make the eval() functions on our const types return the resulting value #115803

RalfJung · 2023-09-12T22:05:04Z

This is a part of #115748 that's hopefully perf-neutral, and that does not depend on #115764.

rustbot · 2023-09-12T22:05:12Z

r? @b-naber

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

rustbot · 2023-09-12T22:05:15Z

Some changes occurred to the CTFE / Miri engine

cc @rust-lang/miri

Some changes occurred in src/tools/clippy

cc @rust-lang/clippy

Some changes might have occurred in exhaustiveness checking

cc @Nadrieril

Some changes occurred in compiler/rustc_codegen_cranelift

cc @bjorn3

Some changes occurred to MIR optimizations

cc @rust-lang/wg-mir-opt

RalfJung · 2023-09-12T22:06:23Z

compiler/rustc_middle/src/mir/mod.rs

+            ConstantKind::Ty(c) => {
+                if let ty::ConstKind::Unevaluated(uv) = c.kind() {
+                    // Avoid the round-trip via valtree, evaluate directly to ConstValue.
+                    uv.expand()


I wanted to start with as small a change as possible, so this avoids the intermediate valtree for now.

RalfJung · 2023-09-12T22:07:06Z

r? @oli-obk
@bors try
@rust-timer queue

bors · 2023-09-12T22:07:15Z

⌛ Trying commit 35fa022 with merge 9aa0bc9...

make the eval() functions on our const types return the resulting value This is a part of rust-lang#115748 that's hopefully perf-neutral, and that does not depend on rust-lang#115764.

compiler/rustc_codegen_ssa/src/debuginfo/type_names.rs

RalfJung · 2023-09-12T22:37:00Z

@bors try

bors · 2023-09-12T22:37:08Z

⌛ Trying commit 6086a68 with merge d9931dc...

make the eval() functions on our const types return the resulting value This is a part of rust-lang#115748 that's hopefully perf-neutral, and that does not depend on rust-lang#115764.

bors · 2023-09-12T23:47:46Z

☀️ Try build successful - checks-actions
Build commit: d9931dc (d9931dc754958a938ba2dd00bb8166923191dcec)

rust-timer · 2023-09-13T03:00:18Z

Finished benchmarking commit (d9931dc): comparison URL.

Overall result: ❌ regressions - ACTION NEEDED

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.

Next Steps: If you can justify the regressions found in this try perf run, please indicate this with @rustbot label: +perf-regression-triaged along with sufficient written justification. If you cannot justify the regressions please fix the regressions and do another perf run. If the next run shows neutral or positive results, the label will be automatically removed.

@bors rollup=never
@rustbot label: -S-waiting-on-perf +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)	0.8%	[0.3%, 1.4%]	12
Regressions ❌ (secondary)	0.8%	[0.4%, 1.3%]	11
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.8%	[0.3%, 1.4%]	12

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)	2.3%	[0.7%, 3.8%]	2
Improvements ✅ (primary)	-0.6%	[-0.6%, -0.6%]	1
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-0.6%	[-0.6%, -0.6%]	1

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%	[1.5%, 5.2%]	5
Regressions ❌ (secondary)	2.2%	[2.2%, 2.2%]	2
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	3.2%	[1.5%, 5.2%]	5

Binary size

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

Bootstrap: 631.022s -> 629.231s (-0.28%)
Artifact size: 317.92 MiB -> 317.34 MiB (-0.18%)

RalfJung · 2023-09-13T05:27:06Z

compiler/rustc_middle/src/ty/consts.rs

+                    tcx.erase_regions(param_env)
+                        .with_reveal_all_normalized(tcx)
+                        .and(tcx.erase_regions(unevaluated))
+                };


This is some magic that only the Valtree path has, maybe that's where the perf diff comes from?

bors · 2023-09-13T05:47:42Z

⌛ Trying commit 2c1c85b with merge 1c3b6c4...

make the eval() functions on our const types return the resulting value This is a part of rust-lang#115748 that's hopefully perf-neutral, and that does not depend on rust-lang#115764.

bors · 2023-09-13T06:59:59Z

☀️ Try build successful - checks-actions
Build commit: 1c3b6c4 (1c3b6c4ebfd55a4d61ff506771dcd8cc9eaa76f1)

bors · 2023-09-13T06:59:59Z

☀️ Try build successful - checks-actions
Build commit: 1c3b6c4 (1c3b6c4ebfd55a4d61ff506771dcd8cc9eaa76f1)

rust-timer · 2023-09-13T09:27:59Z

Finished benchmarking commit (1c3b6c4): comparison URL.

