Specialize some methods of `io::Chain` #105917

a1phyr · 2022-12-19T19:38:44Z

This PR specializes the implementation of some methods of io::Chain, which could bring performance improvements when using it.

rustbot · 2022-12-19T19:38:50Z

r? @m-ou-se

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

rustbot · 2022-12-19T19:38:53Z

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

library/std/src/io/mod.rs

pitaj · 2023-01-27T07:12:28Z

If you resolved all issues brought up in the review, you should use the bot to mark it as waiting-on-review

a1phyr · 2023-01-27T09:21:20Z

@rustbot label -S-waiting-on-author +S-waiting-on-review

a1phyr · 2023-01-27T09:22:07Z

I didn't know I can do that myself ! Thanks !

a1phyr · 2023-06-12T16:19:42Z

I have removed read_to_string specialization: an error would happen if an UTF8 sequence were split between the two readers.
I have also added a test for that case.

tgross35 · 2023-07-14T21:52:39Z

Are you able to run a quick benchmark that shows this being faster? I'm positive it will be faster, but it never hurts to have something to back up added complexity for the sake of performance.

We could do a perf run on this in any case, but I don't expect this change to be exercised much in that run.

(side note, @rustbot ready does the same thing as manually swapping those two labels out)

workingjubilee · 2023-07-26T03:55:20Z

Looks good. Can you do the squishy-squishy for the last two commits?

r? workingjubilee
@rustbot author

workingjubilee · 2023-07-26T03:57:20Z

We could do a perf run on this in any case, but I don't expect this change to be exercised much in that run.

I don't think there's a single instance of this type in the compiler?

workingjubilee · 2023-07-26T18:18:15Z

Mostly it's confusing to see a commit that introduces something and then immediately see its removal while navigating the repo's history, so it's just noise in the blame.

…izations

workingjubilee · 2023-07-28T02:40:27Z

Making sure I get to this soon.
@rustbot ready

workingjubilee · 2023-07-30T22:39:30Z

library/std/src/io/mod.rs

+    #[inline]
+    fn is_read_vectored(&self) -> bool {
+        self.first.is_read_vectored() || self.second.is_read_vectored()
+    }


Shouldn't this require both to be efficient? I could be persuaded otherwise, certainly, but can you explain why you chose this logic instead? If one half of the chain doesn't have an efficient implementation, then why is the chained type efficient? At the extreme end, we're allowed to apply Chain recursively, so this version you propose would yield true even if we'd chained 16 different things together and only 1 reported true...

Suggested change

#[inline]

fn is_read_vectored(&self) -> bool {

self.first.is_read_vectored() || self.second.is_read_vectored()

}

#[inline]

fn is_read_vectored(&self) -> bool {

self.first.is_read_vectored() & self.second.is_read_vectored()

}

I guess you could also use a more complicated answer which checks done_first, but that seems prone to confusion, even if it might be "more accurate".

library/std/src/io/mod.rs

workingjubilee · 2023-07-31T13:59:29Z

Waiting on discussion a bit, so
@rustbot author

workingjubilee · 2023-10-05T08:25:51Z

Also, I can accept this as-is without the is_read_vectored change and that could be handled separately.

JohnCSimon · 2023-11-12T18:31:12Z

@a1phyr

ping from triage - can you post your status on this PR? There hasn't been an update in month - you seem close to getting this into a mergable state.

FYI: when a PR is ready for review, send a message containing
@rustbot ready to switch to S-waiting-on-review so the PR is in the reviewer's backlog.

a1phyr · 2023-12-04T18:44:50Z

Well there is debate over the implementation of is_read_vectored.

I initially used self.first.is_read_vectored() || self.second.is_read_vectored().
@workingjubilee suggested self.first.is_read_vectored() && self.second.is_read_vectored().
And there are the more "correct" implementations based on self.done_first, but I don't expect people calling this function more than once so this may just be confusing.

workingjubilee · 2023-12-05T02:12:51Z

Indeed. I'm not entirely sure what the rationale is for the || version? But I tried to explain why I think the && version is more appropriate. I agree that using the "technically correct" implementation seems pointless/confusing.

workingjubilee · 2023-12-05T02:14:26Z

I'm happy to be persuaded? It just doesn't feel obvious atm.

a1phyr · 2024-01-16T08:55:03Z

Well, the rationale is mostly the same: if we chain 16 types and only one returns false, we would lose a lot of perf.

The main difference is that vectored I/O can significantly improve perf, but has a small cost when used with a single buffer (a indirection on the stack). This is why WASI doesn't have an equivalent of read, only readv (https://docs.rs/wasi/latest/wasi/fn.fd_read.html).

So even with half the chain not vectored, having vectored I/O for the other is usually interesting.

workingjubilee · 2024-02-18T21:59:13Z

Hmm. I see. WASI only providing readv does indeed feel persuasive.

@bors r+

bors · 2024-02-18T21:59:16Z

📌 Commit ebc5970 has been approved by workingjubilee

It is now in the queue for this repository.

bors · 2024-02-19T04:44:00Z

⌛ Testing commit ebc5970 with merge bea5beb...

bors · 2024-02-19T06:42:44Z

☀️ Test successful - checks-actions
Approved by: workingjubilee
Pushing bea5beb to master...

bors · 2024-02-19T06:42:44Z

☀️ Test successful - checks-actions
Approved by: workingjubilee
Pushing bea5beb to master...

rust-timer · 2024-02-19T08:40:29Z

Finished benchmarking commit (bea5beb): comparison URL.

