Add #[rustc_box] and use it inside alloc #97293

est31 · 2022-05-22T21:53:04Z

This commit adds an alternative content boxing syntax, and uses it inside alloc.

#![feature(box_syntax)]

fn foo() {
    let foo = box bar;
}

is equivalent to

#![feature(rustc_attrs)]

fn foo() {
    let foo = #[rustc_box] Box::new(bar);
}

The usage inside the very performance relevant code in
liballoc is the only remaining relevant usage of box syntax
in the compiler (outside of tests, which are comparatively easy to port).

box syntax was originally designed to be used by all Rust
developers. This introduces a replacement syntax more tailored
to only being used inside the Rust compiler, and with it,
lays the groundwork for eventually removing box syntax.

Earlier work by @nbdd0121 to lower Box::new to box during THIR -> MIR building ran into borrow checker problems, requiring the lowering to be adjusted in a way that led to performance regressions. The proposed change in this PR lowers #[rustc_box] Box::new -> box in the AST -> HIR lowering step, which is way earlier in the compiler, and thus should cause less issues both performance wise as well as regarding type inference/borrow checking/etc. Hopefully, future work can move the lowering further back in the compiler, as long as there are no performance regressions.

rust-highfive · 2022-05-22T21:53:06Z

Hey! It looks like you've submitted a new PR for the library teams!

If this PR contains changes to any rust-lang/rust public library APIs then please comment with r? rust-lang/libs-api @rustbot label +T-libs-api -T-libs to request review from a libs-api team reviewer. If you're unsure where your change falls no worries, just leave it as is and the reviewer will take a look and make a decision to forward on if necessary.

Examples of T-libs-api changes:

Stabilizing library features
Introducing insta-stable changes such as new implementations of existing stable traits on existing stable types
Introducing new or changing existing unstable library APIs (excluding permanently unstable features / features without a tracking issue)
Changing public documentation in ways that create new stability guarantees
Changing observable runtime behavior of library APIs

rust-highfive · 2022-05-22T21:53:07Z

r? @kennytm

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

est31 · 2022-05-22T21:53:29Z

Oh right and I'd like a perf run for this.

@bors try

bors · 2022-05-22T21:53:37Z

⌛ Trying commit 21b8f4c5732fa39970597845a14656dd92d5aa81 with merge c40cfb8bd5f0f25521b77d628fc502e48839db62...

compiler-errors · 2022-05-22T23:37:57Z

@bors try @rust-timer queue

rust-timer · 2022-05-22T23:37:59Z

Awaiting bors try build completion.

@rustbot label: +S-waiting-on-perf

bors · 2022-05-22T23:38:05Z

⌛ Trying commit b3da9d276364534550d8f71c5b6f1c21d983b2ac with merge 7a4b510ef15bc589ea430b0bd23134b71f4e25d2...

bors · 2022-05-23T01:28:41Z

☀️ Try build successful - checks-actions
Build commit: 7a4b510ef15bc589ea430b0bd23134b71f4e25d2 (7a4b510ef15bc589ea430b0bd23134b71f4e25d2)

rust-timer · 2022-05-23T01:28:44Z

Queued 7a4b510ef15bc589ea430b0bd23134b71f4e25d2 with parent b2eed72, future comparison URL.

rust-timer · 2022-05-23T02:48:56Z

Finished benchmarking commit (7a4b510ef15bc589ea430b0bd23134b71f4e25d2): comparison url.

Instruction count

Primary benchmarks: no relevant changes found
Secondary benchmarks: 😿 relevant regressions found

	Regressions 😿 (primary)	Regressions 😿 (secondary)	Improvements 🎉 (primary)	Improvements 🎉 (secondary)	All 😿 🎉 (primary)
count¹	0	7	0	0	0
mean²	N/A	0.6%	N/A	N/A	N/A
max	N/A	0.9%	N/A	N/A	N/A

Max RSS (memory usage)

Results

Primary benchmarks: 🎉 relevant improvements found
Secondary benchmarks: mixed results

