Add `-C hint-mostly-unused` to tell rustc that most of a crate will go unused #135656

joshtriplett · 2025-01-17T21:29:50Z

This hint allows the compiler to optimize its operation based on this assumption, in order to compile faster. This is a hint, and does not guarantee any particular behavior.

This option can substantially speed up compilation if applied to a large dependency where the majority of the dependency does not get used. This flag may slow down compilation in other cases.

Currently, this option makes the compiler defer as much code generation as possible from functions in the crate, until later crates invoke those functions. Functions that never get invoked will never have code generated for them. For instance, if a crate provides thousands of functions, but only a few of them will get called, this flag will result in the compiler only doing code generation for the called functions. (This uses the same mechanisms as cross-crate inlining of functions.) This does not affect extern functions, or functions marked as #[inline(never)].

This option has already existed in nightly as -Zcross-crate-inline-threshold=always for some time, and has gotten testing in that form. However, this option is still unstable, to give an opportunity for wider testing in this form.

Some performance numbers, based on a crate with many dependencies having just one large dependency set to -C hint-mostly-unused (using Cargo's profile-rustflags option):

A release build went from 4m32s to 2m06s.

A non-release build went from 2m13s to 1m24s.

rustbot · 2025-01-17T21:29:59Z

r? @GuillaumeGomez

rustbot has assigned @GuillaumeGomez.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

rustbot · 2025-01-17T21:30:00Z

Some changes occurred to MIR optimizations

cc @rust-lang/wg-mir-opt

GuillaumeGomez · 2025-01-17T21:33:02Z

r? compiler

workingjubilee · 2025-01-17T21:44:01Z

This should be added as an unstable option first?

joshtriplett · 2025-01-17T21:49:23Z

@workingjubilee This option has already existed in nightly as -Zcross-crate-inline-threshold=always for some time, and has gotten testing in that form.

The only difference being made between that option and this one is that this one respects #[inline(never)].

compiler/rustc_interface/src/tests.rs

workingjubilee · 2025-01-17T22:53:27Z

@joshtriplett:

@workingjubilee This option has already existed in nightly as -Zcross-crate-inline-threshold=always for some time, and has gotten testing in that form.

The only difference being made between that option and this one is that this one respects #[inline(never)].

I see. I suppose that makes sense, though it brings these thoughts to mind:

In general, T-compiler has a much lower threshold for unstable flags that are not intended for serious or permanent usage downstream. They're often added as debugging aids. I am not sure if this was intended for public usage when it was originally added.
All inlining annotations are hints, and this is also noted as a hint, yet it changes the behavior of another inlining hint, and in that way deviates from its unstable variation.
Notably, inline(never) is often relied on to some degree by programmers as working.
This describes itself as a hint, but also is motivated on the basis that it will be very impactful, which suggests its behavior will be immediately relied on. The combination of relying-on-inlining-but-also-relying-on-not-inlining makes me wonder if it is definitely correct? There are a lot of weird nuances to our inlining rules.

As a result, I am curious if the rest of T-compiler believes changing the behavior with respect to inline(never) is trivial.

Perhaps this should land with a test that demonstrates that it respects #[inline(never)] correctly, now?

workingjubilee · 2025-01-18T00:13:07Z

( I'm also kinda curious if we should adjust how we handle cross-crate inlining by default, tbh. )

lqd · 2025-01-18T00:24:30Z

You may have missed it Jubilee, so hop on to https://rust-lang.zulipchat.com/#narrow/stream/233931-xxx/topic/Add.20.60-C.20hint-mostly-unused.60.20option.20compiler-team.23829 to ask your cool questions and comments :)

…o unused This hint allows the compiler to optimize its operation based on this assumption, in order to compile faster. This is a hint, and does not guarantee any particular behavior. This option can substantially speed up compilation if applied to a large dependency where the majority of the dependency does not get used. This flag may slow down compilation in other cases. Currently, this option makes the compiler defer as much code generation as possible from functions in the crate, until later crates invoke those functions. Functions that never get invoked will never have code generated for them. For instance, if a crate provides thousands of functions, but only a few of them will get called, this flag will result in the compiler only doing code generation for the called functions. (This uses the same mechanisms as cross-crate inlining of functions.) This does not affect `extern` functions, or functions marked as `#[inline(never)]`. Some performance numbers, based on a crate with many dependencies having just *one* large dependency set to `-C hint-mostly-unused` (using Cargo's `profile-rustflags` option): A release build went from 4m32s to 2m06s. A non-release build went from 2m13s to 1m24s.

