Obligation forest tweaks by nnethercote · Pull Request #97674 · rust-lang/rust

nnethercote · 2022-06-03T06:54:57Z

A few minor improvements to the code.

r? @nikomatsakis

nnethercote · 2022-06-03T06:55:32Z

I'm not expecting much performance change, maybe a small improvement to keccak and one or two others.

@bors try @rust-timer queue

rust-timer · 2022-06-03T06:55:34Z

Awaiting bors try build completion.

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

bors · 2022-06-03T06:55:43Z

⌛ Trying commit f673eade735b3e85b1703b11f12c00d715a29c44 with merge 1e46fe42354430de9aca5fb2769e2766f4853903...

nnethercote · 2022-06-03T08:40:51Z

@bors try @rust-timer queue

rust-timer · 2022-06-03T08:40:53Z

Awaiting bors try build completion.

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

bors · 2022-06-03T08:41:00Z

⌛ Trying commit bb5005ff758306b0ed470915ff5cb0508d161967 with merge fbe2676ec277a531d1e97d27b8ab76f06129f797...

bors · 2022-06-03T10:15:29Z

☀️ Try build successful - checks-actions
Build commit: fbe2676ec277a531d1e97d27b8ab76f06129f797 (fbe2676ec277a531d1e97d27b8ab76f06129f797)

rust-timer · 2022-06-03T10:15:31Z

Queued fbe2676ec277a531d1e97d27b8ab76f06129f797 with parent f5507aa, future comparison URL.

rust-timer · 2022-06-03T13:51:12Z

Finished benchmarking commit (fbe2676ec277a531d1e97d27b8ab76f06129f797): comparison url.

Instruction count

Primary benchmarks: 😿 relevant regressions found
Secondary benchmarks: 😿 relevant regressions found

	mean¹	max	count²
Regressions 😿 (primary)	0.9%	0.9%	2
Regressions 😿 (secondary)	2.1%	2.4%	6
Improvements 🎉 (primary)	N/A	N/A	0
Improvements 🎉 (secondary)	N/A	N/A	0
All 😿🎉 (primary)	0.9%	0.9%	2

Max RSS (memory usage)

Results

Primary benchmarks: 😿 relevant regressions found
Secondary benchmarks: no relevant changes found

	mean¹	max	count²
Regressions 😿 (primary)	4.1%	5.8%	2
Regressions 😿 (secondary)	N/A	N/A	0
Improvements 🎉 (primary)	N/A	N/A	0
Improvements 🎉 (secondary)	N/A	N/A	0
All 😿🎉 (primary)	4.1%	5.8%	2

Cycles

Results

Primary benchmarks: mixed results
Secondary benchmarks: no relevant changes found

	mean¹	max	count²
Regressions 😿 (primary)	3.1%	3.1%	1
Regressions 😿 (secondary)	N/A	N/A	0
Improvements 🎉 (primary)	-3.6%	-3.6%	1
Improvements 🎉 (secondary)	N/A	N/A	0
All 😿🎉 (primary)	-0.2%	-3.6%	2

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.

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-review -S-waiting-on-perf +perf-regression

the arithmetic mean of the percent change↩↩² ↩³
number of relevant changes↩↩² ↩³

It is simpler if `ObligationForest` does this itself, rather than the caller having to manage it.

Because it really has two halves: - A read-only part that checks if further work is needed. - The further work part, which is much less hot. This makes things a bit clearer and nicer.

nnethercote · 2022-06-06T05:12:03Z

Ugh, this was a tiny perf win for me locally. PGO's effects again, I suspect. I will try a few more CI perf runs, try to work out the problem.

nnethercote · 2022-06-06T05:12:14Z

@bors try @rust-timer queue

rust-timer · 2022-06-06T05:12:16Z

Awaiting bors try build completion.

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

bors · 2022-06-06T05:12:22Z

⌛ Trying commit 32741d5 with merge 991b508f88d9c528d5dd375f43c1cececea50473...

bors · 2022-06-06T06:46:18Z

☀️ Try build successful - checks-actions
Build commit: 991b508f88d9c528d5dd375f43c1cececea50473 (991b508f88d9c528d5dd375f43c1cececea50473)

rust-timer · 2022-06-06T06:46:20Z

Queued 991b508f88d9c528d5dd375f43c1cececea50473 with parent 760237f, future comparison URL.

rust-timer · 2022-06-06T08:03:45Z

Finished benchmarking commit (991b508f88d9c528d5dd375f43c1cececea50473): comparison url.

