`DepNodeColor` tweaks #147423

nnethercote · 2025-10-07T00:06:20Z

A follow-up to #147293, where I attempted and mostly failed to make things faster again, but I found a few cleanups worth doing.

r? @saethlin

Two functions take arguments of this type, but the `Option` is always `Some`, so we can just pass `&MarkFrame<'_>` instead.

Because it's only ever used for inserting red. (`None` is only used as a starting value, and `Green` is inserted by `try_mark_green` and friends.)

Currently it's binary, either `Green` or `Red`. But it's almost always used within an `Option`. So it's a bit neater, and possibly slightly faster, to make it tri-value with `Unknown` as a first-class variant.

nnethercote · 2025-10-07T00:06:36Z

@bors try @rust-timer queue

`DepNodeColor` tweaks

compiler/rustc_query_system/src/dep_graph/graph.rs

cjgillot · 2025-10-07T01:19:04Z

compiler/rustc_query_system/src/dep_graph/graph.rs

-                DepNodeColor::Red,
-            );
+            colors.insert_red(SerializedDepNodeIndex::from_u32(
+                DepNodeIndex::FOREVER_RED_NODE.as_u32(),


Should we introduce a SERIALIZED_RED constant?

Not sure. Feels orthogonal to this PR.

rust-bors · 2025-10-07T02:21:34Z

☀️ Try build successful (CI)
Build commit: 40c8f48 (40c8f48fa9837632a61cd1bb1dac6118e2d93e1b, parent: 3d8c1c1fc077d04658de63261d8ce2903546db13)

rust-timer · 2025-10-07T03:33:05Z

Finished benchmarking commit (40c8f48): 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)	-0.6%	[-1.5%, -0.2%]	96
Improvements ✅ (secondary)	-0.6%	[-1.1%, -0.2%]	46
All ❌✅ (primary)	-0.6%	[-1.5%, -0.2%]	96

Max RSS (memory usage)

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)	2.4%	[2.4%, 2.4%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

Cycles

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

Binary size

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

Bootstrap: 472.01s -> 471.833s (-0.04%)
Artifact size: 388.39 MiB -> 388.35 MiB (-0.01%)

nnethercote · 2025-10-07T04:15:10Z

Cool, once again the CI results are better than what I saw locally.

The most common `get` case is green. This commit changes `get` to use use `if`/`else` instead of match, so that getting green requires one comparison instead of two.

It uses a different implementation depending on whether the compiler front-end is running single-threaded or multi-threaded. The two implementations are equivalent and I think the multi-threaded one expresses the intent more clearly, and I imagine the perf is similar. So this commit removes the single-threaded code.

nnethercote · 2025-10-07T06:53:56Z

Let's re-run perf to make sure my changes didn't muck things up:

@bors try @rust-timer queue

`DepNodeColor` tweaks

rust-bors · 2025-10-07T09:09:35Z

☀️ Try build successful (CI)
Build commit: bc91772 (bc917725191fb1f5aae04eeb24f8f246cfc5d80f, parent: d773bd07d63a74adcf25ea5f4aae986be94cac5e)

rust-timer · 2025-10-07T10:19:54Z

Finished benchmarking commit (bc91772): 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)	-0.6%	[-1.4%, -0.2%]	100
Improvements ✅ (secondary)	-0.6%	[-1.1%, -0.1%]	42
All ❌✅ (primary)	-0.6%	[-1.4%, -0.2%]	100

Max RSS (memory usage)

Results (secondary 1.7%)

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)	1.7%	[1.4%, 2.0%]	2
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

Cycles

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

Binary size

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

Bootstrap: 473.455s -> 472.567s (-0.19%)
Artifact size: 388.39 MiB -> 388.38 MiB (-0.00%)

nnethercote · 2025-10-07T11:07:03Z

Second CI perf run looks the same as the first. Good to go, I think.

cjgillot · 2025-10-07T13:09:31Z

Thanks!
@bors r+

bors · 2025-10-07T13:09:34Z

📌 Commit 19423b7 has been approved by cjgillot

It is now in the queue for this repository.

`DepNodeColor` tweaks A follow-up to #147293, where I attempted and mostly failed to make things faster again, but I found a few cleanups worth doing. r? `@saethlin`

bors · 2025-10-07T13:32:13Z

⌛ Testing commit 19423b7 with merge 68070eb...

bors · 2025-10-07T14:25:41Z

💔 Test failed - checks-actions

nnethercote · 2025-10-07T20:35:19Z

@bors retry

bors · 2025-10-08T06:28:03Z

⌛ Testing commit 19423b7 with merge 5767910...

bors · 2025-10-08T09:36:19Z

☀️ Test successful - checks-actions
Approved by: cjgillot
Pushing 5767910 to master...

github-actions · 2025-10-08T09:39:24Z

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 4fd3181 (parent) -> 5767910 (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 5767910cbcc9d199bf261a468574d45aa3857599 --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: 7050.9s -> 9219.9s (30.8%)
aarch64-apple: 9726.2s -> 7023.8s (-27.8%)
dist-aarch64-msvc: 6622.2s -> 5349.6s (-19.2%)
dist-apple-various: 3645.5s -> 4041.2s (10.9%)
x86_64-gnu-miri: 4354.7s -> 4798.6s (10.2%)
aarch64-gnu-llvm-20-2: 2437.0s -> 2668.0s (9.5%)
dist-aarch64-apple: 7984.3s -> 7287.0s (-8.7%)
dist-arm-linux-gnueabi: 4949.4s -> 5315.1s (7.4%)
dist-ohos-armv7: 4219.1s -> 3941.5s (-6.6%)
x86_64-rust-for-linux: 2898.8s -> 3086.9s (6.5%)

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 · 2025-10-08T10:47:37Z

Finished benchmarking commit (5767910): 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)	-0.6%	[-1.5%, -0.2%]	98
Improvements ✅ (secondary)	-0.6%	[-1.2%, -0.1%]	47
All ❌✅ (primary)	-0.6%	[-1.5%, -0.2%]	98

Max RSS (memory usage)

Results (primary 0.1%)

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

	mean	range	count
Regressions ❌ (primary)	1.1%	[1.1%, 1.1%]	1
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-0.9%	[-0.9%, -0.9%]	1
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.1%	[-0.9%, 1.1%]	2

Cycles

Results (primary -5.0%, secondary 2.5%)

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)	2.5%	[2.5%, 2.5%]	1
Improvements ✅ (primary)	-5.0%	[-5.0%, -5.0%]	1
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-5.0%	[-5.0%, -5.0%]	1

Binary size

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

Bootstrap: 473.366s -> 473.657s (0.06%)
Artifact size: 388.38 MiB -> 388.42 MiB (0.01%)

nnethercote added 3 commits October 7, 2025 09:32

Avoid passing Option<&MarkFrame<'_>>.

f644f2c

Two functions take arguments of this type, but the `Option` is always `Some`, so we can just pass `&MarkFrame<'_>` instead.

Change DepNodeColorMap::insert to insert_red.

a4162d9

Because it's only ever used for inserting red. (`None` is only used as a starting value, and `Green` is inserted by `try_mark_green` and friends.)

Make DepNodeColor tri-value.

659b758

Currently it's binary, either `Green` or `Red`. But it's almost always used within an `Option`. So it's a bit neater, and possibly slightly faster, to make it tri-value with `Unknown` as a first-class variant.