[PERF] partition shims into CGU of their characteristic types, if possible #141560

compiler-errors · 2025-05-25T20:44:11Z

We throw all non-InstanceKind::Item mono items into the same fallback CGU. This seems to have been a thing since #32779. This PR investigates: what happens to incremental codegen if we didn't do that, but tried a bit harder to partition shims into the same CGU as their characteristic types?

r? @ghost

compiler-errors · 2025-05-25T20:44:29Z

@bors try @rust-timer queue

bors · 2025-05-25T20:45:40Z

⌛ Trying commit 8ce6e6e with merge 01c9c97...

…try> [PERF] partition shims into CGU of their characteristic types, if possible We throw all non-`InstanceKind::Item` mono items into the same fallback CGU. What happens to incremental codegen if we didn't do that, but tried a bit harder to partition shims into the same CGU as their characteristic types? r? `@ghost`

compiler/rustc_monomorphize/src/partitioning.rs

compiler-errors · 2025-05-25T22:04:59Z

@bors try @rust-timer queue

…try> [PERF] partition shims into CGU of their characteristic types, if possible We throw all non-`InstanceKind::Item` mono items into the same fallback CGU. This seems to have been a thing since #32779. This PR investigates: what happens to incremental codegen if we didn't do that, but tried a bit harder to partition shims into the same CGU as their characteristic types? r? `@ghost`

bors · 2025-05-25T22:06:10Z

⌛ Trying commit 3695a33 with merge 155e6b9...

bors · 2025-05-26T00:12:47Z

☀️ Try build successful - checks-actions
Build commit: 155e6b9 (155e6b96fcb60833ff6194a1e370d66371850f47)

rust-timer · 2025-05-26T01:43:59Z

Finished benchmarking commit (155e6b9): comparison URL.

Overall result: ❌✅ regressions and improvements - please read the text below

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.

Next Steps: If you can justify the regressions found in this try perf run, please indicate this with @rustbot label: +perf-regression-triaged along with sufficient written justification. If you cannot justify the regressions please fix the regressions and do another perf run. If the next run shows neutral or positive results, the label will be automatically removed.

@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)	4.9%	[0.2%, 17.9%]	30
Regressions ❌ (secondary)	4.3%	[0.2%, 23.6%]	8
Improvements ✅ (primary)	-2.0%	[-4.2%, -0.3%]	5
Improvements ✅ (secondary)	-1.2%	[-2.7%, -0.2%]	8
All ❌✅ (primary)	3.9%	[-4.2%, 17.9%]	35

Max RSS (memory usage)

Results (primary 0.4%, secondary -0.7%)

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)	3.4%	[1.2%, 9.9%]	8
Regressions ❌ (secondary)	2.5%	[2.3%, 2.8%]	3
Improvements ✅ (primary)	-3.7%	[-5.0%, -1.2%]	6
Improvements ✅ (secondary)	-3.9%	[-5.2%, -2.1%]	3
All ❌✅ (primary)	0.4%	[-5.0%, 9.9%]	14

Cycles

Results (primary 6.6%, secondary 5.3%)

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)	8.2%	[0.8%, 19.7%]	18
Regressions ❌ (secondary)	9.1%	[0.8%, 26.4%]	4
Improvements ✅ (primary)	-2.7%	[-4.2%, -1.7%]	3
Improvements ✅ (secondary)	-2.3%	[-2.9%, -1.7%]	2
All ❌✅ (primary)	6.6%	[-4.2%, 19.7%]	21

Binary size

Results (primary 0.4%, secondary 0.5%)

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%	[0.0%, 1.8%]	59
Regressions ❌ (secondary)	4.6%	[0.0%, 14.4%]	10
Improvements ✅ (primary)	-0.3%	[-0.7%, -0.0%]	25
Improvements ✅ (secondary)	-2.5%	[-5.4%, -0.0%]	14
All ❌✅ (primary)	0.4%	[-0.7%, 1.8%]	84

Bootstrap: 775.708s -> 776.167s (0.06%)
Artifact size: 366.25 MiB -> 366.31 MiB (0.02%)

compiler-errors · 2025-06-13T18:15:08Z

@bors2 try @rust-timer queue

rust-bors · 2025-06-13T18:15:12Z

⌛ Trying commit 2b34e6f with merge 70425c9…

To cancel the try build, run the command @bors2 try cancel.

