2023-12-26 Triage Log

Some of the recent noise swings have been appearing again this week, but luckily less often than before. There were a few actual regressions, but most of them were localized to a single stress test. On the other hand, there were a few very nice wins across the board, especially for check and incremental builds, primarily thanks to #118824 and #119265.

Triage done by @kobzol. Revision range: bf9229a2e366b4c311f059014a4aa08af16de5d8..1ab783112ab4e4807304dbd249b39771246013ef

Summary:

(instructions:u)	mean	range	count
Regressions ❌ (primary)	0.6%	[0.6%, 0.6%]	1
Regressions ❌ (secondary)	2.5%	[0.3%, 4.2%]	10
Improvements ✅ (primary)	-0.8%	[-3.3%, -0.1%]	180
Improvements ✅ (secondary)	-1.2%	[-5.5%, -0.2%]	109
All ❌✅ (primary)	-0.8%	[-3.3%, 0.6%]	181

5 Regressions, 7 Improvements, 2 Mixed; 1 of them in rollups 58 artifact comparisons made in total

Regressions

Make exhaustiveness usable outside of rustc #118842 (Comparison Link)

(instructions:u)	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	0.9%	[0.9%, 1.0%]	6
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-0.4%	[-0.4%, -0.4%]	1
All ❌✅ (primary)	-	-	0

• This PR enables the use of exhaustiveness checking e.g. in rust-analyzer, which is part of the “librarification effort”. The small regression in a single stress-test benchmark is worth it.

mark ty::Const::Error when meet unsupport ty for const generic params #117176 (Comparison Link)

(instructions:u)	mean	range	count
Regressions ❌ (primary)	0.7%	[0.6%, 0.7%]	4
Regressions ❌ (secondary)	0.4%	[0.2%, 0.6%]	19
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.7%	[0.6%, 0.7%]	4

• This PR only added a test, so this has to be noise.
• Already marked as triaged by @jackh726

Exhaustiveness: Improve complexity on some wide matches #118796 (Comparison Link)

(instructions:u)	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	2.9%	[2.6%, 3.1%]	6
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-1.3%	[-1.3%, -1.3%]	1
All ❌✅ (primary)	-	-	0

• This PR fixed a potential hang in the compiler, the regression of the match checking stress-test is unfortunate, but it's more important to avoid the pathological case.
• Already marked as triaged by @apiraino

Subtree sync for rustc_codegen_cranelift #119278 (Comparison Link)

(instructions:u)	mean	range	count
Regressions ❌ (primary)	0.9%	[0.8%, 0.9%]	3
Regressions ❌ (secondary)	0.5%	[0.3%, 0.7%]	20
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.9%	[0.8%, 0.9%]	3

• Noise

fix ICE when using raw ptr in a pattern #119274 (Comparison Link)

(instructions:u)	mean	range	count
Regressions ❌ (primary)	0.7%	[0.7%, 0.7%]	1
Regressions ❌ (secondary)	0.4%	[0.3%, 0.5%]	6
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.7%	[0.7%, 0.7%]	1

• Noise

Improvements

fast path for declared_generic_bounds_from_env #119084 (Comparison Link)

(instructions:u)	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-1.9%	[-2.9%, -0.2%]	8
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-1.9%	[-2.9%, -0.2%]	8

• Very nice improvement to diesel.

use Vec for region constraints instead of BTreeMap #118824 (Comparison Link)

(instructions:u)	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-0.5%	[-1.3%, -0.2%]	39
Improvements ✅ (secondary)	-0.7%	[-1.7%, -0.2%]	29
All ❌✅ (primary)	-0.5%	[-1.3%, -0.2%]	39

• Nice improvements to check builds across the board.

Rollup of 6 pull requests #119237 (Comparison Link)

(instructions:u)	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-0.5%	[-0.8%, -0.3%]	17
Improvements ✅ (secondary)	-0.4%	[-0.6%, -0.2%]	10
All ❌✅ (primary)	-0.5%	[-0.8%, -0.3%]	17

• Nice win for type checking across several benchmarks. It's unclear which PR from the rollup has caused the wins though.

Stabilize ip_in_core feature #119276 (Comparison Link)

(instructions:u)	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-0.9%	[-0.9%, -0.8%]	4
Improvements ✅ (secondary)	-0.5%	[-0.8%, -0.2%]	21
All ❌✅ (primary)	-0.9%	[-0.9%, -0.8%]	4

• This is just noise returning back to steady state.

Unify SourceFile::name_hash and StableSourceFileId #119139 (Comparison Link)

(instructions:u)	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-0.2%	[-0.2%, -0.1%]	3
Improvements ✅ (secondary)	-0.8%	[-0.9%, -0.6%]	3
All ❌✅ (primary)	-0.2%	[-0.2%, -0.1%]	3

• A few small wins for incremental benchmarks.

Fix display of warning block if it is first element of the top doc block #119283 (Comparison Link)

(instructions:u)	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-0.7%	[-0.7%, -0.7%]	2
Improvements ✅ (secondary)	-0.4%	[-0.6%, -0.3%]	5
All ❌✅ (primary)	-0.7%	[-0.7%, -0.7%]	2

• This PR only modified the UI and tests of rustdoc, so this is just noise.

Make closures carry their own ClosureKind #119258 (Comparison Link)

(instructions:u)	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-1.1%	[-1.1%, -1.1%]	1
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-1.1%	[-1.1%, -1.1%]	1

• This is probably just noise.

Mixed

add more niches to rawvec #106790 (Comparison Link)

(instructions:u)	mean	range	count
Regressions ❌ (primary)	0.4%	[0.1%, 0.7%]	20
Regressions ❌ (secondary)	0.6%	[0.2%, 1.0%]	2
Improvements ✅ (primary)	-0.4%	[-2.0%, -0.1%]	27
Improvements ✅ (secondary)	-0.9%	[-5.3%, -0.2%]	16
All ❌✅ (primary)	-0.1%	[-2.0%, 0.7%]	47

• This PR enabled more optimization opportunities for LLVM, which could in theory help with program optimization, but it also caused some binary size increases and compile-time regressions.
• Not marking as triaged yet.

Remove metadata decoding DefPathHash cache #119265 (Comparison Link)

(instructions:u)	mean	range	count
Regressions ❌ (primary)	0.9%	[0.9%, 0.9%]	1
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-0.6%	[-2.1%, -0.2%]	101
Improvements ✅ (secondary)	-1.2%	[-3.2%, -0.2%]	65
All ❌✅ (primary)	-0.6%	[-2.1%, 0.9%]	102

• Very nice check and incremental wins across the board.
• The single regression is probably just noise returning back to steady state.