rustc_span: Optimize span parent get/set methods

[petrochenkov]

Like #125017, but for span parents.

r? @cjgillot

[petrochenkov]

@bors try @rust-timer queue

[bors]

⌛ Trying commit 90b72b5 with merge eab5a5b...

[bors]

☀️ Try build successful - checks-actions
Build commit: eab5a5b (eab5a5bd74528e95727f5e553f1d0fe06bcac85e)

[rust-timer]

Finished benchmarking commit (eab5a5b): 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)	0.3%	[0.3%, 0.3%]	2
Regressions ❌ (secondary)	0.3%	[0.2%, 0.4%]	8
Improvements ✅ (primary)	-0.3%	[-0.4%, -0.2%]	6
Improvements ✅ (secondary)	-2.0%	[-2.0%, -2.0%]	1
All ❌✅ (primary)	-0.1%	[-0.4%, 0.3%]	8

Max RSS (memory usage)

Results (primary -2.6%)

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)	-2.6%	[-2.6%, -2.6%]	1
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-2.6%	[-2.6%, -2.6%]	1

Cycles

Results (primary 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)	2.0%	[2.0%, 2.0%]	1
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	2.0%	[2.0%, 2.0%]	1

Binary size

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

Bootstrap: 672.776s -> 671.266s (-0.22%)
Artifact size: 319.83 MiB -> 320.50 MiB (0.21%)

[petrochenkov]

Statistics (similar to #125017 (comment)).

// from standard library
// apparently incremental is not enabled for stdlib despite `rust.incremental = true` in the config
91546 inline_ctxt: None -> None
   41 partially_interned: None -> None

// from ui test suite
// incremental is typically disabled
364420 inline_ctxt: None -> None
  3957 inline_ctxt: None -> Some
   188 inline_parent: Some -> Some
    93 partially_interned: Some -> Some
    13 partially_interned: None -> None
    11 inline_parent: Some -> None
     2 partially_interned: Some -> None

// from incremental test suite
107656 inline_ctxt: None -> Some
  8983 inline_ctxt: None -> None
  3642 inline_parent: Some -> Some
  2789 partially_interned: Some -> Some
    90 partially_interned: Some -> None
    42 inline_parent: Some -> None
     5 partially_interned: None -> Some

[petrochenkov]

The main conclusion from the statistics here and in #125017 (comment) is that we don't get much perf improvements from avoiding the interner (partially_interned cases are rare).

We instead get the improvements from avoiding the span.data() -> Span::new() arithmetic in the most common inline_ctxt cases, which don't involve accessing the interner at all.

So it's likely that we can optimize just inline_ctxt: zero|small -> zero|small for ctxt updates, and inline_ctxt: None -> None|Some for parent updates, and everything else can just go through the fallback logic to avoid blowing up code size.
Upd: benchmarked in #126591.
Upd2: the benchmarks in #126591 confirm the hypothesis.

[petrochenkov]

Updated in accordance with #126544 (comment).
I also added some comment updates for map_ctxt missing from the accidentally merged #126591.
The questions about dependency tracking are still relevant though.

[cjgillot]

r=me with the points about tracking resolved.

[petrochenkov]

@bors r=cjgillot

[bors]

📌 Commit 4d3b617 has been approved by cjgillot

It is now in the queue for this repository.

[bors]

⌛ Testing commit 4d3b617 with merge a9c8887...

[bors]

☀️ Test successful - checks-actions
Approved by: cjgillot
Pushing a9c8887 to master...

[rust-timer]

Finished benchmarking commit (a9c8887): 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
Regressions ❌ (secondary)	0.7%	[0.6%, 0.8%]	4
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

Max RSS (memory usage)

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

Cycles

Results (secondary 0.3%)

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)	2.8%	[2.8%, 2.8%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-2.2%	[-2.2%, -2.2%]	1
All ❌✅ (primary)	-	-	0

Binary size

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

Bootstrap: 700.231s -> 701.948s (0.25%)
Artifact size: 324.89 MiB -> 325.02 MiB (0.04%)