Optimize query_cache_hit to reduce code size of the query hot path. #107529

Zoxc · 2023-01-31T18:35:10Z

A small tweak which improves performance on check builds by 0.33% and reduces rustc_driver size by 1%.

Benchmark	Before	Before		After
Benchmark	Time	Time	%	Time	%
🟣 clap:check	1.7978s	1.7980s	0.01%	1.7930s	-0.27%
🟣 hyper:check	0.2594s	0.2591s	-0.12%	0.2592s	-0.09%
🟣 syntex_syntax:check	6.2522s	6.2540s	0.03%	6.2358s	-0.26%
🟣 syn:check	1.5889s	1.5880s	-0.05%	1.5799s	-0.57%
🟣 regex:check	0.9941s	0.9939s	-0.02%	0.9893s	-0.49%
Total	10.8925s	10.8930s	0.01%	10.8572s	-0.32%
Summary	1.0000s	0.9997s	-0.03%	0.9967s	-0.33%

r? @cjgillot

compiler-errors · 2023-01-31T18:40:16Z

@bors try @rust-timer queue

bors · 2023-01-31T18:40:26Z

⌛ Trying commit 0f85685e521956584bd60923e614d68353f6fe38 with merge f8c1f62c53f582f5eacfa08617ed8e6fb1385ffd...

Noratrieb · 2023-01-31T20:01:25Z

compiler/rustc_data_structures/src/profiling.rs

@@ -393,7 +393,7 @@ impl SelfProfilerRef {
    }

    /// Record a query in-memory cache hit.
-    #[inline(always)]
+    #[inline(never)]
    pub fn query_cache_hit(&self, query_invocation_id: QueryInvocationId) {


The same should probably be done for the other profiling events as well

The problem was that it generated code for TimingGuard which was unused. The other events do make use of it so it outlining doesn't help.

bors · 2023-01-31T20:53:52Z

☀️ Try build successful - checks-actions
Build commit: f8c1f62c53f582f5eacfa08617ed8e6fb1385ffd (f8c1f62c53f582f5eacfa08617ed8e6fb1385ffd)

rust-timer · 2023-01-31T22:12:06Z

Finished benchmarking commit (f8c1f62c53f582f5eacfa08617ed8e6fb1385ffd): 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
Regressions ❌ (secondary)	0.6%	[0.2%, 0.9%]	6
Improvements ✅ (primary)	-0.3%	[-0.3%, -0.3%]	2
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)	-	-	0
Regressions ❌ (secondary)	2.0%	[2.0%, 2.0%]	1
Improvements ✅ (primary)	-3.7%	[-3.7%, -3.7%]	1
Improvements ✅ (secondary)	-2.2%	[-4.7%, -1.0%]	7
All ❌✅ (primary)	-3.7%	[-3.7%, -3.7%]	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)	1.7%	[1.3%, 2.0%]	3
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	1.7%	[1.3%, 2.0%]	3

Zoxc · 2023-02-01T04:03:07Z

It makes sense that perf regresses since it runs with the profiler enabled. I've optimized query_cache_hit instead to avoid the extra function call and branch in the query system when profiling.

bors · 2023-02-06T12:43:59Z

☔ The latest upstream changes (presumably #107667) made this pull request unmergeable. Please resolve the merge conflicts.

Zoxc · 2023-02-06T13:36:21Z

It looks like this is a bigger win after #107667:

Benchmark	Before	After
Benchmark	Time	Time	%
🟣 clap:check	1.8384s	1.8200s	💚 -1.00%
🟣 hyper:check	0.2624s	0.2604s	-0.74%
🟣 regex:check	1.0245s	1.0113s	💚 -1.28%
🟣 syn:check	1.6461s	1.6298s	-0.99%
🟣 syntex_syntax:check	6.3677s	6.3036s	💚 -1.01%
Total	11.1390s	11.0252s	💚 -1.02%
Summary	1.0000s	0.9900s	💚 -1.00%

cjgillot · 2023-02-07T17:46:27Z

@bors r+

bors · 2023-02-07T17:46:29Z

📌 Commit 9539737 has been approved by cjgillot

It is now in the queue for this repository.

bors · 2023-02-08T10:35:50Z

⌛ Testing commit 9539737 with merge a00e24d...

bors · 2023-02-08T13:28:35Z

☀️ Test successful - checks-actions
Approved by: cjgillot
Pushing a00e24d to master...

bors · 2023-02-08T13:28:35Z

☀️ Test successful - checks-actions
Approved by: cjgillot
Pushing a00e24d to master...

rust-timer · 2023-02-08T14:46:49Z

Finished benchmarking commit (a00e24d): comparison URL.