…try> [PERF] partition shims into CGU of their characteristic types, if possible We throw all non-`InstanceKind::Item` mono items into the same fallback CGU. This seems to have been a thing since #32779. This PR investigates: what happens to incremental codegen if we didn't do that, but tried a bit harder to partition shims into the same CGU as their characteristic types? r? `@ghost`

rust-bors · 2025-06-13T18:36:45Z

💔 Test failed

CI ❌

compiler-errors · 2025-06-13T19:57:23Z

@bors2 try @rust-timer queue

rust-bors · 2025-06-13T19:57:27Z

⌛ Trying commit edf7485 with merge c500484…

To cancel the try build, run the command @bors2 try cancel.

…try> [PERF] partition shims into CGU of their characteristic types, if possible We throw all non-`InstanceKind::Item` mono items into the same fallback CGU. This seems to have been a thing since #32779. This PR investigates: what happens to incremental codegen if we didn't do that, but tried a bit harder to partition shims into the same CGU as their characteristic types? r? `@ghost`

rust-log-analyzer · 2025-06-13T20:36:57Z

The job aarch64-gnu-llvm-19-1 failed! Check out the build log: (web) (plain)

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

---- [codegen-units] tests/codegen-units/partitioning/extern-drop-glue.rs stdout ----

The following items were assigned to wrong codegen units:

fn std::ptr::drop_in_place::<LocalStruct> - shim(Some(LocalStruct))
  expected: extern_drop_glue-fallback.cgu[External] 
  actual:   extern_drop_glue.volatile[External] 

fn std::ptr::drop_in_place::<cgu_extern_drop_glue::Struct> - shim(Some(cgu_extern_drop_glue::Struct))
  expected: extern_drop_glue-fallback.cgu[External] 
  actual:   cgu_extern_drop_glue-in-extern_drop_glue.volatile[External] 

fn std::ptr::drop_in_place::<mod1::LocalStruct> - shim(Some(mod1::LocalStruct))
  expected: extern_drop_glue-fallback.cgu[External] 
  actual:   extern_drop_glue-mod1.volatile[External] 


thread '[codegen-units] tests/codegen-units/partitioning/extern-drop-glue.rs' panicked at src/tools/compiletest/src/runtest/codegen_units.rs:103:13:
explicit panic
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

---- [codegen-units] tests/codegen-units/partitioning/local-drop-glue.rs stdout ----

The following items were assigned to wrong codegen units:

fn <Struct as std::ops::Drop>::drop
  expected: local_drop_glue-fallback.cgu[External] 
  actual:   local_drop_glue[External] 

fn std::ptr::drop_in_place::<(u32, Struct)> - shim(Some((u32, Struct)))
  expected: local_drop_glue-fallback.cgu[External] 
  actual:   local_drop_glue.volatile[External] 

fn std::ptr::drop_in_place::<Outer> - shim(Some(Outer))
  expected: local_drop_glue-fallback.cgu[External] 
  actual:   local_drop_glue.volatile[External] 

fn std::ptr::drop_in_place::<Struct> - shim(Some(Struct))
  expected: local_drop_glue-fallback.cgu[External] 
  actual:   local_drop_glue.volatile[External] 

fn std::ptr::drop_in_place::<mod1::Struct2> - shim(Some(mod1::Struct2))
  expected: local_drop_glue-fallback.cgu[External] 
  actual:   local_drop_glue-mod1.volatile[External] 


thread '[codegen-units] tests/codegen-units/partitioning/local-drop-glue.rs' panicked at src/tools/compiletest/src/runtest/codegen_units.rs:103:13:
explicit panic

