Rework instantiation mode selection in monomorphization #128118

saethlin · 2024-07-23T22:32:37Z

This PR is extremely cooked because I'm just dumping ideas into it that seem good and can bootstrap the compiler without linker errors.

I don't know if all the ideas are good yet.

r? @ghost

saethlin · 2024-07-23T23:54:32Z

@bors try @rust-timer queue

…try> Don't add inlinehint to big closures In an unoptimized incremental build we want as many items as possible to be lowered as GloballyShared, but right now monomorphization makes whole classes of instances all LocalCopy without inspecting the instance's MIR or the session settings. Right now I'm experimenting to see what kind of an impact this can have. I think there's a pretty core design flaw currently because we make monomorphization decisions about Instances, but `cross_crate_inlinable` takes a DefId. r? `@ghost`

bors · 2024-07-23T23:55:44Z

⌛ Trying commit c47d3f4 with merge b53ccf2...

bors · 2024-07-24T01:49:21Z

☀️ Try build successful - checks-actions
Build commit: b53ccf2 (b53ccf21a2cbc0afcf4a3975eb97d4edbdbc3ed9)

rust-timer · 2024-07-24T03:06:06Z

Finished benchmarking commit (b53ccf2): 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)	3.0%	[0.2%, 34.9%]	38
Regressions ❌ (secondary)	2.2%	[0.2%, 9.9%]	25
Improvements ✅ (primary)	-0.6%	[-1.5%, -0.2%]	16
Improvements ✅ (secondary)	-33.3%	[-66.4%, -0.2%]	2
All ❌✅ (primary)	2.0%	[-1.5%, 34.9%]	54

Max RSS (memory usage)

Results (primary -0.6%, secondary 2.8%)

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)	5.2%	[4.0%, 6.5%]	3
Regressions ❌ (secondary)	2.8%	[2.8%, 2.8%]	1
Improvements ✅ (primary)	-3.5%	[-7.7%, -2.0%]	6
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-0.6%	[-7.7%, 6.5%]	9

Cycles

Results (primary 5.8%, secondary -2.0%)

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)	5.8%	[1.6%, 37.9%]	20
Regressions ❌ (secondary)	5.1%	[2.1%, 10.0%]	9
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-34.3%	[-66.6%, -2.1%]	2
All ❌✅ (primary)	5.8%	[1.6%, 37.9%]	20

Binary size

Results (primary 2.3%, secondary 2.0%)

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)	2.9%	[0.0%, 7.7%]	56
Regressions ❌ (secondary)	3.0%	[0.0%, 10.7%]	27
Improvements ✅ (primary)	-0.0%	[-0.2%, -0.0%]	13
Improvements ✅ (secondary)	-0.2%	[-0.6%, -0.0%]	11
All ❌✅ (primary)	2.3%	[-0.2%, 7.7%]	69

Bootstrap: 771.348s -> 777.681s (0.82%)
Artifact size: 328.86 MiB -> 328.99 MiB (0.04%)

saethlin · 2024-07-26T21:55:49Z

@bors try

bors · 2024-07-26T21:57:00Z

⌛ Trying commit 14b76c9 with merge e5cfe0b...

…try> Rework instantiation mode selection in monomorphization This PR is _extremely_ cooked because I'm just dumping ideas into it that seem good and can bootstrap the compiler without linker errors. I don't know if all the ideas are good yet. r? `@ghost`

bors · 2024-07-26T23:47:40Z

☀️ Try build successful - checks-actions
Build commit: e5cfe0b (e5cfe0b0eee8ec2cba6584d4de6cebd69307978b)

saethlin · 2024-07-27T00:04:05Z

@rust-timer build e5cfe0b

rust-timer · 2024-07-27T01:17:46Z

Finished benchmarking commit (e5cfe0b): comparison URL.

Overall result: ❌✅ regressions and improvements - BENCHMARK(S) FAILED

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

❗ ❗ ❗ ❗ ❗
Warning ⚠️: The following benchmark(s) failed to build:

cargo-0.60.0

❗ ❗ ❗ ❗ ❗

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)	12.9%	[0.1%, 175.6%]	130
Regressions ❌ (secondary)	4.9%	[0.2%, 78.6%]	62
Improvements ✅ (primary)	-7.4%	[-28.0%, -0.5%]	15
Improvements ✅ (secondary)	-7.9%	[-67.2%, -0.4%]	13
All ❌✅ (primary)	10.8%	[-28.0%, 175.6%]	145

Max RSS (memory usage)

Results (primary 6.6%, secondary 6.0%)

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)	8.0%	[1.4%, 28.2%]	51
Regressions ❌ (secondary)	7.6%	[3.2%, 13.2%]	13
Improvements ✅ (primary)	-7.8%	[-17.0%, -2.9%]	5
Improvements ✅ (secondary)	-4.5%	[-5.1%, -4.0%]	2
All ❌✅ (primary)	6.6%	[-17.0%, 28.2%]	56