	Regressions 😿 (primary)	Regressions 😿 (secondary)	Improvements 🎉 (primary)	Improvements 🎉 (secondary)	All 😿 🎉 (primary)
count¹	0	1	3	7	3
mean²	N/A	3.7%	-3.0%	-1.6%	-3.0%
max	N/A	3.7%	-3.6%	-3.0%	-3.6%

Cycles

Results

Primary benchmarks: 🎉 relevant improvements found
Secondary benchmarks: 😿 relevant regression found

	Regressions 😿 (primary)	Regressions 😿 (secondary)	Improvements 🎉 (primary)	Improvements 🎉 (secondary)	All 😿 🎉 (primary)
count¹	0	1	2	0	2
mean²	N/A	1.7%	-2.4%	N/A	-2.4%
max	N/A	1.7%	-2.6%	N/A	-2.6%

If you disagree with this performance assessment, please file an issue in rust-lang/rustc-perf.

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-review -S-waiting-on-perf -perf-regression

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

est31 · 2022-05-23T11:06:46Z

Hmm so the only serious regression is in deep-vector, which is basically a single vec! call with a lot of zeroes inside. Otherwise it's all improvements. And the deep-vector regression is not that bad, considering that it used to be 20% in the earlier PR:
#87781 (comment)

Instruction count, with regressions only in deep-vector:

Max RSS, a lot of improvements with regressions in one place:

So this is IMO quite a good result.

est31 · 2022-05-23T22:17:51Z

The weird thing is that these changes are still larger than I expected. After all, the differences are only at the start of the compilation pipeline. So regressions in later phases, which these mostly are, are a bit surprising.

One might think that the ucd library has the memory usage regression because of the many tables that it contains. And indeed it has many tables. But there is not a single usage of the vec! macro inside that library, instead it's statics... Which might mean that it's indeed due to the compiler having become more memory consuming for this specific workload, instead of passing different input to the compiler due to differences in the vec! macro.

For deep-vector, it's regressing in the same passes that it regressed back in the older attempt that just used Box::new plus a MIR pass at the end of the pipeline to recognize and special case it. But thankfully the regressions are much lower now.

Really bad deep-vector regressions in the older attempt

Not so bad deep-vector regressions in the newer attempt upthread:

The weird part is that this PR does not change anything what is being fed to LLVM. So why does LLVM take longer? No idea. I think I need to create a local build of the compiler as of this PR and play around with it to verify that the MIR etc. is all the same. Edit: plus, given that it's LLVM, there is also barely any Rust running in that pass so it also can't be some general issue of the rust-written part compiler being slower now, because it uses the (hypothetically) now-slower vec! or something like that.

est31 · 2022-05-23T23:49:31Z

I've got an idea to narrow the performance changes down: what if one only changes the Box::new implementation? Requesting another perf run for this. And then one might do one with only vec using #[rustc_box].

compiler-errors · 2022-05-24T00:17:18Z

@bors try @rust-timer queue

rust-timer · 2022-05-24T00:17:19Z

Awaiting bors try build completion.

@rustbot label: +S-waiting-on-perf

bors · 2022-05-24T00:17:27Z

⌛ Trying commit 704845b4efe169fa881bef052db59d163c74d433 with merge 6f102539baaab210a6d285cc923d5d22e4959ac0...

est31 · 2022-05-31T22:31:23Z

Alright, the last commit is removed and pushed to a separate branch for possible future use. The PR is ready for review.

compiler/rustc_ast_lowering/src/expr.rs

nbdd0121 · 2022-05-31T22:46:19Z

compiler/rustc_ast_lowering/src/expr.rs

-                            hir::ExprKind::Box(self.lower_expr(&inner))
+                            let kind = hir::ExprKind::Box(self.lower_expr(&inner));
+                            let hir_id = self.lower_node_id(e.id);
+                            self.lower_attrs(hir_id, &e.attrs[1..]);


So this assumes that rustc_box is the first attribute? If rustc_box is assumed to be the only attribute, then we could just skip lowering attributes at all.

My reasoning for assuming it was first was that rustc_box shouldn't appear so often and I wouldn't want to do a possibly expensive search over all attributes. Once we have rustc_box, it should be comparatively inexpensive to take care of the other attributes.

