6%)            /home/njn/.cargo/registry/src/index.crates.io-1949cf8c6b5b557f/hashbrown-0.15.2/src/raw/mod.rs
     -55,540,626  (0.6%)            /home/njn/dev/rustN/library/core/src/num/nonzero.rs
     -23,177,646  (0.2%)            /home/njn/dev/rustN/library/core/src/../../stdarch/crates/core_arch/src/x86/sse2.rs
     -11,719,629  (0.1%)            /home/njn/dev/rustN/library/core/src/ptr/const_ptr.rs

In short, a lot less obligation processing than before.

[nnethercote]

In short, a lot less obligation processing than before.

Probably has something to do with holding a ton of pending _: Sized goals in the fulfillment context, which get repeatedly checked if they're unstalled.

This makes me wonder if there's some kind of domain-specific profiling of obligation processing that we could do/have done to detect this (or similar sub-optimalities). E.g. some kind of logging to identify when obligation processing is taking longer than necessary?