[experimental] Make witnesses more eager #141762

compiler-errors · 2025-05-30T11:21:33Z

r? lcnr

compiler-errors · 2025-05-30T11:21:40Z

@bors try @rust-timer queue

[experimental] Make witnesses more eager r? lcnr

bors · 2025-05-30T11:22:51Z

⌛ Trying commit 7bff9fb with merge aa7f5f1...

lcnr · 2025-05-30T11:26:50Z

compiler/rustc_trait_selection/src/traits/select/mod.rs

@@ -1508,7 +1508,8 @@ impl<'cx, 'tcx> SelectionContext<'cx, 'tcx> {
                defining_opaque_types_and_generators: defining_opaque_types,
            }
            | TypingMode::Borrowck { defining_opaque_types } => {
-                defining_opaque_types.is_empty() || !pred.has_opaque_types()
+                defining_opaque_types.is_empty()
+                    || (!pred.has_opaque_types() && !pred.has_coroutines())


i feel like we prolly want to instead never erase the coroutines, only the defining opaques when creating a TypingEnv from the current infcx

might be expensive to erase on every call to TypingEnv; otherwise i'll keep this and write a comment explaining why it's unsound.

bors · 2025-05-30T13:46:43Z

☀️ Try build successful - checks-actions
Build commit: aa7f5f1 (aa7f5f1743a191388f6b130b1ea65bcfa321c198)

rust-timer · 2025-05-31T01:08:00Z

Finished benchmarking commit (aa7f5f1): comparison URL.

Overall result: ✅ improvements - no action needed

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.

@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)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-0.2%	[-0.2%, -0.2%]	2
Improvements ✅ (secondary)	-0.4%	[-0.6%, -0.2%]	5
All ❌✅ (primary)	-0.2%	[-0.2%, -0.2%]	2

Max RSS (memory usage)

Results (primary -0.4%, secondary -1.3%)

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)	1.6%	[1.6%, 1.6%]	1
Regressions ❌ (secondary)	0.5%	[0.4%, 0.5%]	3
Improvements ✅ (primary)	-1.4%	[-1.6%, -1.3%]	2
Improvements ✅ (secondary)	-1.6%	[-3.9%, -0.6%]	17
All ❌✅ (primary)	-0.4%	[-1.6%, 1.6%]	3

Cycles

Results (secondary -0.1%)

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
Regressions ❌ (secondary)	1.0%	[0.4%, 2.4%]	9
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-1.4%	[-3.7%, -0.5%]	8
All ❌✅ (primary)	-	-	0

Binary size

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

Bootstrap: missing data
Artifact size: 370.26 MiB -> 370.23 MiB (-0.01%)

compiler-errors · 2025-06-09T17:00:24Z

@bors2 try @rust-timer queue

rust-bors · 2025-06-09T17:00:27Z

⌛ Trying commit 5862717 with merge 9676755…

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

[experimental] Make witnesses more eager r? lcnr

rust-bors · 2025-06-09T17:19:16Z

💔 Test failed

CI ❌

compiler-errors · 2025-06-09T20:53:19Z

@bors2 try @rust-timer queue

rust-bors · 2025-06-09T20:53:23Z

⌛ Trying commit ec8dc45 with merge c5243c3…

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

[experimental] Make witnesses more eager r? lcnr

rust-bors · 2025-06-09T23:23:28Z

☀️ Try build successful (CI)
Build commit: c5243c3 (c5243c3809f813c22efa750866301fd67942ef25)

rust-timer · 2025-06-10T01:54:23Z

Finished benchmarking commit (c5243c3): comparison URL.

Overall result: ❌ regressions - 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)	0.5%	[0.4%, 0.6%]	8
Regressions ❌ (secondary)	0.8%	[0.1%, 1.8%]	20
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-0.8%	[-0.9%, -0.6%]	3
All ❌✅ (primary)	0.5%	[0.4%, 0.6%]	8

Max RSS (memory usage)

Results (secondary 1.3%)

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)	1.9%	[1.2%, 3.5%]	5
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-2.2%	[-2.2%, -2.2%]	1
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: 754.885s -> 753.548s (-0.18%)
Artifact size: 372.30 MiB -> 372.33 MiB (0.01%)

compiler-errors · 2025-06-10T02:02:21Z

I think filtering out the opaques from the typing env (for pseudo-canonical inputs) is pretty expensive. Let's see how perf is when using the old strategy to compare.

Maybe we could optimize the "filter out opaques" step, but the global caching may really just not be worth it.

@bors2 try @rust-timer queue

rust-bors · 2025-06-10T02:02:25Z

⌛ Trying commit 5f09e4f with merge 1d115c1…

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

[experimental] Make witnesses more eager r? lcnr

rust-bors · 2025-06-10T04:29:07Z

☀️ Try build successful (CI)
Build commit: 1d115c1 (1d115c1f1702fc6a1d322f845583581dd636f0a4)

rust-timer · 2025-06-10T07:49:35Z

Finished benchmarking commit (1d115c1): 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
Improvements ✅ (secondary)	-0.6%	[-0.6%, -0.6%]	2
All ❌✅ (primary)	-	-	0

Max RSS (memory usage)

Results (primary -0.7%, secondary 1.6%)

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)	1.6%	[1.0%, 2.8%]	5
Improvements ✅ (primary)	-1.6%	[-1.7%, -1.5%]	2
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-0.7%	[-1.7%, 1.1%]	3

Cycles

Results (secondary -1.6%)

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)	-	-	0
Improvements ✅ (secondary)	-1.6%	[-1.6%, -1.6%]	1
All ❌✅ (primary)	-	-	0

Binary size

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

Bootstrap: 755.17s -> 754.366s (-0.11%)
Artifact size: 372.30 MiB -> 372.28 MiB (-0.00%)

compiler-errors · 2025-06-13T18:08:35Z

I think I'd like to land this in the state it is, and then later putting up a PR to remove coroutine witnesses altogether. It's worth checking crater, too.

@craterbot check

craterbot · 2025-06-13T18:08:57Z

👌 Experiment pr-141762 created and queued.
🤖 Automatically detected try build 1d115c1
🔍 You can check out the queue and this experiment's details.

ℹ️ Crater is a tool to run experiments across parts of the Rust ecosystem. Learn more

craterbot · 2025-06-13T18:09:39Z

🚧 Experiment pr-141762 is now running

ℹ️ Crater is a tool to run experiments across parts of the Rust ecosystem. Learn more

craterbot · 2025-06-14T23:53:53Z

🎉 Experiment pr-141762 is completed!
📊 3 regressed and 7 fixed (647393 total)
📰 Open the full report.

⚠️ If you notice any spurious failure please add them to the denylist!
ℹ️ Crater is a tool to run experiments across parts of the Rust ecosystem. Learn more

rustbot assigned lcnr May 30, 2025

rustbot added T-compiler WG-trait-system-refactor labels May 30, 2025