(When making that decision, I thought about doc attributes but function calls obviously don't have doc attributes so I was a bit misguided)

This commit adds an alternative content boxing syntax, and uses it inside alloc. The usage inside the very performance relevant code in liballoc is the only remaining relevant usage of box syntax in the compiler (outside of tests, which are comparatively easy to port). box syntax was originally designed to be used by all Rust developers. This introduces a replacement syntax more tailored to only being used inside the Rust compiler, and with it, lays the groundwork for eventually removing box syntax.

est31 · 2022-06-01T16:46:12Z

r? @oli-obk

oli-obk · 2022-06-02T10:27:35Z

I'm surprised it wasn't possible to do on the THIR. Looking at the links you provided it looks like they didn't use a #[rustc_box] attribute. I'm fine merging this PR, but I think after it we should try again to move it to THIR -> MIR and eliminate the box keyword from HIR, too

oli-obk · 2022-06-02T10:28:14Z

@bors r+

bors · 2022-06-02T10:28:15Z

📌 Commit 0a24b94 has been approved by oli-obk

bors · 2022-06-02T13:20:22Z

⌛ Testing commit 0a24b94 with merge 20976ba...

bors · 2022-06-02T16:04:34Z

☀️ Test successful - checks-actions
Approved by: oli-obk
Pushing 20976ba to master...

rust-timer · 2022-06-02T17:22:19Z

Finished benchmarking commit (20976ba): comparison url.

Instruction count

Primary benchmarks: no relevant changes found
Secondary benchmarks: 😿 relevant regression found

	mean¹	max	count²
Regressions 😿 (primary)	N/A	N/A	0
Regressions 😿 (secondary)	0.2%	0.2%	1
Improvements 🎉 (primary)	N/A	N/A	0
Improvements 🎉 (secondary)	N/A	N/A	0
All 😿🎉 (primary)	N/A	N/A	0

Max RSS (memory usage)

Results

Primary benchmarks: no relevant changes found
Secondary benchmarks: mixed results

	mean¹	max	count²
Regressions 😿 (primary)	N/A	N/A	0
Regressions 😿 (secondary)	2.8%	3.8%	4
Improvements 🎉 (primary)	N/A	N/A	0
Improvements 🎉 (secondary)	-2.7%	-3.4%	4
All 😿🎉 (primary)	N/A	N/A	0

Cycles

Results

Primary benchmarks: no relevant changes found
Secondary benchmarks: mixed results

	mean¹	max	count²
Regressions 😿 (primary)	N/A	N/A	0
Regressions 😿 (secondary)	2.7%	2.7%	1
Improvements 🎉 (primary)	N/A	N/A	0
Improvements 🎉 (secondary)	-3.1%	-3.1%	1
All 😿🎉 (primary)	N/A	N/A	0

If you disagree with this performance assessment, please file an issue in rust-lang/rustc-perf.

@rustbot label: -perf-regression

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

rust-highfive assigned kennytm May 22, 2022

rustbot added T-compiler Relevant to the compiler team, which will review and decide on the PR/issue. T-libs Relevant to the library team, which will review and decide on the PR/issue. labels May 22, 2022

rust-highfive added the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label May 22, 2022

est31 mentioned this pull request May 22, 2022

Tracking issue for box_syntax #49733

Closed

This comment has been minimized.

Sign in to view

est31 force-pushed the remove_box branch from 21b8f4c to b3da9d2 Compare May 22, 2022 22:10

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 May 22, 2022

rustbot removed the S-waiting-on-perf Status: Waiting on a perf run to be completed. label May 23, 2022

est31 force-pushed the remove_box branch from b3da9d2 to 763cf1b Compare May 23, 2022 16:55

rustbot added the S-waiting-on-perf Status: Waiting on a perf run to be completed. label May 24, 2022

est31 force-pushed the remove_box branch from 432ffad to 0256331 Compare May 31, 2022 22:30

nbdd0121 reviewed May 31, 2022

View reviewed changes

est31 added 2 commits June 1, 2022 02:28

Use #[rustc_box] in alloc instead of box syntax

535e28b

est31 force-pushed the remove_box branch from 0256331 to 48f3757 Compare June 1, 2022 00:42

