Tweak the threshold for chunked swapping #111803

scottmcm · 2023-05-20T18:11:57Z

Thanks to @AngelicosPhosphoros for the tests here, which I copied from #98892.

This is an experiment as a simple alternative to that PR that just tweaks the existing threshold, since that PR showed that 3×Align (like String) currently doesn't work as well as it could.

rustbot · 2023-05-20T18:12:03Z

r? @Mark-Simulacrum

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

scottmcm · 2023-05-20T18:53:09Z

@bors try @rust-timer queue

bors · 2023-05-20T18:53:17Z

⌛ Trying commit b50eab5 with merge 54855997187eeec09d7258dff0c3c3f9cad15411...

bors · 2023-05-20T21:01:19Z

☀️ Try build successful - checks-actions
Build commit: 54855997187eeec09d7258dff0c3c3f9cad15411 (54855997187eeec09d7258dff0c3c3f9cad15411)

rust-timer · 2023-05-21T23:13:58Z

Finished benchmarking commit (54855997187eeec09d7258dff0c3c3f9cad15411): comparison URL.

Overall result: ❌ regressions - 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 is a highly reliable metric that was used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	0.3%	[0.3%, 0.3%]	2
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.3%	[0.3%, 0.3%]	2

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)	3.3%	[3.3%, 3.3%]	1
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-2.6%	[-5.2%, -0.1%]	2
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-0.7%	[-5.2%, 3.3%]	3

Cycles

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

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)	0.1%	[0.0%, 1.1%]	35
Regressions ❌ (secondary)	0.0%	[0.0%, 0.0%]	1
Improvements ✅ (primary)	-0.2%	[-0.4%, -0.1%]	9
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.0%	[-0.4%, 1.1%]	44

Bootstrap: 645.683s -> 642.174s (-0.54%)

scottmcm · 2023-05-22T05:22:02Z

Oh, I should have put this on the same reviewer as #98892 to compare between the options:
r? @cuviper

Mark-Simulacrum · 2024-01-20T02:03:10Z

I think we should land this unless things have changed (r=me) -- it seems like a much simpler approach than #98892 and is a good starting point. Tweaking the threshold to make LLVM behave better feels easier than landing a new underlying implementation.

AngelicosPhosphoros · 2024-01-20T02:36:40Z

@Mark-Simulacrum I agree. I would not have time to work on it in near future anyway.

Thanks to 98892 for the tests I brought in here, as it demonstrated that 3×usize is currently suboptimal.

scottmcm · 2024-01-20T07:02:35Z

Rebased to fix the tests. Since it's been over ½ a year, let's double-check

@bors try @rust-timer queue

bors · 2024-01-20T07:03:46Z

⌛ Trying commit 60208a0 with merge 4649ab6...

…<try> Tweak the threshold for chunked swapping Thanks to `@AngelicosPhosphoros` for the tests here, which I copied from rust-lang#98892. This is an experiment as a simple alternative to that PR that just tweaks the existing threshold, since that PR showed that 3×Align (like `String`) currently doesn't work as well as it could.

bors · 2024-01-20T08:29:21Z

☀️ Try build successful - checks-actions
Build commit: 4649ab6 (4649ab6a9f39b3f3fb3f066eb5a59f3cc3bd8312)

rust-timer · 2024-01-20T11:33:01Z

Finished benchmarking commit (4649ab6): comparison URL.

Overall result: ❌ regressions - 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 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
Regressions ❌ (secondary)	2.1%	[2.1%, 2.1%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

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)	4.8%	[2.3%, 6.8%]	4
Regressions ❌ (secondary)	1.4%	[1.0%, 1.8%]	2
Improvements ✅ (primary)	-4.9%	[-9.0%, -2.6%]	6
Improvements ✅ (secondary)	-4.5%	[-4.5%, -4.5%]	1
All ❌✅ (primary)	-1.0%	[-9.0%, 6.8%]	10

Cycles

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

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)	0.5%	[0.5%, 0.5%]	3
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.5%	[0.5%, 0.5%]	3

Bootstrap: 663.642s -> 662.371s (-0.19%)
Artifact size: 308.33 MiB -> 308.36 MiB (0.01%)

scottmcm · 2024-01-20T20:50:37Z

@bors r=Mark-Simulacrum

#111803 (comment)

bors · 2024-01-20T20:50:40Z

📌 Commit 60208a0 has been approved by Mark-Simulacrum

It is now in the queue for this repository.

bors · 2024-01-20T21:54:47Z

⌛ Testing commit 60208a0 with merge 4cb17b4...

bors · 2024-01-20T23:50:52Z

☀️ Test successful - checks-actions
Approved by: Mark-Simulacrum
Pushing 4cb17b4 to master...

rust-timer · 2024-01-21T01:06:22Z

Finished benchmarking commit (4cb17b4): comparison URL.