Urgau · 2025-01-18T10:41:47Z

src/doc/rustc/src/codegen-options/index.md

@@ -172,6 +172,25 @@ values:

 The default if not specified depends on the target.

+## hint-mostly-unused


Given that the flag is now unstable/nightly-gated, this documentation should be moved to the unstable-book (under src/doc/unstable-book/src/compiler-flags) and a tracking issue should be created.

No need for a tracking issue before consensus tho right ? This still needs fcp anyways. (And given the reservations in the thread some more tests and all will be needed before stabilization.)

Sure, the creation of the tracking issue can wait until the MCP is accepted and the PR nearing approval.

fee1-dead · 2025-01-21T23:15:02Z

r? compiler

apiraino · 2025-01-23T17:28:08Z

Discussed in T-compiler triage on Zulip.

MCP#829 still needs to be seconded and there seems to be still a few open design questions. Also, do we have some numbers about the real world impact (both positive and negative)?. Maybe it makes sense to nominate again once design and scope are clearer.

By the way @saethlin do you think you have some capacity to dedicate here?

@rustbot label -I-compiler-nominated

saethlin · 2025-01-24T01:28:59Z

compiler/rustc_mir_transform/src/cross_crate_inline.rs

+    // If the crate is likely to be mostly unused, use cross-crate inlining to defer codegen until
+    // the function is referenced, in order to skip codegen for unused functions.
+    if tcx.sess.opts.cg.hint_mostly_unused {
+        return true;
+    }
+


Placing this logic below the check for InlineAttr::Never does not change whether a call to a #[inline(never)] function can be inlined, it changes when and where #[inline(never)] functions get compiled. With the logic down here, #[inline(never)] functions and their callees (recursively) are not delayed.

The exact motivation for putting this logic here instead of above the check for codegen_fn_attrs.inline should be clearly recorded. I don't rightly remember what it was, but for example panic backtrace trimming still works even if it is moved.

Making sure inline-never functions don't get affected was an intentional goal here. I'll add a comment.

saethlin · 2025-01-28T01:22:51Z

src/doc/rustc/src/codegen-options/index.md

+This option can substantially speed up compilation if applied to a large
+dependency where the majority of the dependency does not get used. This flag


Can we provide any indication of how to do this? I've suggested some people try adding -Zcross-crate-inline-threshold=always to just one dependency and nobody seems to know how to do this offhand. If we're going to suggest people do a thing that's poorly understood I'd like to at least have a link to the Cargo docs about profile overrides.

Sure, will do.

joshtriplett added I-compiletime Issue: Problems and improvements with respect to compile times. T-compiler Relevant to the compiler team, which will review and decide on the PR/issue. labels Jan 17, 2025

rustbot assigned GuillaumeGomez Jan 17, 2025

rustbot added the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label Jan 17, 2025

joshtriplett mentioned this pull request Jan 17, 2025

Add -C hint-mostly-unused option rust-lang/compiler-team#829

Open

3 tasks

rustbot assigned fee1-dead and unassigned GuillaumeGomez Jan 17, 2025

joshtriplett added the relnotes Marks issues that should be documented in the release notes of the next release. label Jan 17, 2025

rustbot mentioned this pull request Jan 17, 2025

Tracking issue for release notes of #135656: Add -C hint-mostly-unused to tell rustc that most of a crate will go unused #135657

Open

3 tasks

workingjubilee added needs-fcp This change is insta-stable, so needs a completed FCP to proceed. I-compiler-nominated Nominated for discussion during a compiler team meeting. labels Jan 17, 2025

This comment was marked as resolved.

