Perform opportunistic simplifications during value numbering #111344

cjgillot · 2023-05-08T09:07:29Z

~~Based on #109597~~
~~Based on #119439~~

Opening mostly for discussion.

rustbot · 2023-05-08T09:07:35Z

r? @petrochenkov

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

cjgillot · 2023-05-08T14:11:57Z

@bors try @rust-timer queue

bors · 2023-05-08T14:12:07Z

⌛ Trying commit 7e731ee19e9a11414f711e82aa3f220342a97532 with merge d0c89341fa5520a74ff1a0c782163343027a602d...

bors · 2023-05-08T15:51:39Z

☀️ Try build successful - checks-actions
Build commit: d0c89341fa5520a74ff1a0c782163343027a602d (d0c89341fa5520a74ff1a0c782163343027a602d)

rust-timer · 2023-05-08T17:07:02Z

Finished benchmarking commit (d0c89341fa5520a74ff1a0c782163343027a602d): comparison URL.

Overall result: ❌✅ regressions and improvements - 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)	0.9%	[0.4%, 1.8%]	6
Regressions ❌ (secondary)	3.0%	[0.2%, 24.1%]	9
Improvements ✅ (primary)	-0.5%	[-1.1%, -0.2%]	75
Improvements ✅ (secondary)	-0.6%	[-3.1%, -0.1%]	42
All ❌✅ (primary)	-0.4%	[-1.1%, 1.8%]	81

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)	5.3%	[2.4%, 8.5%]	6
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-2.2%	[-3.9%, -1.1%]	12
Improvements ✅ (secondary)	-2.7%	[-4.3%, -1.4%]	5
All ❌✅ (primary)	0.3%	[-3.9%, 8.5%]	18

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)	1.4%	[0.7%, 2.0%]	4
Regressions ❌ (secondary)	16.3%	[15.5%, 17.1%]	2
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-3.1%	[-3.1%, -3.1%]	1
All ❌✅ (primary)	1.4%	[0.7%, 2.0%]	4

Binary size

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)	0.4%	[0.0%, 1.5%]	11
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-0.6%	[-5.2%, -0.0%]	99
Improvements ✅ (secondary)	-0.8%	[-4.0%, -0.1%]	32
All ❌✅ (primary)	-0.5%	[-5.2%, 1.5%]	110

Bootstrap: 654.499s -> 657.369s (0.44%)

bugadani · 2023-05-08T19:15:19Z

Optimization contributors just have to have a special place in their hearts for keccak.

bors · 2023-05-17T01:56:01Z

☔ The latest upstream changes (presumably #111556) made this pull request unmergeable. Please resolve the merge conflicts.

…illot,tmiasko Preprocess and cache dominator tree Preprocessing dominators has a very strong effect for rust-lang#111344. That pass checks that assignments dominate their uses repeatedly. Using the unprocessed dominator tree caused a quadratic runtime (number of bbs x depth of the dominator tree). This PR also caches the dominator tree and the pre-processed dominators in the MIR cfg cache. Rebase of rust-lang#107157 cc `@tmiasko`

…asko Preprocess and cache dominator tree Preprocessing dominators has a very strong effect for rust-lang/rust#111344. That pass checks that assignments dominate their uses repeatedly. Using the unprocessed dominator tree caused a quadratic runtime (number of bbs x depth of the dominator tree). This PR also caches the dominator tree and the pre-processed dominators in the MIR cfg cache. Rebase of rust-lang/rust#107157 cc `@tmiasko`

bors · 2023-06-14T18:09:08Z

☔ The latest upstream changes (presumably #112418) made this pull request unmergeable. Please resolve the merge conflicts.

bors · 2023-06-19T08:12:29Z

☔ The latest upstream changes (presumably #112724) made this pull request unmergeable. Please resolve the merge conflicts.

bors · 2023-12-30T14:37:14Z

⌛ Trying commit 935e283 with merge 01915ea...

Perform opportunistic simplifications during value numbering Based on rust-lang#109597 Opening mostly for discussion. In its current form, I think this pass does too much. I want to remove the const-propagation part to make it simpler.

bors · 2023-12-30T16:03:22Z

☀️ Try build successful - checks-actions
Build commit: 01915ea (01915ea9cc2476cc183f598c9e9092878bedc23c)

rust-timer · 2023-12-30T17:39:22Z

Finished benchmarking commit (01915ea): comparison URL.

