Use `DeepRejectCtxt` to quickly reject `ParamEnv` candidates #128776

Bryanskiy · 2024-08-07T11:08:06Z

The description is on the zulip thread

lcnr · 2024-08-07T13:23:24Z

compiler/rustc_next_trait_solver/src/solve/trait_goals.rs

+                {
+                    return Err(NoSolution);
+                }
+


please also add this for NormalizesTo goals

lcnr · 2024-08-07T13:33:30Z

compiler/rustc_type_ir/src/fast_reject.rs

+    pub fn types_may_unify(self, lhs: I::Ty, rhs: I::Ty) -> bool {
+        // Non rigid types may unify with each other.
+        if !lhs.is_known_rigid() && !rhs.is_known_rigid() {
+            return true;
        }


this means we never reject Param(T) == Param(U)

when treating params as rigid, compare their index

lcnr

and also in project::assemble_candidates_from_predicates to also use a fast-path when normalizing in the old solver

lcnr · 2024-08-09T09:03:41Z

compiler/rustc_type_ir/src/fast_reject.rs

+        match ty.kind() {
+            ty::Param(p) => match (self.treat_lhs_params, self.treat_rhs_params) {
+                (TreatParams::AsRigid, TreatParams::AsRigid) => param == p,
+                _ => true,


please exhaustively match here

lcnr · 2024-08-09T09:08:00Z

compiler/rustc_type_ir/src/fast_reject.rs

+            (ty::Placeholder(lhs), ty::Placeholder(rhs)) => lhs == rhs,
+            (ty::Placeholder(_), _) | (_, ty::Placeholder(_)) => false,
+
+            _ => false,


this function has to be an overapproximation, so incorrectly returning false would result in subtle bugs. Please exhaustively match here: (ty::Whatever | ty::Whatever | ...., _) | (_, ty::whatever | .....)

lcnr · 2024-08-09T13:59:14Z

compiler/rustc_type_ir/src/fast_reject.rs

+                    (TreatParams::InstantiateWithInfer, TreatParams::AsRigid)
+                    | (TreatParams::AsRigid, TreatParams::InstantiateWithInfer)
+                    | (TreatParams::InstantiateWithInfer, TreatParams::InstantiateWithInfer) => {


Suggested change

(TreatParams::InstantiateWithInfer, TreatParams::AsRigid)

| (TreatParams::AsRigid, TreatParams::InstantiateWithInfer)

| (TreatParams::InstantiateWithInfer, TreatParams::InstantiateWithInfer) => {

(TreatParams::InstantiateWithInfer, _)

| (_, TreatParams::InstantiateWithInfer) => {

small style preference

lcnr · 2024-08-09T14:01:24Z

compiler/rustc_type_ir/src/fast_reject.rs

-            },
-            ty::Pat(obl_ty, _) => {
+            }
+            (ty::Param(_), ty::Placeholder(_)) | (ty::Placeholder(_), ty::Param(_)) => true,


not really relevant as it doesnt ever happen on stable, but ty::Param and placeholder can't unify when treating params as rigid, so you can remove this branch and just use the self.treat_lhs_params == TreatParams::InstantiateWithInfer branch

lcnr · 2024-08-09T14:03:47Z

@bors try @rust-timer queue

Use `DeepRejectCtxt` to quickly reject `ParamEnv` candidates The description is on the [zulip thread](https://rust-lang.zulipchat.com/#narrow/stream/144729-t-types/topic/.5Basking.20for.20help.5D.20.60DeepRejectCtxt.60.20for.20param.20env.20candidates) r? `@lcnr`

bors · 2024-08-09T14:06:10Z

⌛ Trying commit 77fc853 with merge 7a0b60e...

bors · 2024-08-09T16:00:17Z

☀️ Try build successful - checks-actions
Build commit: 7a0b60e (7a0b60e0de7dfef4ed18ad00b864a144909632d3)

rust-timer · 2024-08-09T18:37:55Z

Finished benchmarking commit (7a0b60e): comparison URL.