Overall result: no relevant changes - no action needed

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-perf -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)	3.9%	[3.9%, 3.9%]	1
Regressions ❌ (secondary)	0.6%	[0.6%, 0.6%]	1
Improvements ✅ (primary)	-2.8%	[-4.1%, -1.5%]	2
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-0.6%	[-4.1%, 3.9%]	3

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)	-	-	0
Regressions ❌ (secondary)	2.2%	[1.9%, 2.4%]	2
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

Binary size

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

Bootstrap: 631.557s -> 631.698s (0.02%)
Artifact size: 317.99 MiB -> 317.79 MiB (-0.06%)

RalfJung · 2023-09-13T09:43:13Z

Nice, so that prepare_for_eval function seems to be the key then. :)

Should be ready to land. This can be further cleaned up, but given how perf-sensitive this is I like landing this in small chunks that can be separately benchmarked.

Further cleanup would be:

In MIR const fn eval, call the ty const fn eval and then do valtree-to-value conversion. But this changes the data flow through our queries, unlike the current PR which should be a complete NOP in terms of queries.
Also use the new fn eval in the interpreter. This needs some ConstValue refactoring #115764 to avoid fresh AllocId generation from going via ConstValue instead of directly via raw constants.

cjgillot · 2023-09-13T10:00:23Z

compiler/rustc_middle/src/mir/mod.rs

+
+    /// Normalizes the constant to a value or an error if possible.
+    #[inline]
+    pub fn normalize(self, tcx: TyCtxt<'tcx>, param_env: ty::ParamEnv<'tcx>) -> Self {


Do we want this method to translate a ValTree into a ConstValue? For normalization, it may be more interesting to keep the ValTree unchanged, as it is already a legal value.

This is a MIR constant, its normal form is a ConstValue. (Also this matches prior behavior.)

If we leave this as a valtree, then every future call to eval has to do the conversion again. That seems like a lot of wasted effort.

oli-obk · 2023-09-13T14:40:56Z

@bors r+

bors · 2023-09-13T14:40:57Z

📌 Commit 8ef6b7a has been approved by oli-obk

It is now in the queue for this repository.

bors · 2023-09-13T14:41:58Z

⌛ Testing commit 8ef6b7a with merge 76e59c7...

bors · 2023-09-13T16:30:00Z

☀️ Test successful - checks-actions
Approved by: oli-obk
Pushing 76e59c7 to master...

rust-timer · 2023-09-14T00:34:03Z

Finished benchmarking commit (76e59c7): 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)	1.0%	[0.5%, 1.5%]	2
Improvements ✅ (primary)	-1.9%	[-2.1%, -1.7%]	4
Improvements ✅ (secondary)	-3.2%	[-3.2%, -3.2%]	1
All ❌✅ (primary)	-1.9%	[-2.1%, -1.7%]	4

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)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-1.0%	[-1.0%, -1.0%]	1
Improvements ✅ (secondary)	-2.8%	[-3.1%, -2.5%]	2
All ❌✅ (primary)	-1.0%	[-1.0%, -1.0%]	1

Binary size

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

Bootstrap: 632.169s -> 632.671s (0.08%)
Artifact size: 317.92 MiB -> 317.85 MiB (-0.02%)

rustbot assigned b-naber Sep 12, 2023

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. T-compiler Relevant to the compiler team, which will review and decide on the PR/issue. T-rustdoc Relevant to the rustdoc team, which will review and decide on the PR/issue. labels Sep 12, 2023

RalfJung commented Sep 12, 2023

View reviewed changes

This comment has been minimized.

Sign in to view

rustbot assigned oli-obk and unassigned b-naber Sep 12, 2023

rustbot added the S-waiting-on-perf Status: Waiting on a perf run to be completed. label Sep 12, 2023

This comment has been minimized.

Sign in to view

RalfJung mentioned this pull request Sep 12, 2023

Debug info for const generics is hashes rather than something nice #115786

Open

RalfJung force-pushed the const-eval branch from 35fa022 to a151f93 Compare September 12, 2023 22:31

RalfJung commented Sep 12, 2023

View reviewed changes

compiler/rustc_codegen_ssa/src/debuginfo/type_names.rs Show resolved Hide resolved

RalfJung force-pushed the const-eval branch from a151f93 to 6086a68 Compare September 12, 2023 22:35

This comment has been minimized.

