New solver opts for diesel (part 3) by nnethercote · Pull Request #165160 · rust-lang/rust

nnethercote · 2026-10-29T23:07:09Z

A sequel to #163981 and #164234. These ones mostly relate to allocations. Details in individual commits.

Local measurements:

solver         icounts (B)
------         -----------
old            36.4
new (orig)     44.5 (1.22x)
new (this PR)  42.4 (1.16x)

Also some small improvements on a range of other benchmarks.

r? @jdonszelmann

nnethercote · 2026-10-29T23:07:35Z

@bors try @rust-timer queue

New solver opts for diesel (part 3)

rust-bors · 2026-10-30T00:07:51Z

☀️ Try build successful (CI)
Build commit: 4f26521 (4f26521093dc3d11b878b8f464dbd082553cf1e6)
Base parent: 92b5b8c (92b5b8c87c61b20adc9001b41baf587634a74734)

rust-timer · 2026-10-30T00:56:48Z

Finished benchmarking commit (4f26521): 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)	0.3%	[0.2%, 0.4%]	12
Improvements ✅ (primary)	-0.1%	[-0.2%, -0.1%]	4
Improvements ✅ (secondary)	-1.4%	[-4.4%, -0.1%]	38
All ❌✅ (primary)	-0.1%	[-0.2%, -0.1%]	4

Max RSS (memory usage)

Results (primary -3.9%, secondary 3.9%)

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)	7.4%	[5.0%, 8.6%]	4
Improvements ✅ (primary)	-3.9%	[-3.9%, -3.9%]	1
Improvements ✅ (secondary)	-0.8%	[-0.9%, -0.8%]	3
All ❌✅ (primary)	-3.9%	[-3.9%, -3.9%]	1

Cycles

Results (secondary -1.5%)

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)	2.9%	[1.7%, 4.5%]	6
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-3.7%	[-7.5%, -2.4%]	12
All ❌✅ (primary)	-	-	0

Binary size

Results (primary -0.2%, 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)	-0.2%	[-0.2%, -0.0%]	32
Improvements ✅ (secondary)	-0.1%	[-0.3%, -0.0%]	44
All ❌✅ (primary)	-0.2%	[-0.2%, -0.0%]	32

Bootstrap: 484.843s -> 489.287s (0.92%)
Artifact size: 406.61 MiB -> 409.03 MiB (0.60%)

nnethercote · 2026-10-30T04:38:55Z

@bors try @rust-timer queue

New solver opts for diesel (part 3)

rust-bors · 2026-10-30T05:42:48Z

☀️ Try build successful (CI)
Build commit: 5018116 (50181163cd9f89d9f09e11f71ab14761f7d89ef1)
Base parent: 92b5b8c (92b5b8c87c61b20adc9001b41baf587634a74734)

rust-timer · 2026-10-30T06:32:37Z

Finished benchmarking commit (5018116): 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)	0.2%	[0.1%, 0.4%]	5
Improvements ✅ (primary)	-0.2%	[-0.2%, -0.1%]	8
Improvements ✅ (secondary)	-1.7%	[-4.3%, -0.1%]	32
All ❌✅ (primary)	-0.2%	[-0.2%, -0.1%]	8

Max RSS (memory usage)

Results (primary -4.6%, 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)	-	-	0
Improvements ✅ (primary)	-4.6%	[-4.6%, -4.6%]	1
Improvements ✅ (secondary)	-1.7%	[-2.3%, -1.1%]	2
All ❌✅ (primary)	-4.6%	[-4.6%, -4.6%]	1

Cycles

Results (primary 2.8%, secondary -3.1%)

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

	mean	range	count
Regressions ❌ (primary)	2.8%	[2.8%, 2.8%]	1
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-3.1%	[-4.0%, -2.2%]	13
All ❌✅ (primary)	2.8%	[2.8%, 2.8%]	1

Binary size

This perf run didn't have relevant results for this metric.