Instruction count

Primary benchmarks: 🎉 relevant improvements found
Secondary benchmarks: 🎉 relevant improvements found

	mean¹	max	count²
Regressions 😿 (primary)	N/A	N/A	0
Regressions 😿 (secondary)	N/A	N/A	0
Improvements 🎉 (primary)	-0.5%	-0.7%	3
Improvements 🎉 (secondary)	-1.5%	-2.0%	7
All 😿🎉 (primary)	-0.5%	-0.7%	3

Max RSS (memory usage)

Results

Primary benchmarks: 😿 relevant regressions found
Secondary benchmarks: no relevant changes found

	mean¹	max	count²
Regressions 😿 (primary)	1.6%	3.1%	2
Regressions 😿 (secondary)	N/A	N/A	0
Improvements 🎉 (primary)	N/A	N/A	0
Improvements 🎉 (secondary)	N/A	N/A	0
All 😿🎉 (primary)	1.6%	3.1%	2

Cycles

Results

Primary benchmarks: no relevant changes found
Secondary benchmarks: mixed results

	mean¹	max	count²
Regressions 😿 (primary)	N/A	N/A	0
Regressions 😿 (secondary)	2.2%	2.2%	1
Improvements 🎉 (primary)	N/A	N/A	0
Improvements 🎉 (secondary)	-3.8%	-4.0%	2
All 😿🎉 (primary)	N/A	N/A	0

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

the arithmetic mean of the percent change↩↩² ↩³
number of relevant changes↩↩² ↩³

nnethercote · 2022-06-06T09:24:45Z

Hooray, those new perf results are much better. @nikomatsakis, review away please! :)

nikomatsakis · 2022-06-20T08:57:44Z

@bors r+

bors · 2022-06-20T08:57:46Z

📌 Commit 32741d5 has been approved by nikomatsakis

bors · 2022-06-20T10:58:59Z

⌛ Testing commit 32741d5 with merge 1d60108...

bors · 2022-06-20T13:24:25Z

☀️ Test successful - checks-actions
Approved by: nikomatsakis
Pushing 1d60108 to master...

rust-timer · 2022-06-20T14:41:23Z

Finished benchmarking commit (1d60108): comparison url.

Instruction count

Primary benchmarks: 😿 relevant regressions found
Secondary benchmarks: 😿 relevant regressions found

	mean¹	max	count²
Regressions 😿 (primary)	0.8%	0.9%	2
Regressions 😿 (secondary)	1.8%	2.4%	7
Improvements 🎉 (primary)	N/A	N/A	0
Improvements 🎉 (secondary)	-0.3%	-0.3%	1
All 😿🎉 (primary)	0.8%	0.9%	2

Max RSS (memory usage)

Results

Primary benchmarks: no relevant changes found
Secondary benchmarks: mixed results

	mean¹	max	count²
Regressions 😿 (primary)	N/A	N/A	0
Regressions 😿 (secondary)	3.7%	5.9%	2
Improvements 🎉 (primary)	N/A	N/A	0
Improvements 🎉 (secondary)	-3.6%	-3.6%	1
All 😿🎉 (primary)	N/A	N/A	0

Cycles

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

If you disagree with this performance assessment, please file an issue in rust-lang/rustc-perf.

Next Steps: If you can justify the regressions found in this perf run, please indicate this with @rustbot label: +perf-regression-triaged along with sufficient written justification. If you cannot justify the regressions please open an issue or create a new PR that fixes the regressions, add a comment linking to the newly created issue or PR, and then add the perf-regression-triaged label to this PR.

@rustbot label: +perf-regression

the arithmetic mean of the percent change↩↩²
number of relevant changes↩↩²

nnethercote · 2022-06-20T22:44:39Z

Sigh, performance regression here. keccak stresses the process obligation code, and it's very easy for minor, seemingly inconsequential changes to change it a couple of percent, because e.g. register allocation in the hot paths change a tiny bit. E.g. here are the results for the past month:

Let's look at the check full results, particularly the instruction counts.

This PR was looking like a win when measured 15 days ago (2022-06-06):

keccak	check	full	-1.96%	10.03x	20,749,612,998	20,343,711,976

Since then, three intervening PRs landed that improved keccak.

