Remove `TypeckResults` from `InferCtxt` #101632

camsteffen · 2022-09-09T20:30:42Z

InferCtxt currently has in_progress_typeck_results which is only used for some diagnostics during typeck. It adds a lifetime which propagates through a lot of code. This PR moves that field into a new helper struct TypeErrCtxt.

rust-highfive · 2022-09-09T20:30:45Z

r? @oli-obk

(rust-highfive has picked a reviewer for you, use r? to override)

rustbot · 2022-09-09T20:30:45Z

Some changes occurred in src/tools/clippy

cc @rust-lang/clippy

Some changes occurred in const_evaluatable.rs

cc @lcnr

Some changes occurred in need_type_info.rs

cc @lcnr

bors · 2022-09-11T00:02:16Z

☔ The latest upstream changes (presumably #98559) made this pull request unmergeable. Please resolve the merge conflicts.

compiler/rustc_infer/src/infer/error_reporting/mod.rs

compiler/rustc_typeck/src/check/inherited.rs

lcnr

really happy about this PR as it is something i've wanted to do myself for quite a while ❤️

some nits and further suggestions, but this looks pretty good already

compiler/rustc_infer/src/infer/mod.rs

compiler/rustc_mir_build/src/thir/pattern/const_to_pat.rs

compiler/rustc_infer/src/infer/combine.rs

lcnr · 2022-09-12T09:29:36Z

r? @lcnr

camsteffen · 2022-09-20T13:59:53Z

I don't know where to start with the failing rustdoc tests 😤

bors · 2022-09-20T22:04:45Z

☔ The latest upstream changes (presumably #102061) made this pull request unmergeable. Please resolve the merge conflicts.

lcnr · 2022-09-21T08:29:35Z

🤔 i also can't tell why rustdoc is failing 😅 maybe look at the actually generated docs for the smallest failing test to check whether there are any user-visible changes? 🤔

camsteffen · 2022-09-29T00:28:12Z

~~Looks like a bit of waiting and doing nothing made the problem go away!~~ nope

camsteffen · 2022-09-29T00:28:40Z

@rustbot ready

src/librustdoc/clean/blanket_impl.rs

compiler/rustc_hir_analysis/src/check/fn_ctxt/mod.rs

lcnr · 2022-10-03T09:45:05Z

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

@@ -63,6 +66,37 @@ pub struct ImplCandidate<'tcx> {
 }

 pub trait InferCtxtExt<'tcx> {


should these be on TypeErrCtxt instead?

Maybe? I was being conservative, only moving stuff to TypeErrCtxt if it needs TypeckResults.

compiler/rustc_trait_selection/src/traits/error_reporting/on_unimplemented.rs

src/tools/clippy/clippy_lints/src/future_not_send.rs

lcnr · 2022-10-03T10:03:29Z

small nits which we can fix in a followup pr

prone to merge conflicts:
@bors r+ p=5 rollup=never

bors · 2022-10-03T10:03:30Z

📌 Commit d68bfb9012f16c29c52dea118439dc1c8d4354a5 has been approved by lcnr

It is now in the queue for this repository.

bors · 2022-10-07T13:54:59Z

⌛ Testing commit 283abbf with merge 43c22af...

bors · 2022-10-07T17:37:33Z

☀️ Test successful - checks-actions
Approved by: lcnr
Pushing 43c22af to master...

rust-timer · 2022-10-07T18:58:45Z

Finished benchmarking commit (43c22af): comparison URL.

Overall result: ❌✅ regressions and improvements - ACTION NEEDED

Next Steps: If you can justify the regressions found in this perf run, please indicate this with @rustbot label: +perf-regression-triaged along with sufficient written justification. If you cannot justify the regressions please open an issue or create a new PR that fixes the regressions, add a comment linking to the newly created issue or PR, and then add the perf-regression-triaged label to this PR.

@rustbot label: +perf-regression
cc @rust-lang/wg-compiler-performance

Instruction count

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

	mean¹	range	count²
Regressions ❌ (primary)	0.7%	[0.5%, 1.2%]	13
Regressions ❌ (secondary)	4.3%	[3.2%, 5.7%]	6
Improvements ✅ (primary)	-0.3%	[-0.6%, -0.2%]	19
Improvements ✅ (secondary)	-0.6%	[-1.6%, -0.2%]	52
All ❌✅ (primary)	0.1%	[-0.6%, 1.2%]	32

Max RSS (memory usage)

Results

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.

	mean¹	range	count²
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	3.5%	[2.1%, 4.5%]	4
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-2.6%	[-2.9%, -2.3%]	2
All ❌✅ (primary)	-	-	0