Overall result: ❌✅ regressions and improvements - 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.

Next Steps: If you can justify the regressions found in this try perf run, please indicate this with @rustbot label: +perf-regression-triaged along with sufficient written justification. If you cannot justify the regressions please fix the regressions and do another perf run. If the next run shows neutral or positive results, the label will be automatically removed.

@bors rollup=never
@rustbot label: -S-waiting-on-perf +perf-regression

Instruction count

This is a highly reliable metric that was used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	6.3%	[0.2%, 12.7%]	18
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-1.8%	[-4.2%, -0.3%]	15
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	2.6%	[-4.2%, 12.7%]	33

Max RSS (memory usage)

Results (primary -3.7%, secondary 2.9%)

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)	2.9%	[2.9%, 2.9%]	1
Improvements ✅ (primary)	-3.7%	[-5.4%, -2.6%]	3
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-3.7%	[-5.4%, -2.6%]	3

Cycles

Results (primary 2.3%)

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)	4.5%	[3.3%, 7.4%]	13
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-1.8%	[-2.5%, -1.4%]	7
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	2.3%	[-2.5%, 7.4%]	20

Binary size

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

Bootstrap: 759.636s -> 760.059s (0.06%)
Artifact size: 337.10 MiB -> 337.13 MiB (0.01%)

lcnr

2 ideas for perf

move all (Rigid, _) | (_, Rigid) => false into a single branch at the end

and if that doesn't fix the regressions make the deep reject ctxt instead be generic over TreatParams 🤔

Bryanskiy · 2024-08-12T11:20:58Z

A small statistical observation:

bitmaps-3.1.0 benchmark (12.70%):

successful rejects in assemble_candidates_from_caller_bounds: 10/58 (~17%)
successful rejects in project: 1/1032 (~0.1%)

typenum-1.17.0 benchmark (9.59%):

successful rejects in assemble_candidates_from_caller_bounds: 7286/12526 (~58%)
successful rejects in project: 115/518 (~22%)

diesel-1.4.8 benchmark (-4.17%):

successful rejects in assemble_candidates_from_caller_bounds: 1405482/1558566 (~90%)
successful rejects in project: 159641/168826 (~95%)

std:

successful rejects in assemble_candidates_from_caller_bounds: 212125/285199 (~74%)
successful rejects in project: 5754/25544 (~22%)

We waste time more often in project and we can try to remove DeepRejectCtxt from there.

lcnr · 2024-08-12T13:44:36Z

the impact on diesel seems quite desirable however. I think we can improve the performance of fast reject significantly so that even the ~20% hit rate from typenum is a perf improvement.

lcnr · 2024-08-12T13:45:30Z

@bors try @rust-timer queue

bors · 2024-08-12T13:46:44Z

⌛ Trying commit ea66e5b with merge bb463aa...

Use `DeepRejectCtxt` to quickly reject `ParamEnv` candidates The description is on the [zulip thread](https://rust-lang.zulipchat.com/#narrow/stream/144729-t-types/topic/.5Basking.20for.20help.5D.20.60DeepRejectCtxt.60.20for.20param.20env.20candidates) r? `@lcnr`

bors · 2024-09-03T10:37:23Z

⌛ Testing commit c51953f with merge 06ab993...

rust-log-analyzer · 2024-09-03T12:17:52Z

The job x86_64-msvc-ext failed! Check out the build log: (web) (plain)

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

[RUSTC-TIMING] miri test:false 4.428
error: failed to remove file `C:\a\rust\rust\build\x86_64-pc-windows-msvc\stage1-tools\x86_64-pc-windows-msvc\release\miri.exe`

Caused by:
  Access is denied. (os error 5)
Command has failed. Rerun with -v to see more details.
  local time: Tue, Sep  3, 2024 12:17:37 PM
  network time: Tue, 03 Sep 2024 12:17:37 GMT
##[error]Process completed with exit code 1.
Post job cleanup.

bors · 2024-09-03T12:19:56Z

💔 Test failed - checks-actions

lqd · 2024-09-03T12:21:03Z

@bors retry

bors · 2024-09-03T20:35:29Z

⌛ Testing commit c51953f with merge c190066...

