Avoid invalidating CFG caches from MirPatch::apply. #146697

cjgillot · 2025-09-18T01:11:03Z

Small effort to reduce invalidating CFG caches.

rustbot · 2025-09-18T01:11:06Z

Some changes occurred to MIR optimizations

cc @rust-lang/wg-mir-opt

rustbot · 2025-09-18T01:11:08Z

rustbot has assigned @lcnr.
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

lcnr · 2025-09-19T10:06:27Z

r=me after perf

@bors try @rust-timer queue

Avoid invalidating CFG caches from MirPatch::apply.

rust-bors · 2025-09-19T12:19:02Z

☀️ Try build successful (CI)
Build commit: 37ed41a (37ed41a84266b30ffc58df946ea5338241662949, parent: 2f4dfc753fd86c672aa4145940db075a8a149f17)

rust-timer · 2025-09-19T13:28:09Z

Finished benchmarking commit (37ed41a): comparison URL.

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

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.

@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)	0.2%	[0.2%, 0.2%]	5
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-0.1%	[-0.1%, -0.1%]	1
All ❌✅ (primary)	-	-	0

Max RSS (memory usage)

Results (secondary -3.2%)

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)	-	-	0
Improvements ✅ (secondary)	-3.9%	[-4.5%, -2.5%]	9
All ❌✅ (primary)	-	-	0

Cycles

Results (secondary 3.4%)

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)	3.4%	[2.7%, 4.4%]	4
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

Binary size

Results (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
Regressions ❌ (secondary)	0.0%	[0.0%, 0.0%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

Bootstrap: 470.95s -> 471.912s (0.20%)
Artifact size: 389.99 MiB -> 387.95 MiB (-0.52%)

lcnr · 2025-09-19T20:39:00Z

compiler/rustc_mir_transform/src/patch.rs

            body.basic_blocks.len()
        );
-        let bbs = if self.term_patch_map.is_empty() && self.new_blocks.is_empty() {
+        let bbs = if self.term_patch_map.iter().all(Option::is_none) && self.new_blocks.is_empty() {


so this small perf regression might be real because we're now iterating over all basic blocks here even if all are None 🤔

🤷 would still be fine with landing this, but we may also consider alternatives, e.g. I would kinda assume that the patch map will be mostly None all the time, so maybe something that's better with sparse entries would perform better here.

anyways, haven't looked much at the relevant code yet

cjgillot · 2025-09-20T14:35:14Z

@bors try @rust-timer queue

Avoid invalidating CFG caches from MirPatch::apply.

rust-bors · 2025-09-20T16:48:10Z

☀️ Try build successful (CI)
Build commit: 5b81809 (5b8180916cd08f43db7975b4bf3491d8735ebdf8, parent: 9f2ef0f14d6028c5108643cafa6e2c617834594b)

rust-timer · 2025-09-20T17:57:34Z

Finished benchmarking commit (5b81809): comparison URL.

Overall result: ✅ improvements - no action needed

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.

@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
Improvements ✅ (primary)	-0.2%	[-0.2%, -0.1%]	7
Improvements ✅ (secondary)	-0.2%	[-0.4%, -0.1%]	12
All ❌✅ (primary)	-0.2%	[-0.2%, -0.1%]	7

Max RSS (memory usage)

Results (primary 0.3%, secondary -0.8%)

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

	mean	range	count
Regressions ❌ (primary)	1.7%	[1.0%, 2.1%]	3
Regressions ❌ (secondary)	0.9%	[0.9%, 0.9%]	1
Improvements ✅ (primary)	-1.8%	[-2.8%, -0.8%]	2
Improvements ✅ (secondary)	-2.5%	[-2.5%, -2.5%]	1
All ❌✅ (primary)	0.3%	[-2.8%, 2.1%]	5

Cycles

Results (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)	-	-	0
Improvements ✅ (secondary)	-2.4%	[-2.8%, -2.0%]	2
All ❌✅ (primary)	-	-	0

Binary size

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

Bootstrap: 471.07s -> 470.492s (-0.12%)
Artifact size: 389.99 MiB -> 389.88 MiB (-0.03%)

compiler/rustc_mir_transform/src/patch.rs

lcnr

hm, we might be able to get another 0.1-0.2% here by doing some counts based profiling looking at how many entries end up in the term_patch_map and then e.g. use some SsoHashMap with an inline len of 1 or sth istead of a HashMap. I expect that slotting in the current SsoHashMap would actually slow things down as TerminatorKind is too large.

r=me after adding the comment

View changes since this review

Co-authored-by: lcnr <rust@lcnr.de>

cjgillot · 2025-09-24T02:31:08Z

@bors r=lcnr

I'll try SsoHashMap in a follow up PR.

bors · 2025-09-24T02:31:11Z

📌 Commit ce677c7 has been approved by lcnr

It is now in the queue for this repository.