Cycles

Results

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.

	mean¹	range	count²
Regressions ❌ (primary)	2.4%	[2.4%, 2.4%]	1
Regressions ❌ (secondary)	3.4%	[3.4%, 3.4%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-3.4%	[-3.4%, -3.4%]	1
All ❌✅ (primary)	2.4%	[2.4%, 2.4%]	1

the arithmetic mean of the percent change ↩ ↩² ↩³
number of relevant changes ↩ ↩² ↩³

lcnr · 2022-10-10T04:25:58Z

oh wow, should have definitely run perf before approving this 😅

i guess this resulted in a lot of small inlining/optimization changes, might be worth to check out the flamegraph comparison of a few regressions to check if we can maybe fix them by putting #[inline] somewhere.

rylev · 2022-10-11T15:44:26Z

@lcnr @camsteffen I ran callgrind diff on one of the benchmarks. Looks specialization_graph::Children::insert is getting called way more.

--------------------------------------------------------------------------------
Ir           file:function
--------------------------------------------------------------------------------
 94,311,966  ???:rustc_trait_selection::traits::coherence::overlapping_impls::<<rustc_middle::traits::specialization_graph::Children as rustc_trait_selection::traits::specialize::specialization_graph::ChildrenExt>::insert::{closure
-66,019,056  ???:<rustc_middle::traits::specialization_graph::Children as rustc_trait_selection::traits::specialize::specialization_graph::ChildrenExt>::insert
  7,957,920  ???:__rust_probestack
 -7,904,085  ???:<rustc_infer::infer::InferCtxtBuilder>::enter::<core::result::Result<alloc::vec::Vec<rustc_middle::ty::Predicate>, rustc_errors::ErrorGuaranteed>, rustc_trait_selection::traits::do_normalize_predicates::{closure
  6,887,397  ???:rustc_trait_selection::traits::do_normalize_predicates
  4,716,702  ???:<rustc_infer::infer::InferCtxt>::probe::<(), <rustc_trait_selection::traits::select::SelectionContext>::assemble_candidates_from_impls::{closure
 -4,586,815  ???:<rustc_trait_selection::traits::select::SelectionContext>::match_impl
  4,052,492  ???:<rustc_infer::infer::InferCtxtBuilder>::build
 -3,625,123  /build/glibc-eX1tMB/glibc-2.31/string/../sysdeps/x86_64/multiarch/memmove-vec-unaligned-erms.S:__memcpy_avx_unaligned_erms

camsteffen · 2022-10-11T19:30:47Z

Thanks @rylev. I think it might be traits::overlapping_impls that needs inline? I see that function in the callgrind diff, and the changes to that function might have made it not inlined anymore.

rylev · 2022-10-11T19:46:30Z

@camsteffen the symbol is a bit hard to parse. It seems like it's not the function but possibly a closure inside of that function? Might require some experimentation.

camsteffen · 2022-10-11T20:35:03Z

Probably this closure, and if overlapping_impls is inlined, then the closure would probably be inlined as well?

camsteffen · 2022-10-13T20:20:28Z

bitmaps perf addressed in #102931. Looks like deeply-nested-multi could use another fix.

nnethercote · 2022-10-13T20:32:17Z

deeply-nested-multi is less important, plus it has been noisy lately, so don't spend too much time on it.

camsteffen · 2022-10-13T20:34:46Z

Thanks @nnethercote. I'll gladly ignore that.

… r=cjgillot Make `overlapping_impls` not generic Trying to win back perf from rust-lang#101632.

rust-highfive assigned oli-obk Sep 9, 2022

rustbot added the T-compiler Relevant to the compiler team, which will review and decide on the PR/issue. label Sep 9, 2022

rust-highfive added the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label Sep 9, 2022

lcnr reviewed Sep 12, 2022

View reviewed changes

compiler/rustc_infer/src/infer/error_reporting/mod.rs Outdated Show resolved Hide resolved

compiler/rustc_infer/src/infer/error_reporting/mod.rs Outdated Show resolved Hide resolved

compiler/rustc_typeck/src/check/inherited.rs Outdated Show resolved Hide resolved

lcnr reviewed Sep 12, 2022

View reviewed changes

rust-highfive assigned lcnr and unassigned oli-obk Sep 12, 2022

lcnr 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-review Status: Awaiting review from the assignee but also interested parties. labels Sep 12, 2022

camsteffen force-pushed the refactor-infer-err branch 2 times, most recently from 4c43c71 to ce7f171 Compare September 20, 2022 13:42

This comment has been minimized.

