[panstromek]

No description provided.

[panstromek]

@bors try @rust-timer queue

[rust-bors]

☀️ Try build successful (CI)
Build commit: 46f8fa0 (46f8fa0ae54824f38293297a0ed124bc4946b6ae)
Base parent: aaf324d (aaf324d50a7dcb17aae7f2a37e15ff143d980c6e)

[rust-timer]

Finished benchmarking commit (46f8fa0): comparison URL.

Overall result: ✅ improvements - no action needed

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

@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
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-0.3%	[-0.5%, -0.0%]	12
All ❌✅ (primary)	-	-	0

Max RSS (memory usage)

Results (primary 0.3%)

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

	mean	range	count
Regressions ❌ (primary)	0.8%	[0.5%, 1.4%]	12
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-2.9%	[-5.4%, -0.5%]	2
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.3%	[-5.4%, 1.4%]	14

Cycles

Results (primary -0.2%, 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)	0.8%	[0.5%, 1.5%]	6
Regressions ❌ (secondary)	2.4%	[0.8%, 4.1%]	2
Improvements ✅ (primary)	-0.8%	[-1.4%, -0.4%]	10
Improvements ✅ (secondary)	-2.2%	[-2.2%, -2.2%]	1
All ❌✅ (primary)	-0.2%	[-1.4%, 1.5%]	16

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%	[0.0%, 0.2%]	67
Regressions ❌ (secondary)	0.1%	[0.0%, 0.3%]	56
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.1%	[0.0%, 0.2%]	67

Bootstrap: 488.092s -> 488.091s (-0.00%)
Artifact size: 405.81 MiB -> 406.14 MiB (0.08%)

[panstromek]

include-blob is noise.

Otherwise, the result is positive, but it's sadly not super visible. This code is only used in opt builds, because it's called by MatchBranchSimplification and JumpThreading MIR optimizations, so the relevant results are here: https://perf.rust-lang.org/compare.html?start=aaf324d50a7dcb17aae7f2a37e15ff143d980c6e&end=46f8fa0ae54824f38293297a0ed124bc4946b6ae&stat=instructions%3Au&clippy=false&check=false&debug=false&doc=false&doc-json=false&nonRelevant=true&incrUnchanged=false&incrPatched=false&incrFull=false

It's a bit unfortunate that for opt builds, almost everything is overshadowed by time spent in LLVM, so we can't easily measure small changes in there, it will end up below significance threshold.

When I run cachegrind diff on diesel, which is a benchmark that calls this function a lot (500k times), the result is pretty much exactly right:

-26,136,452 (100.0%)  PROGRAM TOTALS

  Ir________________________  function:file

> -9,129,848 (34.9%,  34.9%)  <hashbrown::map::HashMap<&rustc_middle::mir::BasicBlock, (), rustc_hash::FxBuildHasher>>::insert:
  -2,323,344  (8.9%)            /cargo/registry/25cdd57fae9f0462/hashbrown-0.17.1/src/map.rs
  -2,033,154  (7.8%)            /cargo/registry/25cdd57fae9f0462/hashbrown-0.17.1/src/raw.rs
  -1,160,760  (4.4%)            /cargo/registry/25cdd57fae9f0462/hashbrown-0.17.1/src/control/tag.rs
    -879,117  (3.4%)            /rustc/e3feeb59cd1bdd011a2bfcc0747b55125d9b8527/library/core/src/../../stdarch/crates/core_arch/src/x86/sse2.rs
    -580,380  (2.2%)            /rustc/e3feeb59cd1bdd011a2bfcc0747b55125d9b8527/library/core/src/num/uint_macros.rs
    -483,650  (1.9%)            /rustc/e3feeb59cd1bdd011a2bfcc0747b55125d9b8527/library/core/src/ptr/mod.rs
    -402,190  (1.5%)            /cargo/registry/25cdd57fae9f0462/hashbrown-0.17.1/src/control/bitmask.rs
    -388,136  (1.5%)            /rustc/e3feeb59cd1bdd011a2bfcc0747b55125d9b8527/library/core/src/intrinsics/mod.rs
    -293,838  (1.1%)            /rustc/e3feeb59cd1bdd011a2bfcc0747b55125d9b8527/library/core/src/ptr/mut_ptr.rs
    -290,190  (1.1%)            /rustc/e3feeb59cd1bdd011a2bfcc0747b55125d9b8527/library/core/src/macros/mod.rs
     -97,946  (0.4%)            /rustc/e3feeb59cd1bdd011a2bfcc0747b55125d9b8527/library/core/src/num/nonzero.rs
     -96,730  (0.4%)            /rustc/e3feeb59cd1bdd011a2bfcc0747b55125d9b8527/library/core/src/ub_checks.rs
     -96,730  (0.4%)            /rustc/e3feeb59cd1bdd011a2bfcc0747b55125d9b8527/library/core/src/convert/num.rs