---- [codegen-units] tests/codegen-units/partitioning/vtable-through-const.rs stdout ----

The following items were assigned to wrong codegen units:

fn <mod1::NeedsDrop as std::ops::Drop>::drop
  expected: vtable_through_const-fallback.cgu[External] 
  actual:   vtable_through_const-mod1[External] 

fn std::ptr::drop_in_place::<mod1::NeedsDrop> - shim(Some(mod1::NeedsDrop))
  expected: vtable_through_const-fallback.cgu[External] 
  actual:   vtable_through_const-mod1.volatile[External] 


thread '[codegen-units] tests/codegen-units/partitioning/vtable-through-const.rs' panicked at src/tools/compiletest/src/runtest/codegen_units.rs:103:13:
explicit panic

rust-bors · 2025-06-13T22:25:39Z

☀️ Try build successful (CI)
Build commit: c500484 (c5004842700e4d9f3f326347878229ff0f2dae11, parent: 0d6ab209c525e276cbe7544cbd39a3c3619b6b18)

rust-timer · 2025-06-14T00:54:37Z

Finished benchmarking commit (c500484): comparison URL.

Overall result: ❌✅ regressions and improvements - please read the text below

Benchmarking this pull request means it may be perf-sensitive – we'll automatically label it not fit for rolling up. You can override this, but we strongly advise not to, due to possible changes in compiler perf.

Next Steps: If you can justify the regressions found in this try perf run, please do so in sufficient writing along with @rustbot label: +perf-regression-triaged. If not, please fix the regressions and do another perf run. If its results are neutral or positive, the label will be automatically removed.

@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)	4.6%	[0.3%, 18.0%]	30
Regressions ❌ (secondary)	3.9%	[0.1%, 23.7%]	9
Improvements ✅ (primary)	-1.8%	[-4.1%, -0.3%]	5
Improvements ✅ (secondary)	-2.3%	[-2.8%, -1.8%]	3
All ❌✅ (primary)	3.7%	[-4.1%, 18.0%]	35

Max RSS (memory usage)

Results (primary 0.9%, secondary -5.7%)

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

	mean	range	count
Regressions ❌ (primary)	2.3%	[1.2%, 4.5%]	6
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-3.6%	[-4.6%, -2.5%]	2
Improvements ✅ (secondary)	-5.7%	[-5.7%, -5.7%]	1
All ❌✅ (primary)	0.9%	[-4.6%, 4.5%]	8

Cycles

Results (primary 6.7%, secondary 6.0%)

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

	mean	range	count
Regressions ❌ (primary)	8.5%	[0.8%, 19.6%]	16
Regressions ❌ (secondary)	10.0%	[2.4%, 26.1%]	4
Improvements ✅ (primary)	-2.6%	[-3.9%, -1.8%]	3
Improvements ✅ (secondary)	-2.0%	[-2.3%, -1.8%]	2
All ❌✅ (primary)	6.7%	[-3.9%, 19.6%]	19

Binary size

Results (primary 0.4%, 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)	0.6%	[0.0%, 1.8%]	61
Regressions ❌ (secondary)	5.2%	[0.1%, 14.3%]	9
Improvements ✅ (primary)	-0.3%	[-1.4%, -0.1%]	26
Improvements ✅ (secondary)	-2.4%	[-5.3%, -0.0%]	14
All ❌✅ (primary)	0.4%	[-1.4%, 1.8%]	87

Bootstrap: 755.964s -> 756.21s (0.03%)
Artifact size: 372.26 MiB -> 372.23 MiB (-0.01%)

rustbot added the T-compiler Relevant to the compiler team, which will review and decide on the PR/issue. label May 25, 2025

compiler-errors changed the title ~~More characteristic def ids~~ [PERF] partition shims into CGU of their characteristic types, if possible May 25, 2025