Zalathar · 2025-09-25T10:34:05Z

Scheduling: interleave rollup=never PRs with rollups.

@bors p=5

Avoid invalidating CFG caches from MirPatch::apply. Small effort to reduce invalidating CFG caches.

bors · 2025-09-25T13:48:53Z

⌛ Testing commit ce677c7 with merge ef5b084...

bors · 2025-09-25T14:02:13Z

💔 Test failed - checks-actions

rust-log-analyzer · 2025-09-25T14:04:20Z

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

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

##[endgroup]
Image input checksum 4ef84ab3b8a59556ac09337f9862a9844bfdfe25dfeff8792afdf568f1697d3819e6cfc0ed7069216c6f4229b9812f4c5544fa54c7a6e6d57bef45de3f3e3a8f
##[group]Building docker image for dist-sparcv9-solaris
Docker version 28.0.4, build b8034c0
Error response from daemon: Get "https://ghcr.io/v2/": Get "https://ghcr.io/token?account=rust-lang&client_id=docker&offline_token=true&service=ghcr.io": context deadline exceeded (Client.Timeout exceeded while awaiting headers) (Client.Timeout exceeded while awaiting headers)
##[error]Process completed with exit code 1.
Post job cleanup.

Zalathar · 2025-09-25T14:15:59Z

@bors retry (flaky)

bors · 2025-09-25T17:19:33Z

⌛ Testing commit ce677c7 with merge eabf390...

bors · 2025-09-25T20:34:41Z

☀️ Test successful - checks-actions
Approved by: lcnr
Pushing eabf390 to master...

github-actions · 2025-09-25T20:37:57Z

What is this?

This is an experimental post-merge analysis report that shows differences in test outcomes between the merged PR and its parent PR.

Comparing 6f34f4e (parent) -> eabf390 (this PR)

Test differences

Show 2 test diffs

2 doctest diffs were found. These are ignored, as they are noisy.

Test dashboard

Run

cargo run --manifest-path src/ci/citool/Cargo.toml -- \
    test-dashboard eabf390b4ceeb34db9f37e97f435134abbcdea92 --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

dist-x86_64-apple: 8055.4s -> 6766.5s (-16.0%)
aarch64-gnu: 7150.7s -> 6102.5s (-14.7%)
pr-check-1: 1562.6s -> 1334.6s (-14.6%)
aarch64-apple: 7594.2s -> 6536.8s (-13.9%)
i686-gnu-nopt-1: 8244.0s -> 7321.3s (-11.2%)
dist-aarch64-msvc: 6666.6s -> 5941.2s (-10.9%)
i686-gnu-2: 6131.0s -> 5492.0s (-10.4%)
x86_64-gnu-stable: 7524.4s -> 6790.8s (-9.8%)
aarch64-gnu-llvm-20-1: 3689.8s -> 3330.6s (-9.7%)
x86_64-gnu-debug: 7384.8s -> 6695.6s (-9.3%)

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.

rust-timer · 2025-09-25T21:44:19Z

Finished benchmarking commit (eabf390): comparison URL.

Overall result: ✅ improvements - no action needed

@rustbot label: -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)	-0.2%	[-0.2%, -0.1%]	3
Improvements ✅ (secondary)	-0.3%	[-0.4%, -0.1%]	10
All ❌✅ (primary)	-0.2%	[-0.2%, -0.1%]	3

Max RSS (memory usage)

Results (primary 1.0%, secondary 0.8%)

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

	mean	range	count
Regressions ❌ (primary)	1.0%	[1.0%, 1.0%]	1
Regressions ❌ (secondary)	1.6%	[0.7%, 2.0%]	4
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-2.5%	[-2.5%, -2.5%]	1
All ❌✅ (primary)	1.0%	[1.0%, 1.0%]	1

Cycles

Results (primary -2.6%)

