Optimize new solver unification table ops - #160801

nnethercote · 2026-08-09T11:08:35Z

The current code uses the ena crate in sub-optimal ways. Improving this gives big speed wins for the new trait solver on some benchmarks. Details in individual commits.

In `is_changed_arg` we do four variations on this (pseudocode-ish) expression: ``` !(probe_value(vid) == Unknown && find(vid) == vid) ``` The first sub-expression means "the var's value is unknown" and the second means "this var is a root". `probe_value` and `find` are both parent-chasing operations, i.e. they duplicate some of their work. We can instead use `try_probe_value` to do this more cheaply: ``` try_probe_value(vid) != Some(Unknown) ``` This has the identical meaning because `try_probe_value` means "get the var's value, but only if it's a root". (Strange but fortuitous!) There are four cases like this in `is_changed_arg`. Three of them (int/float/const vars) are very similar to each other. The ty var case has slightly different syntax but it's the same idea.

In this case we currently have something approximating this: ``` find(vid) != vid ``` which means "is this not a root var?" This commit changes it to basically this: ``` !try_probe_value(vid).is_some() ``` which is equivalent but doesn't require any parent chasing.

nnethercote · 2026-08-09T11:08:50Z

@bors try @rust-timer queue

Optimize new solver unification table ops

rust-bors · 2026-08-09T12:07:52Z

☀️ Try build successful (CI)
Build commit: 6faf9bb (6faf9bb85f7bc21e2a3dd05d1e5defd7b165b4a7)
Base parent: 153ecc4 (153ecc4f74035b709bb3e1eb9546f1d934865042)

rust-timer · 2026-08-09T12:50:36Z

Finished benchmarking commit (6faf9bb): 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 rustc-perf
@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%	[0.2%, 0.5%]	7
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-3.0%	[-6.7%, -0.7%]	10
All ❌✅ (primary)	-	-	0

Max RSS (memory usage)

Results (primary 2.3%, secondary 1.0%)

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

	mean	range	count
Regressions ❌ (primary)	2.3%	[2.3%, 2.4%]	2
Regressions ❌ (secondary)	1.0%	[0.4%, 3.1%]	19
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	2.3%	[2.3%, 2.4%]	2

Cycles

Results (secondary -0.4%)

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)	1.5%	[0.6%, 5.2%]	8
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-1.3%	[-5.0%, -0.4%]	17
All ❌✅ (primary)	-	-	0

Binary size

Results (primary -0.1%, secondary -0.1%)

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)	-0.1%	[-0.1%, -0.0%]	15
Improvements ✅ (secondary)	-0.1%	[-0.1%, -0.0%]	15
All ❌✅ (primary)	-0.1%	[-0.1%, -0.0%]	15

Bootstrap: 458.036s -> 459.953s (0.42%)
Artifact size: 398.61 MiB -> 398.55 MiB (-0.01%)

nnethercote added 2 commits August 9, 2026 20:47

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 Aug 9, 2026

This comment has been minimized.

Sign in to view

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

Auto merge of #160801 - nnethercote:new-solver-probes, r=<try>

6faf9bb

Optimize new solver unification table ops

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 Aug 9, 2026

Navigation Menu

Uh oh!

Uh oh!

Optimize new solver unification table ops - #160801