Lower the assume intrinsic to a MIR statement

[oli-obk]

This makes #96862 (comment) easier and will generally allow us to cheaply insert assume intrinsic calls in mir building.

r? rust-lang/wg-mir-opt

[rust-highfive]

Some changes occured to the CTFE / Miri engine

cc @rust-lang/miri

Some changes occured to rustc_codegen_cranelift

cc @bjorn3

Some changes occurred in src/tools/clippy.

cc @rust-lang/clippy

[rust-highfive]

r? @wesleywiser

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

[oli-obk]

cc @rust-lang/wg-mir-opt

[oli-obk]

@bors try @rust-timer queue

[rust-timer]

Awaiting bors try build completion.

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

[bors]

⌛ Trying commit 15f4f800da144f34374cdaefdf31abeab5465942 with merge ec462c48b5f3225dba4718607e400ee78bdc577d...

[bors]

☀️ Try build successful - checks-actions
Build commit: ec462c48b5f3225dba4718607e400ee78bdc577d (ec462c48b5f3225dba4718607e400ee78bdc577d)

[rust-timer]

Queued ec462c48b5f3225dba4718607e400ee78bdc577d with parent 42dcf70, future comparison URL.

[rust-timer]

Finished benchmarking commit (ec462c48b5f3225dba4718607e400ee78bdc577d): 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)	0.4%	0.4%	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: 🎉 relevant improvement found
• Secondary benchmarks: 😿 relevant regressions found

	mean1	max	count2
Regressions 😿 (primary)	N/A	N/A	0
Regressions 😿 (secondary)	2.1%	3.1%	2
Improvements 🎉 (primary)	-3.5%	-3.5%	1
Improvements 🎉 (secondary)	N/A	N/A	0
All 😿🎉 (primary)	-3.5%	-3.5%	1

Cycles

Results

• Primary benchmarks: 🎉 relevant improvement found
• Secondary benchmarks: 🎉 relevant improvement found

	mean1	max	count2
Regressions 😿 (primary)	N/A	N/A	0
Regressions 😿 (secondary)	N/A	N/A	0
Improvements 🎉 (primary)	-2.1%	-2.1%	1
Improvements 🎉 (secondary)	-2.3%	-2.3%	1
All 😿🎉 (primary)	-2.1%	-2.1%	1

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

Footnotes

1. the arithmetic mean of the percent change↩↩²↩³

2. number of relevant changes↩↩²↩³

[wesleywiser]

I will give this a more thorough review later this week.

[wesleywiser]

cc @JakobDegen in case you'd like to take a look as well 🙂

[JakobDegen]

Thanks for thinking of me Wesley - I actually am on wg-mir-opt though so I had received that ping already 😆

---

Obviously not a concern that should block this (or indeed any one) PR, but just some thoughts: Adding a new variant to StatementKind is always quite a bit of churn, and as MIR matures and as we add more optimizations that do more complicated things, I expect that it will get more and more error prone to do so.

One way to do this is to do less exhaustive matching in optimizations and try and generalize things to helper functions, but not sure if that's really feasible. A more plausible thing that I could imagine though is us adding a

StatementKind::Intrinsic {
inputs: Vec<Place>,
output: Option<Place>,
intrinsic: IntrinsicKind,
}
#[non_exhaustive]
enum IntrinsicKind { Assume, ... }

The idea here is that this is a fairly general "shape" for a statement to take, and it is likely that large classes of statements that we want to add in the future can be brought into this form. If we then add some requirements about the behavior of these statements (eg may only write to the output place), we can have optimizations provide a fall-through case on IntrinsicKind, and so adding a new one is easy. Optimizations that want to give special treatment to a particular intrinsic are of course still free to do so.

I'm now wondering if maybe this kind of definition makes more sense for Rvalue too... I feel like a lot of use cases only care about the list of input and output places/operands, and not about which operation is used to do the computation.

[bors]

☔ The latest upstream changes (presumably #98649) made this pull request unmergeable. Please resolve the merge conflicts.

[oli-obk]

The idea here is that this is a fairly general "shape" for a statement to take, and it is likely that large classes of statements that we want to add in the future can be brought into this form. If we then add some requirements about the behavior of these statements (eg may only write to the output place), we can have optimizations provide a fall-through case on IntrinsicKind, and so adding a new one is easy. Optimizations that want to give special treatment to a particular intrinsic are of course still free to do so.

I'm now wondering if maybe this kind of definition makes more sense for Rvalue too... I feel like a lot of use cases only care about the list of input and output places/operands, and not about which operation is used to do the computation.

Yea, we talked about this when we created CopyNonOverlapping. Though I guess the idea there was more open (basically a Call statement that doesn't unwind like the terminator). Restricting it just to a set of specific functions exhaustively enumerated via an enum, that seems like something we should explore and see if it helps.

[rustbot]

Some changes occurred in compiler/rustc_codegen_cranelift

cc @bjorn3

Some changes occurred to MIR optimizations

cc @rust-lang/wg-mir-opt

Some changes occurred to the CTFE / Miri engine

cc @rust-lang/miri

Some changes occurred in src/tools/clippy

cc @rust-lang/clippy

[wesleywiser]

One small comment. r=me when you're ready

[oli-obk]

@bors r=wesleywiser

[bors]

📌 Commit a0130e6 has been approved by wesleywiser

It is now in the queue for this repository.

[bors]

⌛ Testing commit a0130e6 with merge 2a0411df513195dc79ed466ad848f89ee8269373...

[bors]

💔 Test failed - checks-actions

[rust-log-analyzer]

The job x86_64-msvc-cargo failed! Check out the build log: (web)(plain)

Click to see the possible cause of the failure (guessed by this bot)

Initialized empty Git repository in D:/a/rust/rust/build/ct/iron/.git/
fatal: Could not parse object 'cf056ea5e8052c1feea6141e40ab0306715a2c33'.
remote: Internal Server Error.
remote: fatal: unable to access 'https://github.com/iron/iron/': The requested URL returned error: 500
thread 'main' panicked at 'assertion failed: status.success()', src\tools\cargotest\main.rs:148:13
Build completed unsuccessfully in 0:47:28

[oli-obk]

@bors retry

[bors]

⌛ Testing commit a0130e6 with merge e7c7aa7...

[bors]

☀️ Test successful - checks-actions
Approved by: wesleywiser
Pushing e7c7aa7 to master...

[rust-timer]

Finished benchmarking commit (e7c7aa7): 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.

	mean1	range	count2
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-0.8%	[-0.9%, -0.7%]	6
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-0.8%	[-0.9%, -0.7%]	6

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.

	mean1	range	count2
Regressions ❌ (primary)	3.7%	[3.3%, 4.1%]	2
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	3.7%	[3.3%, 4.1%]	2

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.

	mean1	range	count2
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-3.4%	[-3.4%, -3.4%]	1
All ❌✅ (primary)	-	-	0

Footnotes

1. the arithmetic mean of the percent change↩↩²↩³

2. number of relevant changes↩↩²↩³