optimize slice::Iter::fold #106343

the8472 · 2023-01-01T04:24:07Z

Fixes 2 of 4 cases from #106288

OLD: test slice::fold_to_last                                           ... bench:         248 ns/iter (+/- 3)
NEW: test slice::fold_to_last                                           ... bench:           0 ns/iter (+/- 0)

rustbot · 2023-01-01T04:24:13Z

r? @joshtriplett

(rustbot has picked a reviewer for you, use r? to override)

rustbot · 2023-01-01T04:24:15Z

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 @rustbot label +T-libs-api -T-libs to tag it appropriately. If this PR contains changes to any unstable APIs please edit the PR description to add a link to the relevant API Change Proposal or create one if you haven't already. 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

the8472 · 2023-01-01T04:24:24Z

@bors try @rust-timer queue

bors · 2023-01-01T04:24:33Z

⌛ Trying commit 6a6a41907748388572966d5f214133a8b4efefd5 with merge 3fad16f536f1c5c74f8d8c80d98c3ea988f1e33b...

bors · 2023-01-01T04:28:23Z

💔 Test failed - checks-actions

the8472 · 2023-01-01T04:36:56Z

@bors try

bors · 2023-01-01T04:37:04Z

⌛ Trying commit e42a42b95ba5c8c3a30fc24badcae06bb0c53a3f with merge d8fe6702c1e876077d9db9bdfe7b8756d8d7de6e...

bors · 2023-01-01T06:44:53Z

☀️ Try build successful - checks-actions
Build commit: d8fe6702c1e876077d9db9bdfe7b8756d8d7de6e (d8fe6702c1e876077d9db9bdfe7b8756d8d7de6e)

the8472 · 2023-01-02T04:26:46Z

@bors try @rust-timer queue

jyn514 · 2023-06-13T21:47:40Z

i don't think i'm a good reviewer for libs PRs 😅 the old code was simpler iirc

r? @scottmcm

scottmcm · 2023-06-14T05:01:42Z

library/core/src/slice/iter/macros.rs

+                loop {
+                    // SAFETY: the loop iterates `i in 0..len`, which always is in bounds of
+                    // the slice allocation
+                    acc = f(acc, unsafe { & $( $mut_ )? *self.ptr.add(i).as_ptr() });


How much of the duplication here is essential?

Avoiding the Option loop seems entirely reasonable to me, but how does this approach compare to something much shorter? For example, something like

for _ in 0..len!(self) { acc = f(acc, next_unchecked!(self)); }

Or the same with a while (n --> 0) loop or something, if the Range is not ok.

And if that's not sufficient, it would be nice to have some comments here about why it's written this way.

(Also, if overloading fold is worth it, I expect rfold should be overridden too.)

@rustbot author

If you look at the PR history you'll see that I tried several approaches. A for-in-range loop was one of them. I even used IndexRange instead of Range.

And if that's not sufficient, it would be nice to have some comments here about why it's written this way.

Ok, will do.

(Also, if overloading fold is worth it, I expect rfold should be overridden too.)

Maybe, but that'd be optimizing two things at once which makes assessing perf more complicated.

Added a comment.

@rustbot ready

this seems to produce less IR

scottmcm · 2023-06-15T05:44:51Z

library/core/src/slice/iter/macros.rs

+            {
+                // this implementation consists of the following optimizations compared to the
+                // default implementation:
+                // - do-while loop, as is llvm's preferred loop shape,


It's very weird to me that doing this manually is needed, since LLVM has a loop rotation pass to do this. But I guess it's fine.

scottmcm · 2023-06-15T05:45:50Z

@bors r+ rollup=never

bors · 2023-06-15T05:45:52Z

📌 Commit d90508f has been approved by scottmcm

It is now in the queue for this repository.

klensy · 2023-06-15T08:52:01Z

tests/codegen/vec-shrink-panik.rs

-    // CHECK-NOT: panic
-
-    // Call to panic_cannot_unwind in case of double-panic is expected,
-    // on LLVM 16 and older, but other panics are not.
-    // old: filter
-    // old-NEXT: ; call core::panicking::panic_cannot_unwind
-    // old-NEXT: panic_cannot_unwind
-


Now there panic? Or why this was killed.
Ohh, i see it lower, my bad.

bors · 2023-06-15T09:38:56Z

⌛ Testing commit d90508f with merge 4996b56...

bors · 2023-06-15T13:02:41Z

☀️ Test successful - checks-actions
Approved by: scottmcm
Pushing 4996b56 to master...

bors · 2023-06-15T13:02:41Z

☀️ Test successful - checks-actions
Approved by: scottmcm
Pushing 4996b56 to master...

rust-timer · 2023-06-15T14:22:36Z

Finished benchmarking commit (4996b56): 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)	0.7%	[0.4%, 1.1%]	5
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-0.2%	[-0.4%, -0.2%]	87
Improvements ✅ (secondary)	-0.3%	[-1.2%, -0.1%]	24
All ❌✅ (primary)	-0.2%	[-0.4%, 1.1%]	92

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.

	mean	range	count
Regressions ❌ (primary)	4.3%	[2.1%, 6.0%]	4
Regressions ❌ (secondary)	1.3%	[1.3%, 1.3%]	1
Improvements ✅ (primary)	-4.1%	[-7.1%, -1.8%]	3
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.7%	[-7.1%, 6.0%]	7