Overall result: ❌✅ regressions and improvements - ACTION NEEDED

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
cc @rust-lang/wg-compiler-performance

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%	[0.5%, 0.5%]	1
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-0.3%	[-0.4%, -0.2%]	4
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-0.1%	[-0.4%, 0.5%]	5

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)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-3.2%	[-3.2%, -3.2%]	1
All ❌✅ (primary)	-	-	0

Cycles

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

rylev · 2023-02-14T16:19:40Z

Calling this triaged as the regression is small

@rustbot label: perf-regression-triaged

rustbot assigned cjgillot Jan 31, 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 Jan 31, 2023

Noratrieb reviewed Jan 31, 2023

View reviewed changes

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 Jan 31, 2023

Zoxc changed the title ~~Don't inline query_cache_hit to reduce code size of the query hot path.~~ Optimize query_cache_hit to reduce code size of the query hot path. Feb 1, 2023

Don't inline query_cache_hit to reduce code size of the query hot path.

e60ccfc

Zoxc force-pushed the inline-tweak-profile branch from f9aa140 to 4117898 Compare February 6, 2023 13:32

Make an optimal cold path for query_cache_hit

9539737

Zoxc force-pushed the inline-tweak-profile branch from 4117898 to 9539737 Compare February 6, 2023 14:22

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 Feb 7, 2023

bors added merged-by-bors This PR was explicitly merged by bors. labels Feb 8, 2023

bors merged commit a00e24d into rust-lang:master Feb 8, 2023

rustbot added this to the 1.69.0 milestone Feb 8, 2023

Zoxc deleted the inline-tweak-profile branch February 8, 2023 13:48

rustbot added the perf-regression-triaged The performance regression has been triaged. label Feb 14, 2023

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Optimize query_cache_hit to reduce code size of the query hot path. #107529

Optimize query_cache_hit to reduce code size of the query hot path. #107529

Zoxc commented Jan 31, 2023 •

edited

Loading

compiler-errors commented Jan 31, 2023

This comment has been minimized.

bors commented Jan 31, 2023

Noratrieb Jan 31, 2023

Zoxc Feb 1, 2023

bors commented Jan 31, 2023

This comment has been minimized.

rust-timer commented Jan 31, 2023

Zoxc commented Feb 1, 2023

bors commented Feb 6, 2023

Zoxc commented Feb 6, 2023

cjgillot commented Feb 7, 2023

bors commented Feb 7, 2023

bors commented Feb 8, 2023

bors commented Feb 8, 2023

bors commented Feb 8, 2023

rust-timer commented Feb 8, 2023

rylev commented Feb 14, 2023

Optimize query_cache_hit to reduce code size of the query hot path. #107529

Optimize query_cache_hit to reduce code size of the query hot path. #107529

Conversation

Zoxc commented Jan 31, 2023 • edited Loading

compiler-errors commented Jan 31, 2023

This comment has been minimized.

bors commented Jan 31, 2023

Choose a reason for hiding this comment

Choose a reason for hiding this comment

bors commented Jan 31, 2023

This comment has been minimized.

rust-timer commented Jan 31, 2023

Overall result: ❌✅ regressions and improvements - ACTION NEEDED

Zoxc commented Feb 1, 2023

bors commented Feb 6, 2023

Zoxc commented Feb 6, 2023

cjgillot commented Feb 7, 2023

bors commented Feb 7, 2023

bors commented Feb 8, 2023

bors commented Feb 8, 2023

bors commented Feb 8, 2023

rust-timer commented Feb 8, 2023

Overall result: ❌✅ regressions and improvements - ACTION NEEDED

rylev commented Feb 14, 2023

Zoxc commented Jan 31, 2023 •

edited

Loading