Cycles

Results (primary 15.4%, secondary 5.9%)

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)	17.8%	[0.9%, 187.8%]	99
Regressions ❌ (secondary)	10.0%	[2.3%, 69.0%]	34
Improvements ✅ (primary)	-10.4%	[-23.8%, -2.6%]	9
Improvements ✅ (secondary)	-14.0%	[-67.4%, -2.2%]	7
All ❌✅ (primary)	15.4%	[-23.8%, 187.8%]	108

Binary size

Results (primary 12.1%, secondary 3.4%)

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)	13.6%	[0.1%, 80.9%]	134
Regressions ❌ (secondary)	6.7%	[0.0%, 76.4%]	66
Improvements ✅ (primary)	-1.1%	[-8.3%, -0.0%]	16
Improvements ✅ (secondary)	-4.0%	[-11.1%, -0.0%]	29
All ❌✅ (primary)	12.1%	[-8.3%, 80.9%]	150

Bootstrap: 769.319s -> 765.958s (-0.44%)
Artifact size: 328.91 MiB -> 329.40 MiB (0.15%)

saethlin · 2024-07-27T17:27:15Z

@bors try @rust-timer queue

…try> Rework instantiation mode selection in monomorphization This PR is _extremely_ cooked because I'm just dumping ideas into it that seem good and can bootstrap the compiler without linker errors. I don't know if all the ideas are good yet. r? `@ghost`

bors · 2024-07-27T17:28:26Z

⌛ Trying commit 4f0c766 with merge 422545f...

rust-log-analyzer · 2024-07-27T17:52:20Z

The job x86_64-gnu-llvm-17 failed! Check out the build log: (web) (plain)

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

------
 > importing cache manifest from ghcr.io/rust-lang/rust-ci-cache:3aacb9c90579defe09351ac5e8ee504359f8054da6326ff19038f1b7c90e3cb2aafe33685c6d9b76ee8d2ccbd187ca80c46ab5380485abdd8c0ce7d69cd8d8fd:
------
##[endgroup]
Setting extra environment values for docker:  --env ENABLE_GCC_CODEGEN=1 --env GCC_EXEC_PREFIX=/usr/lib/gcc/
[CI_JOB_NAME=x86_64-gnu-llvm-17]
---
sccache: Starting the server...
##[group]Configure the build
configure: processing command line
configure: 
configure: build.configure-args := ['--build=x86_64-unknown-linux-gnu', '--llvm-root=/usr/lib/llvm-17', '--enable-llvm-link-shared', '--set', 'rust.thin-lto-import-instr-limit=10', '--set', 'change-id=99999999', '--enable-verbose-configure', '--enable-sccache', '--disable-manage-submodules', '--enable-locked-deps', '--enable-cargo-native-static', '--set', 'rust.codegen-units-std=1', '--set', 'dist.compression-profile=balanced', '--dist-compression-formats=xz', '--set', 'rust.lld=false', '--disable-dist-src', '--release-channel=nightly', '--enable-debug-assertions', '--enable-overflow-checks', '--enable-llvm-assertions', '--set', 'rust.verify-llvm-ir', '--set', 'rust.codegen-backends=llvm,cranelift,gcc', '--set', 'llvm.static-libstdcpp', '--enable-new-symbol-mangling']
configure: target.x86_64-unknown-linux-gnu.llvm-config := /usr/lib/llvm-17/bin/llvm-config
configure: llvm.link-shared     := True
configure: rust.thin-lto-import-instr-limit := 10
configure: change-id            := 99999999
---
failures:

---- [codegen] tests/codegen/closures-without-inlinehint.rs stdout ----

error: verification with 'FileCheck' failed
status: exit status: 2
command: "/usr/lib/llvm-17/bin/FileCheck" "--input-file" "/checkout/obj/build/x86_64-unknown-linux-gnu/test/codegen/closures-without-inlinehint/closures-without-inlinehint.ll" "/checkout/tests/codegen/closures-without-inlinehint.rs" "--check-prefix=CHECK" "--check-prefix" "NONMSVC" "--allow-unused-prefixes" "--dump-input-context" "100"
--- stderr -------------------------------
--- stderr -------------------------------
Could not open input file '/checkout/obj/build/x86_64-unknown-linux-gnu/test/codegen/closures-without-inlinehint/closures-without-inlinehint.ll': No such file or directory



failures:

bors · 2024-07-27T19:25:18Z

☀️ Try build successful - checks-actions
Build commit: 422545f (422545f89f7bcffb5262a4211a7d98413e44161d)

rust-timer · 2024-07-27T21:18:41Z

Finished benchmarking commit (422545f): 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)	66.4%	[0.2%, 722.4%]	168
Regressions ❌ (secondary)	37.0%	[0.2%, 592.5%]	115
Improvements ✅ (primary)	-12.1%	[-72.2%, -0.2%]	19
Improvements ✅ (secondary)	-13.9%	[-44.9%, -0.3%]	7
All ❌✅ (primary)	58.4%	[-72.2%, 722.4%]	187