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)	-2.6%	[-2.6%, -2.6%]	1
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-2.6%	[-2.6%, -2.6%]	1

Binary size

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

Bootstrap: 471.619s -> 470.601s (-0.22%)
Artifact size: 388.14 MiB -> 388.11 MiB (-0.01%)

Avoid invalidating from MirPatch::apply.

b216cf3

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. T-compiler Relevant to the compiler team, which will review and decide on the PR/issue. labels Sep 18, 2025

rustbot assigned lcnr Sep 18, 2025

cjgillot changed the title ~~Avoid invalidating from MirPatch::apply.~~ Avoid invalidating CFG caches from MirPatch::apply. Sep 18, 2025

This comment has been minimized.

Sign in to view

rust-bors bot added a commit that referenced this pull request Sep 19, 2025

Auto merge of #146697 - cjgillot:invalidate-patch, r=<try>

37ed41a

Avoid invalidating CFG caches from MirPatch::apply.

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 Sep 19, 2025

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 Sep 19, 2025

lcnr reviewed Sep 19, 2025

View reviewed changes

Make term_patch_map sparse.

3c232fe

This comment has been minimized.

Sign in to view

rust-bors bot added a commit that referenced this pull request Sep 20, 2025

Auto merge of #146697 - cjgillot:invalidate-patch, r=<try>

5b81809

Avoid invalidating CFG caches from MirPatch::apply.

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 Sep 20, 2025

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 Sep 20, 2025

lcnr reviewed Sep 22, 2025

View reviewed changes

compiler/rustc_mir_transform/src/patch.rs Show resolved Hide resolved

lcnr reviewed Sep 22, 2025

View reviewed changes

Update compiler/rustc_mir_transform/src/patch.rs

ce677c7

Co-authored-by: lcnr <rust@lcnr.de>

bors 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 Sep 24, 2025

bors added a commit that referenced this pull request Sep 25, 2025

Auto merge of #146697 - cjgillot:invalidate-patch, r=lcnr

ef5b084

Avoid invalidating CFG caches from MirPatch::apply. Small effort to reduce invalidating CFG caches.

bors added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. labels Sep 25, 2025

bors 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 Sep 25, 2025

bors added the merged-by-bors This PR was explicitly merged by bors. label Sep 25, 2025

bors merged commit eabf390 into rust-lang:master Sep 25, 2025
11 checks passed

rustbot added this to the 1.92.0 milestone Sep 25, 2025

cjgillot deleted the invalidate-patch branch September 25, 2025 23:49

cjgillot mentioned this pull request Sep 26, 2025

Use SsoHashMap in MirPatch. #147052

Closed

Avoid invalidating CFG caches from MirPatch::apply. #146697

Avoid invalidating CFG caches from MirPatch::apply. #146697

Conversation

cjgillot commented Sep 18, 2025

Uh oh!

rustbot commented Sep 18, 2025

Uh oh!

rustbot commented Sep 18, 2025

Uh oh!

lcnr commented Sep 19, 2025

Uh oh!

This comment has been minimized.

This comment has been minimized.

rust-bors bot commented Sep 19, 2025

Uh oh!

This comment has been minimized.

rust-timer commented Sep 19, 2025

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

Uh oh!

Choose a reason for hiding this comment

Uh oh!

cjgillot commented Sep 20, 2025

Uh oh!

This comment has been minimized.

This comment has been minimized.

rust-bors bot commented Sep 20, 2025

Uh oh!

This comment has been minimized.

rust-timer commented Sep 20, 2025

Overall result: ✅ improvements - no action needed

Uh oh!

Uh oh!

lcnr left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

cjgillot commented Sep 24, 2025

Uh oh!

bors commented Sep 24, 2025

Uh oh!

Zalathar commented Sep 25, 2025

Uh oh!

bors commented Sep 25, 2025

Uh oh!

bors commented Sep 25, 2025

Uh oh!

rust-log-analyzer commented Sep 25, 2025

Uh oh!

Zalathar commented Sep 25, 2025

Uh oh!

bors commented Sep 25, 2025

Uh oh!

bors commented Sep 25, 2025

Uh oh!

Uh oh!

github-actions bot commented Sep 25, 2025

Test differences

Job duration changes

Uh oh!

rust-timer commented Sep 25, 2025

Overall result: ✅ improvements - no action needed

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

7 participants

lcnr left a comment •

edited

Loading