Sign in to view

camsteffen force-pushed the refactor-infer-err branch from ce7f171 to 6da620c Compare September 29, 2022 00:24

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Sep 29, 2022

This comment has been minimized.

Sign in to view

camsteffen force-pushed the refactor-infer-err branch from 6da620c to d68bfb9 Compare October 1, 2022 16:39

camsteffen commented Oct 1, 2022

View reviewed changes

src/librustdoc/clean/blanket_impl.rs Show resolved Hide resolved

lcnr reviewed Oct 3, 2022

View reviewed changes

bors added the S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. label Oct 7, 2022

bors added the merged-by-bors This PR was explicitly merged by bors. label Oct 7, 2022

bors merged commit 43c22af into rust-lang:master Oct 7, 2022

rustbot added this to the 1.66.0 milestone Oct 7, 2022

camsteffen deleted the refactor-infer-err branch October 7, 2022 17:38

rustbot added the perf-regression Performance regression. label Oct 7, 2022

camsteffen mentioned this pull request Oct 11, 2022

Make overlapping_impls not generic #102931

Merged

bors added a commit to rust-lang-ci/rust that referenced this pull request Oct 16, 2022

Auto merge of rust-lang#102931 - camsteffen:inline-overlapping-impls,…

75dbd5b

… r=cjgillot Make `overlapping_impls` not generic Trying to win back perf from rust-lang#101632.

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Remove `TypeckResults` from `InferCtxt` #101632

Remove `TypeckResults` from `InferCtxt` #101632

camsteffen commented Sep 9, 2022 •

edited

Loading

rust-highfive commented Sep 9, 2022

rustbot commented Sep 9, 2022

bors commented Sep 11, 2022

lcnr left a comment

lcnr commented Sep 12, 2022

camsteffen commented Sep 20, 2022

This comment has been minimized.

bors commented Sep 20, 2022

lcnr commented Sep 21, 2022

camsteffen commented Sep 29, 2022 •

edited

Loading

camsteffen commented Sep 29, 2022

This comment has been minimized.

lcnr Oct 3, 2022

camsteffen Oct 3, 2022

lcnr commented Oct 3, 2022

bors commented Oct 3, 2022

bors commented Oct 7, 2022

bors commented Oct 7, 2022

rust-timer commented Oct 7, 2022

lcnr commented Oct 10, 2022

rylev commented Oct 11, 2022

camsteffen commented Oct 11, 2022

rylev commented Oct 11, 2022

camsteffen commented Oct 11, 2022

camsteffen commented Oct 13, 2022

nnethercote commented Oct 13, 2022

camsteffen commented Oct 13, 2022

		@@ -63,6 +66,37 @@ pub struct ImplCandidate<'tcx> {
		}

		pub trait InferCtxtExt<'tcx> {

Remove TypeckResults from InferCtxt #101632

Remove TypeckResults from InferCtxt #101632

Conversation

camsteffen commented Sep 9, 2022 • edited Loading

rust-highfive commented Sep 9, 2022

rustbot commented Sep 9, 2022

bors commented Sep 11, 2022

lcnr left a comment

Choose a reason for hiding this comment

lcnr commented Sep 12, 2022

camsteffen commented Sep 20, 2022

This comment has been minimized.

bors commented Sep 20, 2022

lcnr commented Sep 21, 2022

camsteffen commented Sep 29, 2022 • edited Loading

camsteffen commented Sep 29, 2022

This comment has been minimized.

lcnr Oct 3, 2022

Choose a reason for hiding this comment

camsteffen Oct 3, 2022

Choose a reason for hiding this comment

lcnr commented Oct 3, 2022

bors commented Oct 3, 2022

bors commented Oct 7, 2022

bors commented Oct 7, 2022

rust-timer commented Oct 7, 2022

Overall result: ❌✅ regressions and improvements - ACTION NEEDED

Instruction count

Max RSS (memory usage)

Cycles

Footnotes

lcnr commented Oct 10, 2022

rylev commented Oct 11, 2022

camsteffen commented Oct 11, 2022

rylev commented Oct 11, 2022

camsteffen commented Oct 11, 2022

camsteffen commented Oct 13, 2022

nnethercote commented Oct 13, 2022

camsteffen commented Oct 13, 2022

Remove `TypeckResults` from `InferCtxt` #101632

Remove `TypeckResults` from `InferCtxt` #101632

camsteffen commented Sep 9, 2022 •

edited

Loading

camsteffen commented Sep 29, 2022 •

edited

Loading