Max RSS (memory usage)

Results (primary 23.1%, secondary 4.0%)

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)	27.0%	[1.0%, 269.4%]	93
Regressions ❌ (secondary)	7.5%	[2.1%, 18.7%]	19
Improvements ✅ (primary)	-9.7%	[-28.4%, -2.2%]	11
Improvements ✅ (secondary)	-9.3%	[-17.7%, -4.5%]	5
All ❌✅ (primary)	23.1%	[-28.4%, 269.4%]	104

Cycles

Results (primary 82.9%, secondary 65.7%)

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)	90.5%	[0.9%, 724.9%]	138
Regressions ❌ (secondary)	69.8%	[1.5%, 659.3%]	70
Improvements ✅ (primary)	-21.3%	[-73.7%, -1.3%]	10
Improvements ✅ (secondary)	-29.4%	[-41.8%, -5.3%]	3
All ❌✅ (primary)	82.9%	[-73.7%, 724.9%]	148

Binary size

Results (primary 159.5%, secondary 54.7%)

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)	170.2%	[0.1%, 924.5%]	151
Regressions ❌ (secondary)	62.3%	[0.0%, 367.7%]	104
Improvements ✅ (primary)	-1.5%	[-8.3%, -0.0%]	10
Improvements ✅ (secondary)	-1.5%	[-8.0%, -0.0%]	14
All ❌✅ (primary)	159.5%	[-8.3%, 924.5%]	161

Bootstrap: 772.472s -> 768.225s (-0.55%)
Artifact size: 329.05 MiB -> 328.80 MiB (-0.08%)

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 Jul 23, 2024

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 Jul 23, 2024

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 Jul 24, 2024

saethlin force-pushed the closures-can-be-shared branch from c47d3f4 to 14b76c9 Compare July 26, 2024 21:49

saethlin changed the title ~~Don't add inlinehint to big closures~~ Rework instantiation mode selection in monomorphization Jul 26, 2024

This comment has been minimized.

Sign in to view

saethlin added 3 commits July 27, 2024 08:04

Rework instantiation mode selection

91fe3b7

Fix gcc

aa82970

Change CGU partitioning to put one item in each CGU

5254f98

saethlin force-pushed the closures-can-be-shared branch from 14b76c9 to 5254f98 Compare July 27, 2024 17:17

saethlin added 2 commits July 27, 2024 13:18

Fix

c042073

Fix

4f0c766

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 Jul 27, 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 Jul 27, 2024

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Rework instantiation mode selection in monomorphization #128118

Rework instantiation mode selection in monomorphization #128118

saethlin commented Jul 23, 2024 •

edited

Loading

saethlin commented Jul 23, 2024

This comment has been minimized.

bors commented Jul 23, 2024

bors commented Jul 24, 2024

This comment has been minimized.

rust-timer commented Jul 24, 2024

saethlin commented Jul 26, 2024

bors commented Jul 26, 2024

This comment has been minimized.

bors commented Jul 26, 2024

saethlin commented Jul 27, 2024

This comment has been minimized.

rust-timer commented Jul 27, 2024

saethlin commented Jul 27, 2024

This comment has been minimized.

bors commented Jul 27, 2024

rust-log-analyzer commented Jul 27, 2024

bors commented Jul 27, 2024

This comment has been minimized.

rust-timer commented Jul 27, 2024

Rework instantiation mode selection in monomorphization #128118

Are you sure you want to change the base?

Rework instantiation mode selection in monomorphization #128118

Conversation

saethlin commented Jul 23, 2024 • edited Loading

saethlin commented Jul 23, 2024

This comment has been minimized.

bors commented Jul 23, 2024

bors commented Jul 24, 2024

This comment has been minimized.

rust-timer commented Jul 24, 2024

Overall result: ❌✅ regressions and improvements - ACTION NEEDED

saethlin commented Jul 26, 2024

bors commented Jul 26, 2024

This comment has been minimized.

bors commented Jul 26, 2024

saethlin commented Jul 27, 2024

This comment has been minimized.

rust-timer commented Jul 27, 2024

Overall result: ❌✅ regressions and improvements - BENCHMARK(S) FAILED

saethlin commented Jul 27, 2024

This comment has been minimized.

bors commented Jul 27, 2024

rust-log-analyzer commented Jul 27, 2024

bors commented Jul 27, 2024

This comment has been minimized.

rust-timer commented Jul 27, 2024

Overall result: ❌✅ regressions and improvements - ACTION NEEDED

saethlin commented Jul 23, 2024 •

edited

Loading