Use `DeepRejectCtxt` to quickly reject `ParamEnv` candidates The description is on the [zulip thread](https://rust-lang.zulipchat.com/#narrow/stream/144729-t-types/topic/.5Basking.20for.20help.5D.20.60DeepRejectCtxt.60.20for.20param.20env.20candidates) r? `@lcnr`

rust-log-analyzer · 2024-09-03T22:13:45Z

The job x86_64-msvc-ext failed! Check out the build log: (web) (plain)

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

[RUSTC-TIMING] miri test:false 4.501
error: failed to remove file `C:\a\rust\rust\build\x86_64-pc-windows-msvc\stage1-tools\x86_64-pc-windows-msvc\release\miri.exe`

Caused by:
  Access is denied. (os error 5)
Command has failed. Rerun with -v to see more details.
  local time: Tue, Sep  3, 2024 10:13:31 PM
  network time: Tue, 03 Sep 2024 22:13:31 GMT
##[error]Process completed with exit code 1.
Post job cleanup.

bors · 2024-09-03T22:14:25Z

💔 Test failed - checks-actions

lcnr · 2024-09-04T07:22:52Z

failed to remove file C:\a\rust\rust\build\x86_64-pc-windows-msvc\stage1-tools\x86_64-pc-windows-msvc\release\miri.exe
2024-09-03T22:13:31.6487555Z
2024-09-03T22:13:31.6487659Z Caused by:
2024-09-03T22:13:31.6487920Z Access is denied. (os error 5)

I think that's also a spurious error? cc @rust-lang/infra

Kobzol · 2024-09-04T07:24:37Z

Sadly, yes.

@bors retry

bors · 2024-09-06T19:50:52Z

⌛ Testing commit c51953f with merge 26b5599...

bors · 2024-09-06T22:13:34Z

☀️ Test successful - checks-actions
Approved by: lcnr
Pushing 26b5599 to master...

rust-timer · 2024-09-06T23:29:21Z

Finished benchmarking commit (26b5599): comparison URL.

Overall result: ✅ improvements - no action needed

@rustbot label: -perf-regression

Instruction count

This is a highly reliable metric that was used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-1.8%	[-4.7%, -0.3%]	17
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-1.8%	[-4.7%, -0.3%]	17

Max RSS (memory usage)

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

Cycles

Results (primary -1.8%)

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)	-	-	0
Improvements ✅ (primary)	-1.8%	[-2.2%, -1.5%]	7
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-1.8%	[-2.2%, -1.5%]	7

Binary size

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

Bootstrap: 754.723s -> 756.217s (0.20%)
Artifact size: 341.14 MiB -> 341.14 MiB (0.00%)

Relevant upstream PR: rust-lang/rust#128776: The `TreatParams` enum variants were renamed. Resolves #3503 By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 and MIT licenses.

rustbot assigned lcnr Aug 7, 2024

lcnr reviewed Aug 7, 2024

View reviewed changes

Bryanskiy force-pushed the deep-reject-ctxt branch from 5520ecc to a36b90f Compare August 8, 2024 18:45

lcnr reviewed Aug 9, 2024

View reviewed changes

Bryanskiy force-pushed the deep-reject-ctxt branch from a36b90f to 77fc853 Compare August 9, 2024 13:33

