Shrink mir::Statement to 40 bytes by panstromek · Pull Request #159012 · rust-lang/rust

panstromek · 2026-07-09T12:59:06Z

We do this by using (old-style) ThinVec for storing debuginfo.

This collection is almost always empty, so we can optimize for that case. I think we could do better by storing the debuginfo elsewhere, but this is the least invasive change for now.

panstromek · 2026-07-09T13:00:14Z

@bors try @rust-timer queue

shrink mir::Statement to 40 bytes

rust-bors · 2026-07-09T15:10:59Z

☀️ Try build successful (CI)
Build commit: 5aa011a (5aa011acd7e9ca4e63fb318d6e94dbf24e019ec7)
Base parent: 561ea2b (561ea2bfa3022226082e2db39d6ccb241e9e34bc)

rust-timer · 2026-07-09T15:51:33Z

Finished benchmarking commit (5aa011a): 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
@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%	[0.2%, 0.3%]	5
Regressions ❌ (secondary)	0.5%	[0.2%, 0.7%]	4
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-0.1%	[-0.2%, -0.1%]	3
All ❌✅ (primary)	0.3%	[0.2%, 0.3%]	5

Max RSS (memory usage)

Results (primary -1.4%, secondary -2.3%)

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)	2.7%	[2.7%, 2.7%]	1
Improvements ✅ (primary)	-1.4%	[-2.5%, -0.6%]	37
Improvements ✅ (secondary)	-2.5%	[-4.1%, -1.0%]	31
All ❌✅ (primary)	-1.4%	[-2.5%, -0.6%]	37

Cycles

Results (primary -2.9%)

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)	-	-	0
Improvements ✅ (primary)	-2.9%	[-2.9%, -2.9%]	1
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-2.9%	[-2.9%, -2.9%]	1

Binary size

Results (primary 0.0%, secondary 0.0%)

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

	mean	range	count
Regressions ❌ (primary)	0.0%	[0.0%, 0.1%]	2
Regressions ❌ (secondary)	0.0%	[0.0%, 0.1%]	2
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.0%	[0.0%, 0.1%]	2

Bootstrap: 489.147s -> 487.269s (-0.38%)
Artifact size: 388.47 MiB -> 388.44 MiB (-0.01%)

panstromek · 2026-07-09T16:10:29Z

Good Max RSS numbers, but I'm a bit confused by the unicode-normalization icount result. Debug-IncrFull is the one I was testing this on initially and it never actually pushes anything into the debuginfos list. Maybe I'm looking into this incorrectly, I'll try to figure out what's going on there.

panstromek · 2026-07-09T18:23:46Z

Looks like this is just overhead of all the new management code (unicode-normalization truly never allocates anything there). I tried to put a few inlines where it helped based on cachegrind diffs. Let's see if it helps. if not, then I'll probably have to change strategy and move the branching out of the datastructure.

@bors try @rust-timer queue

shrink mir::Statement to 40 bytes

rust-bors · 2026-07-09T20:30:51Z

☀️ Try build successful (CI)
Build commit: cc54650 (cc54650a1dcb449e1c44e013a4ffbf87f8d4895a)
Base parent: 3664b37 (3664b37017c529cad0f0ed259769f15743c21a6a)

rust-timer · 2026-07-09T21:31:03Z

Finished benchmarking commit (cc54650): 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
@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)	0.3%	[0.0%, 0.7%]	4
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-0.2%	[-0.3%, -0.1%]	6
All ❌✅ (primary)	-	-	0

Max RSS (memory usage)

Results (primary -1.3%, secondary -2.4%)

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)	3.5%	[3.5%, 3.5%]	1
Improvements ✅ (primary)	-1.3%	[-2.4%, -0.6%]	29
Improvements ✅ (secondary)	-2.6%	[-4.7%, -0.8%]	30
All ❌✅ (primary)	-1.3%	[-2.4%, -0.6%]	29

Cycles

Results (secondary -4.3%)

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)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-4.3%	[-4.3%, -4.3%]	1
All ❌✅ (primary)	-	-	0

Binary size

Results (primary 0.0%, secondary 0.0%)

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

	mean	range	count
Regressions ❌ (primary)	0.0%	[0.0%, 0.1%]	40
Regressions ❌ (secondary)	0.0%	[0.0%, 0.1%]	26
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.0%	[0.0%, 0.1%]	40

Bootstrap: 487.967s -> 487.075s (-0.18%)
Artifact size: 388.37 MiB -> 388.36 MiB (-0.00%)

panstromek · 2026-07-10T04:04:22Z

This looks better. coercions and include-blob looks like bimodal noise. Max RSS gains persist. I'll clean this up a bit and undraft.

panstromek · 2026-07-10T18:03:52Z

Actually, I want to try one more thing. Let's try to be a lot more aggressive about this inline/outline split. This reduced the overhead a bit in my local cachegrind measurements of unicode-normalization.