Overall result: ❌✅ regressions and improvements - 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)	0.5%	[0.3%, 0.8%]	7
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-0.6%	[-1.0%, -0.2%]	9
Improvements ✅ (secondary)	-0.8%	[-1.2%, -0.4%]	11
All ❌✅ (primary)	-0.1%	[-1.0%, 0.8%]	16

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)	3.0%	[0.7%, 6.8%]	5
Regressions ❌ (secondary)	3.8%	[0.5%, 7.2%]	10
Improvements ✅ (primary)	-6.1%	[-15.8%, -0.5%]	5
Improvements ✅ (secondary)	-3.0%	[-4.7%, -1.6%]	5
All ❌✅ (primary)	-1.6%	[-15.8%, 6.8%]	10

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)	1.0%	[1.0%, 1.0%]	3
Regressions ❌ (secondary)	2.1%	[2.0%, 2.1%]	2
Improvements ✅ (primary)	-0.5%	[-0.6%, -0.4%]	3
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.3%	[-0.6%, 1.0%]	6

Binary size

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)	0.4%	[0.0%, 3.7%]	55
Regressions ❌ (secondary)	0.6%	[0.0%, 1.3%]	6
Improvements ✅ (primary)	-0.3%	[-1.4%, -0.0%]	13
Improvements ✅ (secondary)	-0.7%	[-3.9%, -0.2%]	21
All ❌✅ (primary)	0.3%	[-1.4%, 3.7%]	68

Bootstrap: 668.324s -> 669.679s (0.20%)
Artifact size: 311.76 MiB -> 311.83 MiB (0.02%)

Implement constant propagation on top of MIR SSA analysis This implements the idea I proposed in rust-lang/rust#110719 (comment) Based on rust-lang/rust#109597 The value numbering "GVN" pass formulates each rvalue that appears in MIR with an abstract form (the `Value` enum), and assigns an integer `VnIndex` to each. This abstract form can be used to deduplicate values, reusing an earlier local that holds the same value instead of recomputing. This part is proposed in #109597. From this abstract representation, we can perform more involved simplifications, for example in rust-lang/rust#111344. With the abstract representation `Value`, we can also attempt to evaluate each to a constant using the interpreter. This builds a `VnIndex -> OpTy` map. From this map, we can opportunistically replace an operand or a rvalue with a constant if their value has an associated `OpTy`. The most relevant commit is [Evaluated computed values to constants.](rust-lang/rust@2767c49)" r? `@oli-obk`

Fold arithmetic identities in GVN Extracted from rust-lang#111344 This PR implements a few arithmetic folds for unary and binary operations. This should take care of the missed optimizations introduced by rust-lang#116012.

Fold arithmetic identities in GVN Extracted from rust-lang/rust#111344 This PR implements a few arithmetic folds for unary and binary operations. This should take care of the missed optimizations introduced by rust-lang/rust#116012.

cjgillot · 2024-01-20T00:18:16Z

@bors try @rust-timer queue

Perform opportunistic simplifications during value numbering ~Based on rust-lang#109597 ~Based on rust-lang#119439 Opening mostly for discussion.

bors · 2024-01-20T00:19:26Z

⌛ Trying commit 0167761 with merge 2cd7dda...

bors · 2024-01-20T01:44:56Z

☀️ Try build successful - checks-actions
Build commit: 2cd7dda (2cd7ddaa6c29d00461f1185aaaba85330f616648)

rust-timer · 2024-01-20T02:59:49Z

Finished benchmarking commit (2cd7dda): comparison URL.

Overall result: ❌✅ regressions and improvements - 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)	0.7%	[0.7%, 0.7%]	2
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-0.6%	[-0.9%, -0.4%]	5
Improvements ✅ (secondary)	-1.0%	[-1.3%, -0.7%]	6
All ❌✅ (primary)	-0.2%	[-0.9%, 0.7%]	7

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)	3.7%	[0.2%, 6.1%]	6
Regressions ❌ (secondary)	1.6%	[0.8%, 2.2%]	4
Improvements ✅ (primary)	-5.0%	[-10.3%, -1.6%]	4
Improvements ✅ (secondary)	-3.6%	[-3.6%, -3.6%]	1
All ❌✅ (primary)	0.3%	[-10.3%, 6.1%]	10

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)	-	-	0
Improvements ✅ (secondary)	-1.8%	[-1.8%, -1.8%]	1
All ❌✅ (primary)	-	-	0

Binary size

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)	0.9%	[0.4%, 3.6%]	12
Regressions ❌ (secondary)	0.8%	[0.6%, 1.3%]	6
Improvements ✅ (primary)	-1.8%	[-1.8%, -1.8%]	1
Improvements ✅ (secondary)	-1.0%	[-2.2%, -0.6%]	4
All ❌✅ (primary)	0.7%	[-1.8%, 3.6%]	13

Bootstrap: 663.584s -> 667.018s (0.52%)
Artifact size: 308.31 MiB -> 308.34 MiB (0.01%)

rustbot assigned petrochenkov May 8, 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. labels May 8, 2023

cjgillot unassigned petrochenkov May 8, 2023

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 May 8, 2023

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 May 8, 2023

cjgillot force-pushed the gvn-simplify branch from 7e731ee to f18d0f7 Compare May 14, 2023 09:29

This comment has been minimized.

