rustdoc: Compute some fields of `clean::Crate` on-demand to reduce size by camelid · Pull Request #90391 · rust-lang/rust

camelid · 2021-10-29T04:04:24Z

clean::Crate is frequently moved by-value -- for example, in DocFolder
implementations -- so reducing its size should improve performance.

This PR reduces the size of clean::Crate from 168 bytes to 104 bytes.

r? @jyn514

camelid · 2021-10-29T04:06:59Z

@bors try @rust-timer queue

rust-timer · 2021-10-29T04:07:00Z

Awaiting bors try build completion.

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

bors · 2021-10-29T04:07:07Z

⌛ Trying commit a92c191ad66c182bee3f619ececce20fc216c51b with merge a95fcc6c76fa042939aae01abaf13368297003c2...

jyn514

This is great, thanks! Just a couple nits.

camelid · 2021-10-29T04:13:09Z

Note to self: remove this field as well.

Update: In order to remove this field, Cache::populate will need to take a DocContext, which I think is possible but will require some changes. I'm also working on that as part of #90365, so I'll save removing this field for a future PR.

bors · 2021-10-29T05:43:16Z

☀️ Try build successful - checks-actions
Build commit: a95fcc6c76fa042939aae01abaf13368297003c2 (a95fcc6c76fa042939aae01abaf13368297003c2)

rust-timer · 2021-10-29T05:43:17Z

Queued a95fcc6c76fa042939aae01abaf13368297003c2 with parent df76418, future comparison URL.

rust-timer · 2021-10-29T07:59:09Z

Finished benchmarking commit (a95fcc6c76fa042939aae01abaf13368297003c2): comparison url.

Summary: This benchmark run did not return any relevant changes.

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 led to changes in compiler perf.

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

jyn514 · 2021-10-29T14:04:15Z

Perf says this was actually a regression on max-rss?? That seems very strange, I'm not sure how to interpret it.

camelid · 2021-10-29T16:56:01Z

That is really weird. Let's try rerunning perf to see if that was spurious.

@bors try @rust-timer queue

rust-timer · 2021-10-29T16:56:03Z

Awaiting bors try build completion.

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

bors · 2021-10-29T16:56:14Z

⌛ Trying commit a92c191ad66c182bee3f619ececce20fc216c51b with merge 73c6d8ebffee3c722b0c6f21fb78c137711a3f3f...

bors · 2021-10-29T18:34:32Z

☀️ Try build successful - checks-actions
Build commit: 73c6d8ebffee3c722b0c6f21fb78c137711a3f3f (73c6d8ebffee3c722b0c6f21fb78c137711a3f3f)

rust-timer · 2021-10-29T18:34:33Z

Queued 73c6d8ebffee3c722b0c6f21fb78c137711a3f3f with parent 9ed5b94, future comparison URL.

rust-timer · 2021-10-29T20:41:54Z

Finished benchmarking commit (73c6d8ebffee3c722b0c6f21fb78c137711a3f3f): comparison url.

Summary: This benchmark run did not return any relevant changes.

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 led to changes in compiler perf.

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

jyn514 · 2021-10-29T21:14:04Z

... apparently that made the RSS regression go away? how strange ...

camelid · 2021-10-29T23:10:32Z

... apparently that made the RSS regression go away? how strange ...

In my experience, deep-vector often has spurious results, so it's not super surprising. For example, on the most recent perf run above, deep-vector opt had a max-rss regression of 4.85%, which is certainly spurious since this PR did not touch rustc.

jyn514 · 2021-10-30T01:25:44Z

@camelid I would rather not add back name, it's cheap to compute and rustdoc should be doing things on-demand.

camelid · 2021-10-30T01:41:37Z

@camelid I would rather not add back name, it's cheap to compute and rustdoc should be doing things on-demand.

Ok, sounds good. I'll try removing some of the other fields too.

camelid · 2021-10-30T03:45:51Z

(Currently squashing.)

camelid · 2021-10-30T04:02:22Z

Ok, should be good for a new review!

jyn514

r=me with those nits fixed :)

It is only used in one place; `src` was about a third of `Crate`'s total size; `Crate` is frequently moved by-value; and `src` can be easily computed on-demand.

It is not as large as `Crate.src` was, but it's still 8 bytes, and `clean::Crate` is moved by-value a lot.

Now that rust-lang#73423 is fixed, sorting should no longer be necessary. See also this discussion [1]. [1]: rust-lang#86889 (comment)

camelid · 2021-10-30T18:18:57Z

@bors r=jyn514 rollup=never

bors · 2021-10-30T18:18:59Z

📌 Commit a58e214 has been approved by jyn514

GuillaumeGomez · 2021-10-30T18:45:34Z


+// `Crate` is frequently moved by-value. Make sure it doesn't unintentionally get bigger.
+#[cfg(all(target_arch = "x86_64", target_pointer_width = "64"))]
+rustc_data_structures::static_assert_size!(Crate, 104);


Nice idea to add this check! 👍

bors · 2021-10-31T05:52:44Z

⌛ Testing commit a58e214 with merge 38b01d9...

bors · 2021-10-31T09:07:04Z

☀️ Test successful - checks-actions
Approved by: jyn514
Pushing 38b01d9 to master...

rust-timer · 2021-10-31T10:38:03Z

Finished benchmarking commit (38b01d9): comparison url.