> -4,028,190 (15.4%,  50.3%)  <hashbrown::raw::RawTable<(&rustc_middle::mir::BasicBlock, ())>>::reserve_rehash::<hashbrown::map::make_hasher<&rustc_middle::mir::BasicBlock, (), rustc_hash::FxBuildHasher>::{closure#0}>:
  -2,579,455  (9.9%)            /cargo/registry/25cdd57fae9f0462/hashbrown-0.17.1/src/raw.rs
    -459,355  (1.8%)            /rustc/e3feeb59cd1bdd011a2bfcc0747b55125d9b8527/library/core/src/num/uint_macros.rs
    -318,015  (1.2%)            /rustc/e3feeb59cd1bdd011a2bfcc0747b55125d9b8527/library/core/src/ptr/mod.rs
    -212,010  (0.8%)            /cargo/registry/25cdd57fae9f0462/hashbrown-0.17.1/src/alloc.rs
    -141,340  (0.5%)            /rustc/e3feeb59cd1bdd011a2bfcc0747b55125d9b8527/library/core/src/intrinsics/mod.rs
    -141,340  (0.5%)            /rustc/e3feeb59cd1bdd011a2bfcc0747b55125d9b8527/library/alloc/src/alloc.rs
     -70,670  (0.3%)            /rustc/e3feeb59cd1bdd011a2bfcc0747b55125d9b8527/library/core/src/cmp.rs
     -35,335  (0.1%)            /rustc/e3feeb59cd1bdd011a2bfcc0747b55125d9b8527/library/core/src/ub_checks.rs
     -35,335  (0.1%)            /rustc/e3feeb59cd1bdd011a2bfcc0747b55125d9b8527/library/core/src/ptr/mut_ptr.rs
     -35,335  (0.1%)            /cargo/registry/25cdd57fae9f0462/hashbrown-0.17.1/src/scopeguard.rs

> -1,972,066  (7.5%,  57.9%)  _int_free:./malloc/./malloc/malloc.c

> -1,683,275  (6.4%,  64.3%)  <core::iter::adapters::map::Map<core::slice::iter::Iter<rustc_middle::mir::BasicBlock>, <hashbrown::set::HashSet<&rustc_middle::mir::BasicBlock, rustc_hash::FxBuildHasher> as core::iter::traits::collect::Extend<&rustc_middle::mir::BasicBlock>>::extend<core::slice::iter::Iter<rustc_middle::mir::BasicBlock>>::{closure#0}> as core::iter::traits::iterator::Iterator>::fold::<(), core::iter::traits::iterator::Iterator::for_each::call<(&rustc_middle::mir::BasicBlock, ()), <hashbrown::map::HashMap<&rustc_middle::mir::BasicBlock, (), rustc_hash::FxBuildHasher> as core::iter::traits::collect::Extend<(&rustc_middle::mir::BasicBlock, ())>>::extend<core::iter::adapters::map::Map<core::slice::iter::Iter<rustc_middle::mir::BasicBlock>, <hashbrown::set::HashSet<&rustc_middle::mir::BasicBlock, rustc_hash::FxBuildHasher> as core::iter::traits::collect::Extend<&rustc_middle::mir::BasicBlock>>::extend<core::slice::iter::Iter<rustc_middle::mir::BasicBlock>>::{closure#0}>>::{closure#0}>::{closure#0}>:
    -598,930  (2.3%)            /rustc/e3feeb59cd1bdd011a2bfcc0747b55125d9b8527/library/core/src/slice/iter/macros.rs
    -424,020  (1.6%)            /rustc/e3feeb59cd1bdd011a2bfcc0747b55125d9b8527/library/core/src/iter/adapters/map.rs
    -290,190  (1.1%)            /cargo/registry/25cdd57fae9f0462/hashbrown-0.17.1/src/map.rs
    -193,460  (0.7%)            /rustc/e3feeb59cd1bdd011a2bfcc0747b55125d9b8527/library/core/src/num/uint_macros.rs
    -141,340  (0.5%)            /rustc/e3feeb59cd1bdd011a2bfcc0747b55125d9b8527/library/core/src/ptr/const_ptr.rs
     -35,335  (0.1%)            /rustc/e3feeb59cd1bdd011a2bfcc0747b55125d9b8527/library/core/src/ptr/non_null.rs
...

[rustbot]

r? @chenyukang

rustbot has assigned @chenyukang.
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 77 candidates
• Random selection from 21 candidates

[tmiasko]

@bors r+

[rust-bors]

📌 Commit f0974e1 has been approved by tmiasko

It is now in the queue for this repository.

[JonathanBrouwer]

As panstromek mentioned, include-blob perf is noise.
@bors rollup=maybe

(The diesel result looks good tho)

[panstromek]

I should say the cachegrind diff is from local ./x perf run, so it looks nicer. The run with CI artifacts has a lot more unrelated inlining noise. The reductions are there but they show up elsewhere (where they get inlined in mir opt pass code).