Sign in to view

Urgau reviewed Jan 17, 2025

View reviewed changes

compiler/rustc_interface/src/tests.rs Outdated Show resolved Hide resolved

lqd reviewed Jan 17, 2025

View reviewed changes

compiler/rustc_interface/src/tests.rs Outdated Show resolved Hide resolved

joshtriplett force-pushed the hint-mostly-unused branch from 4d1b2cb to e96e15f Compare January 18, 2025 05:19

Make -C hint-mostly-unused require -Zunstable-options for now

d3f3a68

joshtriplett removed relnotes Marks issues that should be documented in the release notes of the next release. needs-fcp This change is insta-stable, so needs a completed FCP to proceed. labels Jan 18, 2025

Urgau reviewed Jan 18, 2025

View reviewed changes

rustbot assigned BoxyUwU and unassigned fee1-dead Jan 21, 2025

rustbot removed the I-compiler-nominated Nominated for discussion during a compiler team meeting. label Jan 23, 2025

saethlin reviewed Jan 24, 2025

View reviewed changes

saethlin assigned saethlin and unassigned BoxyUwU Jan 24, 2025

saethlin reviewed Jan 28, 2025

View reviewed changes

saethlin 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 Feb 7, 2025

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add `-C hint-mostly-unused` to tell rustc that most of a crate will go unused #135656

Add `-C hint-mostly-unused` to tell rustc that most of a crate will go unused #135656

joshtriplett commented Jan 17, 2025 •

edited

Loading

rustbot commented Jan 17, 2025

rustbot commented Jan 17, 2025

GuillaumeGomez commented Jan 17, 2025

workingjubilee commented Jan 17, 2025

joshtriplett commented Jan 17, 2025

This comment was marked as resolved.

This comment was marked as resolved.

workingjubilee commented Jan 17, 2025

workingjubilee commented Jan 18, 2025

lqd commented Jan 18, 2025

Urgau Jan 18, 2025

lqd Jan 18, 2025

Urgau Jan 18, 2025

fee1-dead commented Jan 21, 2025

apiraino commented Jan 23, 2025

saethlin Jan 24, 2025

joshtriplett Jan 28, 2025

saethlin Jan 28, 2025

joshtriplett Jan 28, 2025

		@@ -172,6 +172,25 @@ values:

		The default if not specified depends on the target.

		## hint-mostly-unused

		This option can substantially speed up compilation if applied to a large
		dependency where the majority of the dependency does not get used. This flag

Add -C hint-mostly-unused to tell rustc that most of a crate will go unused #135656

Are you sure you want to change the base?

Add -C hint-mostly-unused to tell rustc that most of a crate will go unused #135656

Conversation

joshtriplett commented Jan 17, 2025 • edited Loading

rustbot commented Jan 17, 2025

rustbot commented Jan 17, 2025

GuillaumeGomez commented Jan 17, 2025

workingjubilee commented Jan 17, 2025

joshtriplett commented Jan 17, 2025

This comment was marked as resolved.

This comment was marked as resolved.

workingjubilee commented Jan 17, 2025

workingjubilee commented Jan 18, 2025

lqd commented Jan 18, 2025

Urgau Jan 18, 2025

Choose a reason for hiding this comment

lqd Jan 18, 2025

Choose a reason for hiding this comment

Urgau Jan 18, 2025

Choose a reason for hiding this comment

fee1-dead commented Jan 21, 2025

apiraino commented Jan 23, 2025

saethlin Jan 24, 2025

Choose a reason for hiding this comment

joshtriplett Jan 28, 2025

Choose a reason for hiding this comment

saethlin Jan 28, 2025

Choose a reason for hiding this comment

joshtriplett Jan 28, 2025

Choose a reason for hiding this comment

Add `-C hint-mostly-unused` to tell rustc that most of a crate will go unused #135656

Add `-C hint-mostly-unused` to tell rustc that most of a crate will go unused #135656

joshtriplett commented Jan 17, 2025 •

edited

Loading