Sign in to view

cjgillot mentioned this pull request May 17, 2023

Preprocess and cache dominator tree #111673

Merged

cjgillot force-pushed the gvn-simplify branch from f18d0f7 to 5c1a123 Compare May 17, 2023 09:35

cjgillot force-pushed the gvn-simplify branch from 5c1a123 to 33bbb76 Compare June 2, 2023 15:30

This comment has been minimized.

Sign in to view

cjgillot force-pushed the gvn-simplify branch from 6e9f935 to 5c7433d Compare June 9, 2023 17:19

cjgillot force-pushed the gvn-simplify branch from 5c7433d to fdf770d Compare June 17, 2023 18:46

This comment has been minimized.

Sign in to view

rustbot removed the S-waiting-on-perf Status: Waiting on a perf run to be completed. label Dec 30, 2023

deltragon mentioned this pull request Jan 5, 2024

RangeInclusive performance regression in beta + nightly #119643

Closed

cjgillot mentioned this pull request Jan 6, 2024

Fold arithmetic identities in GVN #119670

Merged

cjgillot force-pushed the gvn-simplify branch from 935e283 to dd79fd1 Compare January 16, 2024 23:50

cjgillot force-pushed the gvn-simplify branch from dd79fd1 to 0167761 Compare January 19, 2024 22:52

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 Jan 20, 2024

This comment has been minimized.

Sign in to view

rustbot removed the S-waiting-on-perf Status: Waiting on a perf run to be completed. label Jan 20, 2024

Dylan-DPC added S-experimental Status: Ongoing experiment that does not require reviewing and won't be merged in its current state. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Mar 23, 2024

cjgillot added 2 commits April 7, 2024 14:29

Simplify operands in unions.

69d51fc

Perform GVN into debuginfo.

d272ab3

cjgillot force-pushed the gvn-simplify branch from 0167761 to d272ab3 Compare April 7, 2024 15:01

cjgillot closed this Jun 27, 2024

cjgillot deleted the gvn-simplify branch June 27, 2024 09:44

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Perform opportunistic simplifications during value numbering #111344

Perform opportunistic simplifications during value numbering #111344

cjgillot commented May 8, 2023 •

edited

Loading

rustbot commented May 8, 2023

cjgillot commented May 8, 2023

This comment has been minimized.

bors commented May 8, 2023

bors commented May 8, 2023

This comment has been minimized.

rust-timer commented May 8, 2023

bugadani commented May 8, 2023

This comment has been minimized.

bors commented May 17, 2023

This comment has been minimized.

bors commented Jun 14, 2023

This comment has been minimized.

bors commented Jun 19, 2023

bors commented Dec 30, 2023

bors commented Dec 30, 2023

This comment has been minimized.

rust-timer commented Dec 30, 2023

cjgillot commented Jan 20, 2024

This comment has been minimized.

bors commented Jan 20, 2024

bors commented Jan 20, 2024

This comment has been minimized.

rust-timer commented Jan 20, 2024

Perform opportunistic simplifications during value numbering #111344

Perform opportunistic simplifications during value numbering #111344

Conversation

cjgillot commented May 8, 2023 • edited Loading

rustbot commented May 8, 2023

cjgillot commented May 8, 2023

This comment has been minimized.

bors commented May 8, 2023

bors commented May 8, 2023

This comment has been minimized.

rust-timer commented May 8, 2023

Overall result: ❌✅ regressions and improvements - ACTION NEEDED

bugadani commented May 8, 2023

This comment has been minimized.

bors commented May 17, 2023

This comment has been minimized.

bors commented Jun 14, 2023

This comment has been minimized.

bors commented Jun 19, 2023

bors commented Dec 30, 2023

bors commented Dec 30, 2023

This comment has been minimized.

rust-timer commented Dec 30, 2023

Overall result: ❌✅ regressions and improvements - ACTION NEEDED

cjgillot commented Jan 20, 2024

This comment has been minimized.

bors commented Jan 20, 2024

bors commented Jan 20, 2024

This comment has been minimized.

rust-timer commented Jan 20, 2024

Overall result: ❌✅ regressions and improvements - ACTION NEEDED

cjgillot commented May 8, 2023 •

edited

Loading