Bootstrap: 484.843s -> 482.182s (-0.55%)
Artifact size: 406.61 MiB -> 406.66 MiB (0.01%)

nnethercote · 2026-10-30T07:07:42Z

LLM disclosure: some of the optimization ideas came from LLM analysis. I wrote the code and text myself.

rustbot · 2026-10-30T07:08:00Z

jdonszelmann is currently at their maximum review capacity.
They may take a while to respond.

rustbot · 2026-10-30T07:08:03Z

rust-analyzer is developed in its own repository. If possible, consider making this change to rust-lang/rust-analyzer instead.

cc @rust-lang/rust-analyzer

Some changes occurred to the core trait solver

cc @rust-lang/initiative-trait-system-refactor

rust-bors · 2026-10-30T16:04:42Z

📌 Commit 09ac7ab has been approved by jdonszelmann

It is now in the queue for this repository.

JonathanBrouwer · 2026-10-30T18:25:21Z

@bors rollup=iffy
Making a rollup of perf sensitive prs

@jdonszelmann

…zelmann New solver opts for diesel (part 3) A sequel to rust-lang#163981 and rust-lang#164234. These ones mostly relate to allocations. Details in individual commits. Local measurements: ``` solver icounts (B) ------ ----------- old 36.4 new (orig) 44.5 (1.22x) new (this PR) 42.4 (1.16x) ``` Also some small improvements on a range of other benchmarks. r? @jdonszelmann

…uwer Rollup of 7 perf-sensitive pull requests Successful merges: - #163339 (MIR move elimination [5/6]: Storage reconstruction) - #163525 (Move lang item collection to the resolver) - #163928 (do not prefer trivially true candidates with `-Znext-solver` ) - #164341 (Expand an early return case in `coerce_unsized`) - #164484 (Use adaptive sparse encoding for metadata tables) - #164607 (destringlyfy lint infrastructure) - #165160 (New solver opts for diesel (part 3)) > [!WARNING] > This rollup conflicts with pending auto build #165178 and may need to be recreated if the pending build succeeds.

We still use `ThinVec<I::GenericArg>` in `GoalStalledOn`, where the narrowness of the type is important. But using `Vec<I::GenericArg>` elsewhere is a perf win for `CanonicalizerState::variables`, which has `clone_from` called on it a lot. (`Vec::clone_from` is better optimized than `ThinVec::clone_from`.)

Returning requires `take`, which means the next user of `CanonicalizerState` has to allocate a new vec. `finalize` has two callers, but only one of them needs `variables`. This commit makes it so that single caller just clones the vec out of `CanonicalizerState` itself.

Currently, `InferCtxt` has a `CanonicalizerState` that gets reused by each canonicalizer. This gives a good amount of reuse, but we can increase it by moving the state to `GlobalCaches`. And we actually need a vector of `CanonicalizerState`, though it can only get multiple elements if we are using multiple threads.

`compute_query_response_instantiation_values_in_universe` computes an `opt_values` vector but then only sometimes consults the contents. This commit makes it lazy for a small perf win.

`CanonicalVarValues::instantiate` builds a `var_values` using a closure to generate each element. That closure is passed the under-construction `var_values`, which is unusual. This commit renames the parameters in the closures to `prev_var_values` to make this unusual behaviour more obvious. The commit looks bigger than it really is because of formatting changes.

rustbot · 2026-11-02T02:03:09Z

This PR was rebased onto a different main commit. Here's a range-diff highlighting what actually changed.

Rebasing is a normal part of keeping PRs up to date, so no action is needed—this note is just to help reviewers.

nnethercote · 2026-11-02T02:03:39Z

I rebased.

@bors r=jdonszelmann

rust-bors · 2026-11-02T02:03:43Z

📌 Commit d3dca22 has been tentatively approved by jdonszelmann

It will be put into the queue for this repository once PR CI succeeds.

