(EXPERIMENT) Make ChunkedBitSet's bitwise_changes more unroller/autovectorizer-friendly

[Zalathar]

This idea came to me after thinking about #147619, though I don't yet know whether it's worthwhile.

The basic idea is that loops with early exits tend to be slower than necessary, because the compiler doesn't know that it would be fine to batch up a few extra iterations and early-exit a bit later.

r? ghost

[Zalathar]

@bors try @rust-timer queue

[rust-bors]

☀️ Try build successful (CI)
Build commit: 2cef582 (2cef582e05346b4c1717169044d39d23b8c8fa7f, parent: 53a741fc4b8cf2d8e7b1b2336ed8edf889db84f4)

[rust-timer]

Finished benchmarking commit (2cef582): comparison URL.

Overall result: ✅ improvements - no action needed

Benchmarking this pull request means it may be perf-sensitive – we'll automatically label it not fit for rolling up. You can override this, but we strongly advise not to, due to possible changes in compiler perf.

@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
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 (secondary -3.6%)

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)	-3.6%	[-4.4%, -2.3%]	3
All ❌✅ (primary)	-	-	0

Cycles

Results (primary -2.3%, secondary 4.1%)

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%	[2.3%, 5.9%]	2
Improvements ✅ (primary)	-2.3%	[-2.3%, -2.3%]	1
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-2.3%	[-2.3%, -2.3%]	1

Binary size

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

Bootstrap: 475.105s -> 474.148s (-0.20%)
Artifact size: 390.35 MiB -> 390.47 MiB (0.03%)

[Zalathar]

A measurable improvement, but not very dramatic, perhaps because the equality check has already claimed most of the low-hanging fruit.

Since I don't have benchmarking set up locally, iterating on this would be a hassle, so I'll just close.