Remove #[rustc_box] attr during lowering

0a24b94

est31 force-pushed the remove_box branch from 48f3757 to 0a24b94 Compare June 1, 2022 00:44

rust-highfive assigned oli-obk and unassigned kennytm Jun 1, 2022

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 Jun 2, 2022

bors added the merged-by-bors This PR was explicitly merged by bors. label Jun 2, 2022

bors merged commit 20976ba into rust-lang:master Jun 2, 2022

rustbot added this to the 1.63.0 milestone Jun 2, 2022

bors mentioned this pull request Jun 2, 2022

Remove NtExpr. #96724

Closed

petrochenkov mentioned this pull request Jun 6, 2022

[WIP] Also lint Box::new in unused_allocation #97774

Closed

fbstj mentioned this pull request Jun 9, 2022

rustdoc: make the Type enum smaller #97929

Closed

est31 mentioned this pull request Aug 7, 2022

Remove even more box syntax uses from src/test #100194

Merged

est31 mentioned this pull request Jan 18, 2023

Remove unused #![feature(box_syntax)] in alloc #106992

Merged

est31 mentioned this pull request Feb 26, 2023

Remove box_syntax #108471

Merged

est31 mentioned this pull request Apr 23, 2023

Moving #[rustc_box] to move_val_init intrinsic #110715

Closed

oyvindln mentioned this pull request Mar 3, 2024

Avoid using default in HashBuffers::reset Frommi/miniz_oxide#147

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add #[rustc_box] and use it inside alloc #97293

Add #[rustc_box] and use it inside alloc #97293

est31 commented May 22, 2022 •

edited

Loading

rust-highfive commented May 22, 2022

rust-highfive commented May 22, 2022

est31 commented May 22, 2022

bors commented May 22, 2022

This comment has been minimized.

This comment has been minimized.

compiler-errors commented May 22, 2022

rust-timer commented May 22, 2022

bors commented May 22, 2022

bors commented May 23, 2022

rust-timer commented May 23, 2022

rust-timer commented May 23, 2022

est31 commented May 23, 2022

est31 commented May 23, 2022 •

edited

Loading

est31 commented May 23, 2022

compiler-errors commented May 24, 2022

rust-timer commented May 24, 2022

bors commented May 24, 2022

est31 commented May 31, 2022

nbdd0121 May 31, 2022

est31 Jun 1, 2022

est31 Jun 1, 2022

est31 commented Jun 1, 2022

oli-obk commented Jun 2, 2022

oli-obk commented Jun 2, 2022

bors commented Jun 2, 2022

bors commented Jun 2, 2022

bors commented Jun 2, 2022

rust-timer commented Jun 2, 2022

Add #[rustc_box] and use it inside alloc #97293

Add #[rustc_box] and use it inside alloc #97293

Conversation

est31 commented May 22, 2022 • edited Loading

rust-highfive commented May 22, 2022

rust-highfive commented May 22, 2022

est31 commented May 22, 2022

bors commented May 22, 2022

This comment has been minimized.

This comment has been minimized.

compiler-errors commented May 22, 2022

rust-timer commented May 22, 2022

bors commented May 22, 2022

bors commented May 23, 2022

rust-timer commented May 23, 2022

rust-timer commented May 23, 2022

Footnotes

est31 commented May 23, 2022

est31 commented May 23, 2022 • edited Loading

est31 commented May 23, 2022

compiler-errors commented May 24, 2022

rust-timer commented May 24, 2022

bors commented May 24, 2022

est31 commented May 31, 2022

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

est31 commented Jun 1, 2022

oli-obk commented Jun 2, 2022

oli-obk commented Jun 2, 2022

bors commented Jun 2, 2022

bors commented Jun 2, 2022

bors commented Jun 2, 2022

rust-timer commented Jun 2, 2022

Footnotes

est31 commented May 22, 2022 •

edited

Loading

est31 commented May 23, 2022 •

edited

Loading