…uwer Rollup of 2 pull requests Successful merges: - #165140 (More derive(Debug) cleanups for structs) - #165160 (New solver opts for diesel (part 3))

@jdonszelmann

Rollup merge of #165160 - nnethercote:diesel-opts-3, r=jdonszelmann New solver opts for diesel (part 3) A sequel to #163981 and #164234. These ones mostly relate to allocations. Details in individual commits. Local measurements: ``` solver icounts (B) ------ ----------- old 36.4 new (orig) 44.5 (1.22x) new (this PR) 42.4 (1.16x) ``` Also some small improvements on a range of other benchmarks. r? @jdonszelmann

…uwer Rollup of 2 pull requests Successful merges: - rust-lang/rust#165140 (More derive(Debug) cleanups for structs) - rust-lang/rust#165160 (New solver opts for diesel (part 3))

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 Oct 29, 2026

rust-bors Bot pushed a commit that referenced this pull request Oct 29, 2026

Auto merge of #165160 - nnethercote:diesel-opts-3, r=<try>

4f26521

New solver opts for diesel (part 3)

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 Oct 30, 2026

nnethercote force-pushed the diesel-opts-3 branch from 471be56 to 09ac7ab Compare October 30, 2026 04:38

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 Oct 30, 2026

This comment has been minimized.

Sign in to view

rust-bors Bot pushed a commit that referenced this pull request Oct 30, 2026

Auto merge of #165160 - nnethercote:diesel-opts-3, r=<try>

5018116

New solver opts for diesel (part 3)

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 Oct 30, 2026

rustbot assigned jdonszelmann Oct 30, 2026

nnethercote marked this pull request as ready for review October 30, 2026 07:07

rustbot added the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label Oct 30, 2026

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 Oct 30, 2026

JonathanBrouwer mentioned this pull request Oct 30, 2026

Rollup of 7 perf-sensitive pull requests #165205

Closed

rust-bors Bot added S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. and removed S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. labels Oct 31, 2026

This comment has been minimized.

Sign in to view

nnethercote added 5 commits November 2, 2026 12:47

Compute opt_values lazily

ad4a8b5

`compute_query_response_instantiation_values_in_universe` computes an `opt_values` vector but then only sometimes consults the contents. This commit makes it lazy for a small perf win.

nnethercote force-pushed the diesel-opts-3 branch from 09ac7ab to d3dca22 Compare November 2, 2026 02:03

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-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Nov 2, 2026

JonathanBrouwer mentioned this pull request Nov 2, 2026

Rollup of 2 pull requests #165320

Merged

rust-bors Bot merged commit c96d25f into rust-lang:main Nov 2, 2026
14 checks passed

rustbot added this to the 1.101.0 milestone Nov 2, 2026

nnethercote deleted the diesel-opts-3 branch November 2, 2026 10:11

Conversation

nnethercote commented Oct 29, 2026 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

nnethercote commented Oct 29, 2026

Uh oh!

This comment has been minimized.

This comment has been minimized.

rust-bors Bot commented Oct 30, 2026

Uh oh!

This comment has been minimized.

rust-timer commented Oct 30, 2026

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

Uh oh!

nnethercote commented Oct 30, 2026

Uh oh!

This comment has been minimized.

This comment has been minimized.

rust-bors Bot commented Oct 30, 2026

Uh oh!

This comment has been minimized.

rust-timer commented Oct 30, 2026

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

Uh oh!

nnethercote commented Oct 30, 2026

Uh oh!

rustbot commented Oct 30, 2026

Uh oh!

rustbot commented Oct 30, 2026

Uh oh!

rust-bors Bot commented Oct 30, 2026

Uh oh!

JonathanBrouwer commented Oct 30, 2026

Uh oh!

This comment has been minimized.

rustbot commented Nov 2, 2026

Uh oh!

nnethercote commented Nov 2, 2026

Uh oh!

rust-bors Bot commented Nov 2, 2026

Uh oh!

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

nnethercote commented Oct 29, 2026 •

edited by rustbot

Loading