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.

	mean	range	count
Regressions ❌ (primary)	1.2%	[1.2%, 1.2%]	1
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	1.2%	[1.2%, 1.2%]	1

Binary size

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.

	mean	range	count
Regressions ❌ (primary)	0.2%	[0.0%, 0.6%]	45
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-0.4%	[-0.4%, -0.4%]	7
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.1%	[-0.4%, 0.6%]	52

Bootstrap: 646.601s -> 647.802s (0.19%)

…<try> Update `slice::Iter::rfold` to match `slice::Iter::fold` Adds a new codegen test for `rfold`, like the one from rust-lang#106343, and makes a similar fix, updating `rfold` to work via indices too.

rustbot assigned joshtriplett Jan 1, 2023

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. T-libs Relevant to the library team, which will review and decide on the PR/issue. labels Jan 1, 2023

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 Jan 1, 2023

bors 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 Jan 1, 2023

the8472 force-pushed the slice-iter-fold branch from 6a6a419 to e42a42b Compare January 1, 2023 04:35

This comment has been minimized.

Sign in to view

This comment was marked as outdated.

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 Jan 1, 2023

the8472 mentioned this pull request Jan 1, 2023

slice::Iter::fold optimizes poorly for some niche optimized types. #106288

Closed

the8472 closed this Jan 1, 2023

the8472 reopened this Jan 2, 2023

the8472 marked this pull request as draft January 2, 2023 04:19

the8472 force-pushed the slice-iter-fold branch from e42a42b to d5801f7 Compare January 2, 2023 04:21

This comment has been minimized.

Sign in to view

rustbot assigned scottmcm and unassigned jyn514 Jun 13, 2023

scottmcm reviewed Jun 14, 2023

View reviewed changes

rustbot 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 Jun 14, 2023

use indexed loop instead of ptr bumping

d90508f

this seems to produce less IR

the8472 force-pushed the slice-iter-fold branch from 500f081 to d90508f Compare June 14, 2023 20:23

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 Jun 14, 2023

scottmcm reviewed Jun 15, 2023

View reviewed changes

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 15, 2023

klensy reviewed Jun 15, 2023

View reviewed changes

bors added merged-by-bors This PR was explicitly merged by bors. labels Jun 15, 2023

bors merged commit 4996b56 into rust-lang:master Jun 15, 2023

rustbot added this to the 1.72.0 milestone Jun 15, 2023

bors mentioned this pull request Jun 15, 2023

Remove box_free lang item #100036

Merged

scottmcm mentioned this pull request Jun 22, 2023

iterator for_each performance regression #112911

Open

boulanlo mentioned this pull request Sep 6, 2023

Iterator inlining/optimization regression in 1.72 release #115601

Open

scottmcm mentioned this pull request Dec 22, 2023

Update slice::Iter::rfold to match slice::Iter::fold #119207

Closed

the8472 mentioned this pull request Jan 28, 2024

core::slice::Iter and core::slice::IterMut could be replaced with safe code #120438

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

optimize slice::Iter::fold #106343

optimize slice::Iter::fold #106343

the8472 commented Jan 1, 2023 •

edited

Loading

rustbot commented Jan 1, 2023

rustbot commented Jan 1, 2023

the8472 commented Jan 1, 2023

This comment has been minimized.

bors commented Jan 1, 2023

bors commented Jan 1, 2023

the8472 commented Jan 1, 2023

bors commented Jan 1, 2023

This comment has been minimized.

bors commented Jan 1, 2023

This comment has been minimized.

This comment was marked as outdated.

the8472 commented Jan 2, 2023

This comment has been minimized.

jyn514 commented Jun 13, 2023

scottmcm Jun 14, 2023

the8472 Jun 14, 2023

the8472 Jun 14, 2023

scottmcm Jun 15, 2023

scottmcm commented Jun 15, 2023

bors commented Jun 15, 2023

klensy Jun 15, 2023 •

edited

Loading

bors commented Jun 15, 2023

bors commented Jun 15, 2023

bors commented Jun 15, 2023

rust-timer commented Jun 15, 2023

optimize slice::Iter::fold #106343

optimize slice::Iter::fold #106343

Conversation

the8472 commented Jan 1, 2023 • edited Loading

rustbot commented Jan 1, 2023

rustbot commented Jan 1, 2023

the8472 commented Jan 1, 2023

This comment has been minimized.

bors commented Jan 1, 2023

bors commented Jan 1, 2023

the8472 commented Jan 1, 2023

bors commented Jan 1, 2023

This comment has been minimized.

bors commented Jan 1, 2023

This comment has been minimized.

This comment was marked as outdated.

the8472 commented Jan 2, 2023

This comment has been minimized.

jyn514 commented Jun 13, 2023

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

scottmcm commented Jun 15, 2023

bors commented Jun 15, 2023

klensy Jun 15, 2023 • edited Loading

Choose a reason for hiding this comment

bors commented Jun 15, 2023

bors commented Jun 15, 2023

bors commented Jun 15, 2023

rust-timer commented Jun 15, 2023

Overall result: ✅ improvements - no action needed

the8472 commented Jan 1, 2023 •

edited

Loading

klensy Jun 15, 2023 •

edited

Loading