After #97447 (2022-06-08):

keccak	check	full	-2.06%	10.83x	20,718,223,737	20,291,756,873

After #97778 (2022-06-11):

keccak	check	full	-0.79%	6.05x	20,284,132,956	20,123,775,316

After #96591 (2022-06-14):

keccak	check	full	-1.61%	13.79x	20,111,191,751	19,787,718,379

So that was nice.

And now, after this PR merged yesterday (2022-06-20):

keccak	check	full	2.38%	24.28x	19,797,304,578	20,268,024,059

(All of the above is also true for cranelift-codegen-0.82.1, to a lesser extent.)

I'm inclined to accept this and move on, because:

this PR made the code nicer, and
keccak's instruction count is still lower than it was three weeks ago (3 steps forward, 1 step back, and I was responsible for two of those steps), and
the results here are so easy to perturb, and
this PR didn't affect cycle counts.

I may try one more time to recover the loss, but this code has already been optimized to the nth degree, so I probably won't spend long on it.

@rustbot label: +perf-regression-triaged

rustbot added the T-compiler Relevant to the compiler team, which will review and decide on the PR/issue. label Jun 3, 2022

rust-highfive assigned nikomatsakis Jun 3, 2022

rust-highfive added the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label Jun 3, 2022

rustbot added the S-waiting-on-perf Status: Waiting on a perf run to be completed. label Jun 3, 2022

nnethercote force-pushed the oblig-forest-tweaks branch from f673ead to bb5005fCompare June 3, 2022 08:40

rustbot added perf-regression Performance regression. and removed S-waiting-on-perf Status: Waiting on a perf run to be completed. labels Jun 3, 2022

nnethercote added 3 commits June 6, 2022 08:47

Streamline active_cache to done_cache transfer.
cdb446f

Handle stalling within ObligationForest.
281229a
It is simpler if `ObligationForest` does this itself, rather than the caller having to manage it.

Split process_obligation in two.
32741d5
Because it really has two halves: - A read-only part that checks if further work is needed. - The further work part, which is much less hot. This makes things a bit clearer and nicer.

nnethercote force-pushed the oblig-forest-tweaks branch from bb5005f to 32741d5Compare June 6, 2022 05:10

rustbot added the S-waiting-on-perf Status: Waiting on a perf run to be completed. label Jun 6, 2022

rustbot removed S-waiting-on-perf Status: Waiting on a perf run to be completed. perf-regression Performance regression. labels Jun 6, 2022

nikomatsakis approved these changes Jun 20, 2022
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 Jun 20, 2022

bors added the merged-by-bors This PR was explicitly merged by bors. label Jun 20, 2022

bors merged commit 1d60108 into rust-lang:masterJun 20, 2022

rustbot added this to the 1.63.0 milestone Jun 20, 2022

bors mentioned this pull request Jun 20, 2022
wf-check generators #97183
Merged

rustbot added the perf-regression Performance regression. label Jun 20, 2022

nnethercote deleted the oblig-forest-tweaks branch June 20, 2022 21:45

rustbot added the perf-regression-triaged The performance regression has been triaged. label Jun 20, 2022

Uh oh!

Uh oh!

Conversation

nnethercote commented Jun 3, 2022

Uh oh!

nnethercote commented Jun 3, 2022

Uh oh!

rust-timer commented Jun 3, 2022

Uh oh!

bors commented Jun 3, 2022

Uh oh!

nnethercote commented Jun 3, 2022

Uh oh!

rust-timer commented Jun 3, 2022

Uh oh!

bors commented Jun 3, 2022

Uh oh!

bors commented Jun 3, 2022

Uh oh!

rust-timer commented Jun 3, 2022

Uh oh!

rust-timer commented Jun 3, 2022

Footnotes

Uh oh!

nnethercote commented Jun 6, 2022

Uh oh!

nnethercote commented Jun 6, 2022

Uh oh!

rust-timer commented Jun 6, 2022

Uh oh!

bors commented Jun 6, 2022

Uh oh!

bors commented Jun 6, 2022

Uh oh!

rust-timer commented Jun 6, 2022

Uh oh!

rust-timer commented Jun 6, 2022

Footnotes

Uh oh!

nnethercote commented Jun 6, 2022

Uh oh!

nikomatsakis commented Jun 20, 2022

Uh oh!

bors commented Jun 20, 2022

Uh oh!

bors commented Jun 20, 2022

Uh oh!

bors commented Jun 20, 2022

Uh oh!

rust-timer commented Jun 20, 2022

Footnotes

Uh oh!

nnethercote commented Jun 20, 2022

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants