Move fast paths around to gain a bit of new-solver perf

[jdonszelmann]

Change in % compared to previous commit on 4 benchmarks. All measured locally on ragdoll (AMD RYZEN 3990X, x86_64). These benchmarks have been quite repeatable in the past, but don't exactly match perf runs. All benchmarks show change in instruction count
First change is compared to baseline
Commit ids may change on rebases, so these links may become outdated.

	nalgebra [check,debug,opt]	serde [check,debug, opt]	syn [check, debug, opt]	wg-grammar [check, debug, opt]
move overflow check	$${\color{green}-0.04\%}$$ $${\color{green}-0.03\%}$$ $${\color{green}-0.02\%}$$	$${\color{green}-0.15\%}$$ $${\color{green}-0.12\%}$$ $${\color{green}-0.11\%}$$	$${\color{green}-0.08\%}$$ $${\color{green}-0.04\%}$$ $${\color{green}-0.01\%}$$	$${\color{green}-0.32\%}$$ $${\color{green}-0.31\%}$$ $${\color{green}-0.29\%}$$
move fast path into evaluate_goal	$${\color{red}+0.61\%}$$ $${\color{red}+0.60\%}$$ $${\color{red}+0.43\%}$$	$${\color{red}+0.75\%}$$ $${\color{red}+0.69\%}$$ $${\color{red}+0.65\%}$$	$${\color{red}+0.55\%}$$ $${\color{red}+0.36\%}$$ $${\color{red}+0.14\%}$$	$${\color{red}+0.13\%}$$ $${\color{red}+0.21\%}$$ $${\color{red}+0.15\%}$$
remove fastpath from fullfilment loop	$${\color{red}+0.56\%}$$ $${\color{red}+0.50\%}$$ $${\color{red}+0.35\%}$$	$${\color{red}+3.33\%}$$ $${\color{red}+2.76\%}$$ $${\color{red}+2.55\%}$$	$${\color{red}+1.65\%}$$ $${\color{red}+0.93\%}$$ $${\color{red}+0.33\%}$$	$${\color{green}-5.06\%}$$ $${\color{green}-5.00\%}$$ $${\color{green}-4.65\%}$$
return stalled info from fast path	$${\color{red}+0.16\%}$$ $${\color{red}+0.16\%}$$ $${\color{red}+0.12\%}$$	$${\color{red}+0.01\%}$$ $${\color{red}+0.01\%}$$ $${\color{red}+0.01\%}$$	$${\color{red}+0.24\%}$$ $${\color{red}+0.15\%}$$ $${\color{red}+0.05\%}$$	$${\color{green}-0.24\%}$$ $${\color{green}-0.24\%}$$ $${\color{green}-0.25\%}$$
move goal fast path into evaluate_goal_cold	$${\color{red}+0.01\%}$$ $${\color{red}+0.01\%}$$ $${\color{red}+0.01\%}$$	$${\color{red}+0.02\%}$$ $${\color{red}+0.02\%}$$ $${\color{red}+0.02\%}$$	$${\color{red}+0.02\%}$$ $${\color{red}+0.02\%}$$ $${\color{red}+0.01\%}$$	$${\color{green}-0.01\%}$$ $${\color{green}-0.01\%}$$ $${\color{green}-0.01\%}$$
shuffle fastpaths around to before creating an evalctxt	$${\color{green}-0.44\%}$$ $${\color{green}-0.41\%}$$ $${\color{green}-0.30\%}$$	$${\color{green}-2.31\%}$$ $${\color{green}-1.93\%}$$ $${\color{green}-1.80\%}$$	$${\color{green}-1.29\%}$$ $${\color{green}-0.77\%}$$ $${\color{green}-0.29\%}$$	$${\color{green}-6.06\%}$$ $${\color{green}-5.90\%}$$ $${\color{green}-5.48\%}$$
fast path when adding goals	$${\color{green}-0.62\%}$$ $${\color{green}-0.58\%}$$ $${\color{green}-0.41\%}$$	$${\color{green}-0.78\%}$$ $${\color{green}-0.66\%}$$ $${\color{green}-0.61\%}$$	$${\color{green}-0.36\%}$$ $${\color{green}-0.28\%}$$ $${\color{green}-0.10\%}$$	$${\color{green}-0.12\%}$$ $${\color{green}-0.09\%}$$ $${\color{green}-0.04\%}$$
overall from baseline	$${\color{red}+0.24\%}$$ $${\color{red}+0.23\%}$$ $${\color{red}+0.16\%}$$	$${\color{red}+0.79\%}$$ $${\color{red}+0.70\%}$$ $${\color{red}+0.66\%}$$	$${\color{red}+0.60\%}$$ $${\color{red}+0.36\%}$$ $${\color{red}+0.13\%}$$	$${\color{green}-11.31\%}$$ $${\color{green}-11.00\%}$$ $${\color{green}-10.27\%}$$

View all comments

[jdonszelmann]

This tests a bunch of changes together. I've done diffs locally suggesting ~5% speedup across these 7 commits. Moving fast path checks to evaluate_goal_raw ~2%, making it return stalled_on another 3% or so and using TyOrConstInferVar ~0.5%

@bors try @rust-timer queue

[rust-bors]

☀️ Try build successful (CI)
Build commit: 3f4f44d (3f4f44d99c410d3aadeb2731ddd6c5446c89eb72)
Base parent: 9030e34 (9030e345fe92df1ccefd0a8cdf61a9a9a5b73cb0)

[rust-timer]

Finished benchmarking commit (3f4f44d): comparison URL.

Overall result: ❌✅ regressions and improvements - please read:

Benchmarking means the PR may be perf-sensitive. It's automatically marked not fit for rolling up. Overriding is possible but disadvised: it risks changing compiler perf.

Next, please: If you can, justify the regressions found in this try perf run in writing along with @rustbot label: +perf-regression-triaged. If not, fix the regressions and do another perf run. Neutral or positive results will clear the label automatically.

@bors rollup=never
@rustbot label: -S-waiting-on-perf +perf-regression

Instruction count

Our most reliable metric. Used to determine the overall result above. However, even this metric can be noisy.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	1.1%	[0.1%, 3.7%]	18
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-2.9%	[-4.9%, -0.0%]	5
All ❌✅ (primary)	-	-	0

Max RSS (memory usage)

Results (primary 0.6%, secondary 1.6%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	2.7%	[2.7%, 2.7%]	1
Regressions ❌ (secondary)	2.2%	[1.7%, 2.7%]	9
Improvements ✅ (primary)	-1.4%	[-1.4%, -1.4%]	1
Improvements ✅ (secondary)	-1.2%	[-1.2%, -1.1%]	2
All ❌✅ (primary)	0.6%	[-1.4%, 2.7%]	2

Cycles

Results (secondary 3.2%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	4.1%	[2.1%, 6.2%]	6
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-2.3%	[-2.3%, -2.3%]	1
All ❌✅ (primary)	-	-	0

Binary size

Results (primary 0.0%, secondary 0.1%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	0.0%	[0.0%, 0.0%]	4
Regressions ❌ (secondary)	0.1%	[0.0%, 0.1%]	20
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.0%	[0.0%, 0.0%]	4

Bootstrap: 503.321s -> 505.38s (0.41%)
Artifact size: 353.40 MiB -> 353.04 MiB (-0.10%)

[jdonszelmann]

@bors try @rust-timer queue

[rust-bors]

☀️ Try build successful (CI)
Build commit: b2fad67 (b2fad672d15ed85bb352e25461a99b01bf10a0de)
Base parent: 973ad0d (973ad0d0ab149bde2e96422833c1265c7a5be217)

[rust-timer]

Finished benchmarking commit (b2fad67): comparison URL.

Overall result: ❌✅ regressions and improvements - please read:

Benchmarking means the PR may be perf-sensitive. It's automatically marked not fit for rolling up. Overriding is possible but disadvised: it risks changing compiler perf.

Next, please: If you can, justify the regressions found in this try perf run in writing along with @rustbot label: +perf-regression-triaged. If not, fix the regressions and do another perf run. Neutral or positive results will clear the label automatically.

@bors rollup=never
@rustbot label: -S-waiting-on-perf +perf-regression

Instruction count

Our most reliable metric. Used to determine the overall result above. However, even this metric can be noisy.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	0.6%	[0.2%, 1.2%]	12
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-6.9%	[-11.7%, -2.0%]	6
All ❌✅ (primary)	-	-	0

Max RSS (memory usage)

Results (primary 4.0%, secondary 3.3%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	4.0%	[3.4%, 4.7%]	2
Regressions ❌ (secondary)	4.8%	[0.9%, 22.7%]	7
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-1.6%	[-2.4%, -0.9%]	2
All ❌✅ (primary)	4.0%	[3.4%, 4.7%]	2

Cycles

Results (secondary -6.6%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-6.6%	[-8.2%, -2.1%]	4
All ❌✅ (primary)	-	-	0

Binary size

Results (secondary 0.0%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	0.0%	[0.0%, 0.0%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

Bootstrap: 509.16s -> 507.92s (-0.24%)
Artifact size: 353.81 MiB -> 353.75 MiB (-0.02%)

[rustbot]

Some changes occurred to the core trait solver

cc @rust-lang/initiative-trait-system-refactor

[rustbot]

r? @dingxiangfei2009

rustbot has assigned @dingxiangfei2009.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

Why was this reviewer chosen?

The reviewer was selected based on:

• Owners of files modified in this PR: compiler
• compiler expanded to 73 candidates
• Random selection from 18 candidates

[jdonszelmann]

r? @lcnr