Overall result: no relevant changes - no action needed

@rustbot label: -perf-regression

Instruction count

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

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)	-	-	0
Regressions ❌ (secondary)	3.5%	[2.5%, 4.4%]	2
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-2.0%	[-2.0%, -2.0%]	1
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.

	mean	range	count
Regressions ❌ (primary)	3.2%	[2.8%, 3.8%]	3
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-2.8%	[-3.5%, -2.1%]	3
All ❌✅ (primary)	3.2%	[2.8%, 3.8%]	3

Binary size

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

Bootstrap: 640.888s -> 640.916s (0.00%)
Artifact size: 308.82 MiB -> 308.79 MiB (-0.01%)

rustbot assigned m-ou-se Dec 19, 2022

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 Dec 19, 2022

a1phyr changed the title ~~Read chain more impls~~ Specialize some methods of io::Chain Dec 19, 2022

This comment has been minimized.

Sign in to view

a1phyr force-pushed the read_chain_more_impls branch from bb16dd8 to 57718e5 Compare December 19, 2022 22:56

m-ou-se mentioned this pull request Dec 28, 2022

Tracking Issue for Read::is_read_vectored/Write::is_write_vectored. #69941

Open

1 task

m-ou-se reviewed Dec 28, 2022

View reviewed changes

library/std/src/io/mod.rs Show resolved Hide resolved

m-ou-se reviewed Dec 28, 2022

View reviewed changes

library/std/src/io/mod.rs Outdated Show resolved Hide resolved

m-ou-se reviewed Dec 28, 2022

View reviewed changes

library/std/src/io/mod.rs Outdated Show resolved Hide resolved

m-ou-se 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 Dec 28, 2022

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 Jan 27, 2023

a1phyr requested a review from m-ou-se January 31, 2023 23:13

anden3 added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels May 14, 2023

rustbot assigned workingjubilee and unassigned m-ou-se Jul 26, 2023

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 Jul 26, 2023

a1phyr added 3 commits July 26, 2023 23:28

io::Chain: specialize some Read methods

cada71e

io::Chain: specialize some BufRead methods

cba6e10

Add tests and comments about read_to_string and read_line special…

ebc5970

…izations

a1phyr force-pushed the read_chain_more_impls branch from dfbcdcd to ebc5970 Compare July 26, 2023 21:31

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 Jul 28, 2023

workingjubilee reviewed Jul 30, 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 Jul 31, 2023

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-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Feb 18, 2024

bors added the merged-by-bors This PR was explicitly merged by bors. label Feb 19, 2024

bors merged commit bea5beb into rust-lang:master Feb 19, 2024
12 checks passed

rustbot added this to the 1.78.0 milestone Feb 19, 2024

a1phyr deleted the read_chain_more_impls branch March 12, 2024 17:12

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Specialize some methods of `io::Chain` #105917

Specialize some methods of `io::Chain` #105917

a1phyr commented Dec 19, 2022

rustbot commented Dec 19, 2022

rustbot commented Dec 19, 2022

This comment has been minimized.

pitaj commented Jan 27, 2023

a1phyr commented Jan 27, 2023

a1phyr commented Jan 27, 2023

a1phyr commented Jun 12, 2023

tgross35 commented Jul 14, 2023

workingjubilee commented Jul 26, 2023

workingjubilee commented Jul 26, 2023

workingjubilee commented Jul 26, 2023

workingjubilee commented Jul 28, 2023

workingjubilee Jul 30, 2023 •

edited

Loading

workingjubilee Jul 30, 2023

workingjubilee commented Jul 31, 2023

workingjubilee commented Oct 5, 2023

JohnCSimon commented Nov 12, 2023

a1phyr commented Dec 4, 2023

workingjubilee commented Dec 5, 2023

workingjubilee commented Dec 5, 2023

a1phyr commented Jan 16, 2024

workingjubilee commented Feb 18, 2024

bors commented Feb 18, 2024

bors commented Feb 19, 2024

bors commented Feb 19, 2024

bors commented Feb 19, 2024

rust-timer commented Feb 19, 2024

Specialize some methods of io::Chain #105917

Specialize some methods of io::Chain #105917

Conversation

a1phyr commented Dec 19, 2022

rustbot commented Dec 19, 2022

rustbot commented Dec 19, 2022

This comment has been minimized.

pitaj commented Jan 27, 2023

a1phyr commented Jan 27, 2023

a1phyr commented Jan 27, 2023

a1phyr commented Jun 12, 2023

tgross35 commented Jul 14, 2023

workingjubilee commented Jul 26, 2023

workingjubilee commented Jul 26, 2023

workingjubilee commented Jul 26, 2023

workingjubilee commented Jul 28, 2023

workingjubilee Jul 30, 2023 • edited Loading

Choose a reason for hiding this comment

workingjubilee Jul 30, 2023

Choose a reason for hiding this comment

workingjubilee commented Jul 31, 2023

workingjubilee commented Oct 5, 2023

JohnCSimon commented Nov 12, 2023

a1phyr commented Dec 4, 2023

workingjubilee commented Dec 5, 2023

workingjubilee commented Dec 5, 2023

a1phyr commented Jan 16, 2024

workingjubilee commented Feb 18, 2024

bors commented Feb 18, 2024

bors commented Feb 19, 2024

bors commented Feb 19, 2024

bors commented Feb 19, 2024

rust-timer commented Feb 19, 2024

Overall result: no relevant changes - no action needed

Instruction count

Max RSS (memory usage)

Cycles

Binary size

Specialize some methods of `io::Chain` #105917

Specialize some methods of `io::Chain` #105917

workingjubilee Jul 30, 2023 •

edited

Loading