lcnr reviewed Aug 9, 2024

View reviewed changes

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, 2024

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, 2024

lcnr reviewed Aug 12, 2024

View reviewed changes

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 12, 2024

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 2, 2024

lcnr mentioned this pull request Sep 3, 2024

privacy: normalize associated types before visiting #126076

Open

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 3, 2024

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 3, 2024

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 3, 2024

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 4, 2024

bors added the merged-by-bors This PR was explicitly merged by bors. label Sep 6, 2024

bors merged commit 26b5599 into rust-lang:master Sep 6, 2024
7 checks passed

rustbot added this to the 1.83.0 milestone Sep 6, 2024

zhassan-aws mentioned this pull request Sep 8, 2024

Upgrade toolchain to 2024-09-07 model-checking/kani#3504

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Use `DeepRejectCtxt` to quickly reject `ParamEnv` candidates #128776

Use `DeepRejectCtxt` to quickly reject `ParamEnv` candidates #128776

Bryanskiy commented Aug 7, 2024

lcnr Aug 7, 2024

Bryanskiy Aug 8, 2024

lcnr Aug 7, 2024 •

edited

Loading

Bryanskiy Aug 8, 2024

lcnr left a comment

lcnr Aug 9, 2024

lcnr Aug 9, 2024

lcnr Aug 9, 2024

lcnr Aug 9, 2024

lcnr commented Aug 9, 2024

This comment has been minimized.

bors commented Aug 9, 2024

bors commented Aug 9, 2024

This comment has been minimized.

rust-timer commented Aug 9, 2024

lcnr left a comment

Bryanskiy commented Aug 12, 2024

lcnr commented Aug 12, 2024 •

edited

Loading

lcnr commented Aug 12, 2024

This comment has been minimized.

bors commented Aug 12, 2024

bors commented Sep 3, 2024

rust-log-analyzer commented Sep 3, 2024

bors commented Sep 3, 2024

lqd commented Sep 3, 2024

bors commented Sep 3, 2024

rust-log-analyzer commented Sep 3, 2024

bors commented Sep 3, 2024

lcnr commented Sep 4, 2024 •

edited

Loading

Kobzol commented Sep 4, 2024

bors commented Sep 6, 2024

bors commented Sep 6, 2024

rust-timer commented Sep 6, 2024

Use DeepRejectCtxt to quickly reject ParamEnv candidates #128776

Use DeepRejectCtxt to quickly reject ParamEnv candidates #128776

Conversation

Bryanskiy commented Aug 7, 2024

Choose a reason for hiding this comment

Choose a reason for hiding this comment

lcnr Aug 7, 2024 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

lcnr left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

lcnr commented Aug 9, 2024

This comment has been minimized.

bors commented Aug 9, 2024

bors commented Aug 9, 2024

This comment has been minimized.

rust-timer commented Aug 9, 2024

Overall result: ❌✅ regressions and improvements - ACTION NEEDED

lcnr left a comment

Choose a reason for hiding this comment

Bryanskiy commented Aug 12, 2024

lcnr commented Aug 12, 2024 • edited Loading

lcnr commented Aug 12, 2024

This comment has been minimized.

bors commented Aug 12, 2024

bors commented Sep 3, 2024

rust-log-analyzer commented Sep 3, 2024

bors commented Sep 3, 2024

lqd commented Sep 3, 2024

bors commented Sep 3, 2024

rust-log-analyzer commented Sep 3, 2024

bors commented Sep 3, 2024

lcnr commented Sep 4, 2024 • edited Loading

Kobzol commented Sep 4, 2024

bors commented Sep 6, 2024

bors commented Sep 6, 2024

rust-timer commented Sep 6, 2024

Overall result: ✅ improvements - no action needed

Use `DeepRejectCtxt` to quickly reject `ParamEnv` candidates #128776

Use `DeepRejectCtxt` to quickly reject `ParamEnv` candidates #128776

lcnr Aug 7, 2024 •

edited

Loading

lcnr commented Aug 12, 2024 •

edited

Loading

lcnr commented Sep 4, 2024 •

edited

Loading