Summary: This benchmark run did not return any relevant changes.

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

@rustbot label: -perf-regression

rust-highfive assigned jyn514 Oct 29, 2021

rust-highfive added the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label Oct 29, 2021

rustbot added the S-waiting-on-perf Status: Waiting on a perf run to be completed. label Oct 29, 2021

jyn514 reviewed Oct 29, 2021

View reviewed changes

Comment thread src/librustdoc/clean/utils.rs Outdated

Comment thread src/librustdoc/clean/types.rs Outdated

camelid added the S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. label Oct 29, 2021

camelid commented Oct 29, 2021

View reviewed changes

rustbot removed the S-waiting-on-perf Status: Waiting on a perf run to be completed. label Oct 29, 2021

rustbot added the S-waiting-on-perf Status: Waiting on a perf run to be completed. label Oct 29, 2021

rustbot removed the S-waiting-on-perf Status: Waiting on a perf run to be completed. label Oct 29, 2021

camelid mentioned this pull request Oct 29, 2021

[EXPERIMENT] rustdoc: Mark Context::tcx() as #[inline(always)] #90411

Closed

rustdoc: Add static size assertion for clean::Crate

27d47d9

camelid force-pushed the crate-size branch from bac42a7 to 62129e1 Compare October 30, 2021 03:54

camelid removed the S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. label Oct 30, 2021

jyn514 suggested changes Oct 30, 2021

View reviewed changes

Comment thread src/librustdoc/clean/types.rs Outdated

Comment thread src/librustdoc/clean/types.rs Outdated

jyn514 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 Oct 30, 2021

camelid added 4 commits October 30, 2021 11:15

rustdoc: Document that Crate is always local

47b0059

rustdoc: Remove Crate.src and instead compute it on-demand

85f8ae8

It is only used in one place; `src` was about a third of `Crate`'s total size; `Crate` is frequently moved by-value; and `src` can be easily computed on-demand.

rustdoc: Remove Crate.name and instead compute it on-demand

ebe9a11

It is not as large as `Crate.src` was, but it's still 8 bytes, and `clean::Crate` is moved by-value a lot.

rustdoc: Stop sorting external crates

a58e214

Now that rust-lang#73423 is fixed, sorting should no longer be necessary. See also this discussion [1]. [1]: rust-lang#86889 (comment)

camelid force-pushed the crate-size branch from 6619ca4 to a58e214 Compare October 30, 2021 18:18

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 Oct 30, 2021

GuillaumeGomez reviewed Oct 30, 2021

View reviewed changes

bors added the merged-by-bors This PR was explicitly merged by bors. label Oct 31, 2021

bors merged commit 38b01d9 into rust-lang:master Oct 31, 2021

rustbot added this to the 1.58.0 milestone Oct 31, 2021

bors mentioned this pull request Oct 31, 2021

[EXPERIMENT] rustdoc: Cleanup external_traits #90365

Closed

camelid deleted the crate-size branch October 31, 2021 17:00

Uh oh!

Uh oh!

Conversation

camelid commented Oct 29, 2021 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

camelid commented Oct 29, 2021

Uh oh!

rust-timer commented Oct 29, 2021

Uh oh!

bors commented Oct 29, 2021

Uh oh!

jyn514 left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

camelid Oct 29, 2021

Choose a reason for hiding this comment

Uh oh!

camelid Oct 30, 2021

Choose a reason for hiding this comment

Uh oh!

bors commented Oct 29, 2021

Uh oh!

rust-timer commented Oct 29, 2021

Uh oh!

rust-timer commented Oct 29, 2021

Uh oh!

jyn514 commented Oct 29, 2021

Uh oh!

camelid commented Oct 29, 2021

Uh oh!

rust-timer commented Oct 29, 2021

Uh oh!

bors commented Oct 29, 2021

Uh oh!

bors commented Oct 29, 2021

Uh oh!

rust-timer commented Oct 29, 2021

Uh oh!

rust-timer commented Oct 29, 2021

Uh oh!

jyn514 commented Oct 29, 2021

Uh oh!

camelid commented Oct 29, 2021

Uh oh!

jyn514 commented Oct 30, 2021 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

camelid commented Oct 30, 2021

Uh oh!

camelid commented Oct 30, 2021

Uh oh!

camelid commented Oct 30, 2021

Uh oh!

jyn514 left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

camelid commented Oct 30, 2021

Uh oh!

bors commented Oct 30, 2021

Uh oh!

GuillaumeGomez Oct 30, 2021

Choose a reason for hiding this comment

Uh oh!

camelid Oct 30, 2021

Choose a reason for hiding this comment

Uh oh!

bors commented Oct 31, 2021

Uh oh!

bors commented Oct 31, 2021

Uh oh!

rust-timer commented Oct 31, 2021

Uh oh!

Reviewers

Assignees

Labels

camelid commented Oct 29, 2021 •

edited

Loading

jyn514 commented Oct 30, 2021 •

edited

Loading