@bors try @rust-timer queue

shrink mir::Statement to 40 bytes

rustbot · 2026-07-11T06:20:15Z

r? @JohnTitor

rustbot has assigned @JohnTitor.
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 75 candidates
Random selection from 19 candidates

panstromek · 2026-07-11T06:33:54Z

Part of the icount result looks like noise (especially include-blob), but unicode-normalization and tuple-stress spends a lot of time in mir so at least part of it is probably legit. unicode-normalization also improved in my local cachegrind runs. max-rss improvements persist.

I squashed the commits and added a comment.

JohnTitor

Cool, r=me once addressing the typo above.

View changes since this review

We do this by using old-style ThinVec for storing debuginfos collection, which is almost always empty.

panstromek · 2026-07-11T07:14:31Z

Done. I actually don't know whether I have rights to do this. Let's try.

@bors r=JohnTitor

rust-bors · 2026-07-11T07:14:35Z

@panstromek: 🔑 Insufficient privileges: not in review users

panstromek · 2026-07-11T07:31:06Z

Yea, I can't do this. @JohnTitor can you please r+ this yourself? Sorry for the churn.

JohnTitor · 2026-07-11T11:30:41Z

Oh sorry I thought you had it...
Anyway, @bors r+

rust-bors · 2026-07-11T11:30:45Z

📌 Commit 74dc8e7 has been approved by JohnTitor

It is now in the queue for this repository.

JonathanBrouwer · 2026-07-11T13:11:54Z

@bors rollup=iffy
Will to throw this into a small rollup as the perf effect is recognizable enough

…ller, r=JohnTitor Shrink mir::Statement to 40 bytes We do this by using (old-style) ThinVec for storing debuginfo. This collection is almost always empty, so we can optimize for that case. I think we could do better by storing the debuginfo elsewhere, but this is the least invasive change for now.

…uwer Rollup of 8 pull requests Successful merges: - #159126 (Fix PR number in bootstrap's change tracker) - #156968 (diagnostics: suggest generic_const_args for const ops) - #159012 (Shrink mir::Statement to 40 bytes) - #158182 (std: use `OnceLock` for SGX argument storage) - #159114 (Add regression test for $-prefixed fragment specifier in repetition) - #159060 (Look for the cdb architecture that corresponds to the target triple) - #159089 (riscv: update c-variadic test for LLVM changes) - #159093 (pretty-print: use inline asm's actual macro name)

…uwer Rollup of 14 pull requests Successful merges: - #159126 (Fix PR number in bootstrap's change tracker) - #155811 (Include AtomicU128/AtomicI128 in docs for any target) - #156968 (diagnostics: suggest generic_const_args for const ops) - #159012 (Shrink mir::Statement to 40 bytes) - #156618 (rustdoc: test ignoring rustc lints in CLI) - #158182 (std: use `OnceLock` for SGX argument storage) - #159114 (Add regression test for $-prefixed fragment specifier in repetition) - #158523 (Fix feature gate for `repr(simd)`) - #158876 (Fix multiple logic bugs in `Arc::make_mut`) - #158928 (std: fix Xous UDP send_to length mismatch and truncation) - #159060 (Look for the cdb architecture that corresponds to the target triple) - #159089 (riscv: update c-variadic test for LLVM changes) - #159093 (pretty-print: use inline asm's actual macro name) - #159122 (doc: use ptr::addr in offset_from docs) Failed merges: - #158732 (Apply MCP 1003 and move diagnostics.rs into its own module)

JonathanBrouwer · 2026-07-12T05:20:21Z

@rust-timer build ca5b0d2

rust-timer · 2026-07-12T06:19:33Z

Finished benchmarking commit (ca5b0d2): comparison URL.

Overall result: ❌✅ regressions and 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.

@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)	0.4%	[0.4%, 0.4%]	1
Improvements ✅ (primary)	-0.3%	[-0.3%, -0.3%]	2
Improvements ✅ (secondary)	-0.2%	[-0.2%, -0.2%]	1
All ❌✅ (primary)	-0.3%	[-0.3%, -0.3%]	2

Max RSS (memory usage)

Results (primary -1.3%, secondary -2.4%)

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)	-	-	0
Improvements ✅ (primary)	-1.3%	[-3.1%, -0.5%]	30
Improvements ✅ (secondary)	-2.4%	[-3.9%, -0.9%]	28
All ❌✅ (primary)	-1.3%	[-3.1%, -0.5%]	30

Cycles

Results (secondary -2.7%)