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 Sep 13, 2023

RalfJung commented Sep 13, 2023

View reviewed changes

RalfJung added 2 commits September 13, 2023 07:29

make the eval() functions on our const types return the resulting value

6e4779a

make the set of methods between our two Const types more consistent

11a4a24

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 Sep 13, 2023

This comment has been minimized.

Sign in to view

rustbot removed S-waiting-on-perf Status: Waiting on a perf run to be completed. perf-regression Performance regression. labels Sep 13, 2023

restore the old logic adjusting ty::UnevaluatedConst before evaluation

8ef6b7a

RalfJung force-pushed the const-eval branch from 2c1c85b to 8ef6b7a Compare September 13, 2023 09:47

cjgillot reviewed Sep 13, 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 Sep 13, 2023

bors added the merged-by-bors This PR was explicitly merged by bors. label Sep 13, 2023

bors merged commit 76e59c7 into rust-lang:master Sep 13, 2023
12 checks passed

rustbot added this to the 1.74.0 milestone Sep 13, 2023

This was referenced Sep 13, 2023

Remove -Zfuel. #115293

Open

Safe Transmute: Fix ICE (Inconsistent is_transmutable result) #113867

Closed

RalfJung deleted the const-eval branch September 14, 2023 05:22

RalfJung mentioned this pull request Sep 14, 2023

consistently pass ty::Const through valtrees #115804

Merged

RalfJung mentioned this pull request Nov 26, 2023

ICE error in Rust compiler: ty::Const::eval called on a non-valtree-compatible type #118285

Closed

Nadrieril mentioned this pull request May 28, 2024

Bump rustc version hacspec/hax#690

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

make the eval() functions on our const types return the resulting value #115803

make the eval() functions on our const types return the resulting value #115803

RalfJung commented Sep 12, 2023

rustbot commented Sep 12, 2023

rustbot commented Sep 12, 2023

RalfJung Sep 12, 2023

RalfJung commented Sep 12, 2023

This comment has been minimized.

bors commented Sep 12, 2023

This comment has been minimized.

RalfJung commented Sep 12, 2023

bors commented Sep 12, 2023

bors commented Sep 12, 2023

This comment has been minimized.

rust-timer commented Sep 13, 2023

RalfJung Sep 13, 2023

This comment has been minimized.

bors commented Sep 13, 2023

bors commented Sep 13, 2023

bors commented Sep 13, 2023

This comment has been minimized.

rust-timer commented Sep 13, 2023

RalfJung commented Sep 13, 2023

cjgillot Sep 13, 2023

RalfJung Sep 13, 2023 •

edited

Loading

oli-obk commented Sep 13, 2023

bors commented Sep 13, 2023

bors commented Sep 13, 2023

bors commented Sep 13, 2023

rust-timer commented Sep 14, 2023

make the eval() functions on our const types return the resulting value #115803

make the eval() functions on our const types return the resulting value #115803

Conversation

RalfJung commented Sep 12, 2023

rustbot commented Sep 12, 2023

rustbot commented Sep 12, 2023

Choose a reason for hiding this comment

RalfJung commented Sep 12, 2023

This comment has been minimized.

bors commented Sep 12, 2023

This comment has been minimized.

RalfJung commented Sep 12, 2023

bors commented Sep 12, 2023

bors commented Sep 12, 2023

This comment has been minimized.

rust-timer commented Sep 13, 2023

Overall result: ❌ regressions - ACTION NEEDED

Choose a reason for hiding this comment

This comment has been minimized.

bors commented Sep 13, 2023

bors commented Sep 13, 2023

bors commented Sep 13, 2023

This comment has been minimized.

rust-timer commented Sep 13, 2023

Overall result: no relevant changes - no action needed

RalfJung commented Sep 13, 2023

Choose a reason for hiding this comment

RalfJung Sep 13, 2023 • edited Loading

Choose a reason for hiding this comment

oli-obk commented Sep 13, 2023

bors commented Sep 13, 2023

bors commented Sep 13, 2023

bors commented Sep 13, 2023

rust-timer commented Sep 14, 2023

Overall result: no relevant changes - no action needed

RalfJung Sep 13, 2023 •

edited

Loading