Remove SSE2 path from rustc_span analyze_source_file

[real-eren]

Follow-up to #136460. When the SSE2 optimization was introduced, the generic path recorded NonNarrowChars for ASCII control characters. Nowadays, analyze_source_file only deals with newlines and multi-byte chars.

The point of this PR is to see, via a perf run, whether the SSE2 path still provides a meaningful improvement over the generic path. If it doesn't, it could be removed.

The function can be simplified further after inlining; I left it as-is for the initial perf run so that it's easier to see that the behavior is unchanged.

r? @Noratrieb

[workingjubilee]

@bors try @rust-timer queue

[bors]

⌛ Trying commit f48eb45 with merge b29600e...

[bors]

☀️ Try build successful - checks-actions
Build commit: b29600e (b29600eb90aece291a0bd84744383ca90737fff9)

[rust-timer]

Finished benchmarking commit (b29600e): comparison URL.

Overall result: ❌ regressions - please read the text below

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 the most reliable metric that we have; it was used to determine the overall result at the top of this comment. However, even this metric can sometimes exhibit noise.

	mean	range	count
Regressions ❌ (primary)	0.3%	[0.1%, 0.6%]	89
Regressions ❌ (secondary)	0.5%	[0.0%, 2.0%]	68
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.3%	[0.1%, 0.6%]	89

Max RSS (memory usage)

Results (primary 2.7%, secondary -3.5%)

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)	2.7%	[2.7%, 2.7%]	1
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-3.5%	[-3.5%, -3.5%]	1
All ❌✅ (primary)	2.7%	[2.7%, 2.7%]	1

Cycles

Results (primary 2.8%, 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)	2.8%	[2.8%, 2.8%]	1
Regressions ❌ (secondary)	2.4%	[2.4%, 2.4%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-2.9%	[-2.9%, -2.9%]	1
All ❌✅ (primary)	2.8%	[2.8%, 2.8%]	1

Binary size

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

Bootstrap: 789.592s -> 789.563s (-0.00%)
Artifact size: 350.02 MiB -> 350.03 MiB (0.00%)

[Noratrieb]

A tiny regression, but a regression nevertheless. Probably not worth removing then, as I'm not aware of any issues caused by this code.

[real-eren]

Thanks for the perf run.

When I ran the profile command for ripgrep locally, I only got a 0.2% regression, averaged over 3 runs. Still enough to justify keeping the relatively small SSE2 path, but not enough to warrant trying to improve it further (based on microbenchmarks for very similar routines, it can be ~50% faster).