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)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-2.7%	[-2.7%, -2.7%]	1
All ❌✅ (primary)	-	-	0

Binary size

Results (primary 0.0%, secondary 0.0%)

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

	mean	range	count
Regressions ❌ (primary)	0.0%	[0.0%, 0.1%]	2
Regressions ❌ (secondary)	0.0%	[0.0%, 0.1%]	2
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.0%	[0.0%, 0.1%]	2

Bootstrap: 489.809s -> 489.264s (-0.11%)
Artifact size: 389.31 MiB -> 389.32 MiB (0.00%)

rustbot added S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. T-compiler Relevant to the compiler team, which will review and decide on the PR/issue. labels Jul 9, 2026

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 Jul 9, 2026

This comment has been minimized.

Sign in to view

rust-bors Bot pushed a commit that referenced this pull request Jul 9, 2026

Auto merge of #159012 - panstromek:make-stmtdebuginfos-smaller, r=<try>

5aa011a

shrink mir::Statement to 40 bytes

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 Jul 9, 2026

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 Jul 9, 2026

This comment has been minimized.

Sign in to view

rust-bors Bot pushed a commit that referenced this pull request Jul 9, 2026

Auto merge of #159012 - panstromek:make-stmtdebuginfos-smaller, r=<try>

cc54650

shrink mir::Statement to 40 bytes

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 Jul 9, 2026

This comment has been minimized.

Sign in to view

rust-bors Bot pushed a commit that referenced this pull request Jul 10, 2026

Auto merge of #159012 - panstromek:make-stmtdebuginfos-smaller, r=<try>

69e234a

shrink mir::Statement to 40 bytes

rustbot added the S-waiting-on-perf Status: Waiting on a perf run to be completed. label Jul 10, 2026

panstromek marked this pull request as ready for review July 11, 2026 06:20

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 Jul 11, 2026

rustbot assigned JohnTitor Jul 11, 2026

panstromek changed the title ~~shrink mir::Statement to 40 bytes~~ Shrink mir::Statement to 40 bytes Jul 11, 2026

panstromek commented Jul 11, 2026

View reviewed changes

Comment thread compiler/rustc_middle/src/mir/statement.rs Outdated

JohnTitor approved these changes Jul 11, 2026

View reviewed changes

shrink mir::Statement to 40 bytes

74dc8e7

We do this by using old-style ThinVec for storing debuginfos collection, which is almost always empty.

panstromek force-pushed the make-stmtdebuginfos-smaller branch from 053c2b9 to 74dc8e7 Compare July 11, 2026 07:04

rust-bors Bot added S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Jul 11, 2026

JonathanBrouwer mentioned this pull request Jul 11, 2026

Rollup of 8 pull requests #159132

Closed

JonathanBrouwer mentioned this pull request Jul 11, 2026

Rollup of 14 pull requests #159142

Merged

rust-bors Bot merged commit 8c60f8c into rust-lang:main Jul 11, 2026
13 checks passed

rustbot added this to the 1.99.0 milestone Jul 11, 2026

This comment has been minimized.

Sign in to view

Uh oh!

Uh oh!

Conversation

panstromek commented Jul 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

panstromek commented Jul 9, 2026

Uh oh!

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

rust-bors Bot commented Jul 9, 2026

Uh oh!

This comment has been minimized.

rust-timer commented Jul 9, 2026

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

Uh oh!

panstromek commented Jul 9, 2026

Uh oh!

panstromek commented Jul 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

rust-bors Bot commented Jul 9, 2026

Uh oh!

This comment has been minimized.

rust-timer commented Jul 9, 2026

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

Uh oh!

panstromek commented Jul 10, 2026

Uh oh!

panstromek commented Jul 10, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

This comment has been minimized.

This comment has been minimized.

rustbot commented Jul 11, 2026

Uh oh!

panstromek commented Jul 11, 2026

Uh oh!

Uh oh!

JohnTitor 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!

panstromek commented Jul 11, 2026

Uh oh!

rust-bors Bot commented Jul 11, 2026

Uh oh!

panstromek commented Jul 11, 2026

Uh oh!

JohnTitor commented Jul 11, 2026

Uh oh!

rust-bors Bot commented Jul 11, 2026

Uh oh!

JonathanBrouwer commented Jul 11, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

JonathanBrouwer commented Jul 12, 2026

Uh oh!

This comment has been minimized.

rust-timer commented Jul 12, 2026

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

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

panstromek commented Jul 9, 2026 •

edited

Loading

panstromek commented Jul 9, 2026 •

edited

Loading

panstromek commented Jul 10, 2026 •

edited

Loading

JohnTitor left a comment •

edited by rustbot

Loading

JonathanBrouwer commented Jul 11, 2026 •

edited

Loading