Overall result: ❌ regressions - 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.

	mean	range	count
Regressions ❌ (primary)	0.8%	[0.8%, 0.8%]	1
Regressions ❌ (secondary)	1.6%	[1.6%, 1.6%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.8%	[0.8%, 0.8%]	1

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)	2.4%	[1.2%, 3.9%]	3
Regressions ❌ (secondary)	2.5%	[2.5%, 2.5%]	1
Improvements ✅ (primary)	-6.4%	[-8.5%, -4.3%]	2
Improvements ✅ (secondary)	-2.4%	[-2.4%, -2.4%]	1
All ❌✅ (primary)	-1.1%	[-8.5%, 3.9%]	5

Cycles

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

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)	0.5%	[0.5%, 0.5%]	1
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-0.5%	[-0.5%, -0.5%]	1
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.0%	[-0.5%, 0.5%]	2

Bootstrap: 665.187s -> 661.791s (-0.51%)
Artifact size: 308.34 MiB -> 308.38 MiB (0.01%)

Let codegen decide when to `mem::swap` with immediates Making `libcore` decide this is silly; the backend has so much better information about when it's a good idea. Thus this PR introduces a new `typed_swap` intrinsic with a fallback body, and replaces that fallback implementation when swapping immediates or scalar pairs. r? oli-obk Replaces rust-lang#111744, and means we'll never need more libs PRs like rust-lang#111803 or rust-lang#107140

rustbot assigned Mark-Simulacrum May 20, 2023

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. T-libs Relevant to the library team, which will review and decide on the PR/issue. labels May 20, 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 20, 2023

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 May 21, 2023

rustbot assigned cuviper and unassigned Mark-Simulacrum May 22, 2023

scottmcm force-pushed the simple-swap-alternative branch from b50eab5 to e871acb Compare January 20, 2024 06:30

Tweak the threshold for chunked swapping

60208a0

Thanks to 98892 for the tests I brought in here, as it demonstrated that 3×usize is currently suboptimal.

scottmcm force-pushed the simple-swap-alternative branch from e871acb to 60208a0 Compare January 20, 2024 07:00

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

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 Jan 20, 2024

bors added the merged-by-bors This PR was explicitly merged by bors. label Jan 20, 2024

bors merged commit 4cb17b4 into rust-lang:master Jan 20, 2024
12 checks passed

rustbot added this to the 1.77.0 milestone Jan 20, 2024

scottmcm deleted the simple-swap-alternative branch January 22, 2024 11:09

celinval mentioned this pull request Jan 30, 2024

Upgrade Rust toolchain to nightly-2024-01-23 model-checking/kani#2983

Merged

scottmcm mentioned this pull request Mar 16, 2024

Let codegen decide when to mem::swap with immediates #122582

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Tweak the threshold for chunked swapping #111803

Tweak the threshold for chunked swapping #111803

scottmcm commented May 20, 2023

rustbot commented May 20, 2023

scottmcm commented May 20, 2023

This comment has been minimized.

bors commented May 20, 2023

bors commented May 20, 2023

This comment has been minimized.

rust-timer commented May 21, 2023

scottmcm commented May 22, 2023

Mark-Simulacrum commented Jan 20, 2024 •

edited

Loading

AngelicosPhosphoros commented Jan 20, 2024

scottmcm 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

scottmcm commented Jan 20, 2024

bors commented Jan 20, 2024

bors commented Jan 20, 2024

bors commented Jan 20, 2024

rust-timer commented Jan 21, 2024

Tweak the threshold for chunked swapping #111803

Tweak the threshold for chunked swapping #111803

Conversation

scottmcm commented May 20, 2023

rustbot commented May 20, 2023

scottmcm commented May 20, 2023

This comment has been minimized.

bors commented May 20, 2023

bors commented May 20, 2023

This comment has been minimized.

rust-timer commented May 21, 2023

Overall result: ❌ regressions - no action needed

scottmcm commented May 22, 2023

Mark-Simulacrum commented Jan 20, 2024 • edited Loading

AngelicosPhosphoros commented Jan 20, 2024

scottmcm 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 - no action needed

scottmcm commented Jan 20, 2024

bors commented Jan 20, 2024

bors commented Jan 20, 2024

bors commented Jan 20, 2024

rust-timer commented Jan 21, 2024

Overall result: ❌ regressions - no action needed

Mark-Simulacrum commented Jan 20, 2024 •

edited

Loading