Tell rustc about unused bits in Span.

[nnethercote]

This shrinks the size of Option<Span> and similar types, which shrinks several AST and HIR nodes. The most important of these are hir::Expr and ast::Ty.

This change was first attempted in #93747 where it had little effect. But the improved niche-filling implemented by #94075 means this change now has a much bigger effect.

r? @cjgillot

[nnethercote]

@bors try @rust-timer queue

[rust-timer]

Awaiting bors try build completion.

@rustbot label: +S-waiting-on-perf

[bors]

⌛ Trying commit 63596860eb24e80b8e7d024d11f983296a262f2c with merge 9b1b504285e7b4440ae811dc71fbffdc7be2ec92...

[bors]

☀️ Try build successful - checks-actions
Build commit: 9b1b504285e7b4440ae811dc71fbffdc7be2ec92 (9b1b504285e7b4440ae811dc71fbffdc7be2ec92)

[rust-timer]

Queued 9b1b504285e7b4440ae811dc71fbffdc7be2ec92 with parent 2019147, future comparison URL.

[rust-timer]

Finished benchmarking commit (9b1b504285e7b4440ae811dc71fbffdc7be2ec92): comparison URL.

Overall result: ❌ regressions - 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-review -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.

	mean1	range	count2
Regressions ❌ (primary)	0.4%	[0.2%, 1.2%]	103
Regressions ❌ (secondary)	0.6%	[0.2%, 1.1%]	66
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-0.6%	[-0.6%, -0.6%]	1
All ❌✅ (primary)	0.4%	[0.2%, 1.2%]	103

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.

	mean1	range	count2
Regressions ❌ (primary)	1.4%	[0.7%, 2.7%]	25
Regressions ❌ (secondary)	4.2%	[4.2%, 4.2%]	1
Improvements ✅ (primary)	-1.0%	[-1.4%, -0.6%]	8
Improvements ✅ (secondary)	-2.4%	[-5.0%, -1.0%]	20
All ❌✅ (primary)	0.9%	[-1.4%, 2.7%]	33

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.

	mean1	range	count2
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-4.0%	[-4.0%, -4.0%]	1
All ❌✅ (primary)	-	-	0

Footnotes

1. the arithmetic mean of the percent change ↩ ↩² ↩³

2. number of relevant changes ↩ ↩² ↩³

[nnethercote]

Hmm.

• instruction counts: lots of 1% or smaller regressions.
• max-rss: some decent wins for secondary benchmarks that are HIR-heavy, but also some regressions on real-world crates, WTF.
• cycles and wall-time look neutral, though they are noisy enough that it's hard to tell for sure.

[cjgillot]

I marking this PR as blocked on improvements to niche optimization tracked by #101872

[nnethercote]

Let's try a perf run again, just to see what happens.

@bors try @rust-timer queue

[rust-timer]

Awaiting bors try build completion.

@rustbot label: +S-waiting-on-perf

[bors]

⌛ Trying commit 273fb2a with merge 5445699cee92d57d2c75196ec6b23d56f65dbdca...

[bors]

☀️ Try build successful - checks-actions
Build commit: 5445699cee92d57d2c75196ec6b23d56f65dbdca (5445699cee92d57d2c75196ec6b23d56f65dbdca)

[rust-timer]

Queued 5445699cee92d57d2c75196ec6b23d56f65dbdca with parent 9c2797d, future comparison URL.

[rust-timer]

Finished benchmarking commit (5445699cee92d57d2c75196ec6b23d56f65dbdca): comparison URL.

Overall result: ❌ regressions - 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-review -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.

	mean1	range	count2
Regressions ❌ (primary)	0.5%	[0.2%, 1.2%]	147
Regressions ❌ (secondary)	0.7%	[0.2%, 1.9%]	83
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.5%	[0.2%, 1.2%]	147

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.

	mean1	range	count2
Regressions ❌ (primary)	1.6%	[0.7%, 3.0%]	35
Regressions ❌ (secondary)	2.3%	[1.6%, 3.5%]	7
Improvements ✅ (primary)	-0.8%	[-1.5%, -0.4%]	7
Improvements ✅ (secondary)	-2.8%	[-5.1%, -1.1%]	6
All ❌✅ (primary)	1.2%	[-1.5%, 3.0%]	42

Cycles

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

Footnotes

1. the arithmetic mean of the percent change ↩ ↩²

2. number of relevant changes ↩ ↩²

[apiraino]

visiting and checking progress. according to this comment this is waiting on #101872 which in turn is waiting on an LLVM optimization.

Correct @cjgillot @nnethercote ?

[nnethercote]

That's correct.

[nnethercote]

@cjgillot: I just tried to update this, but things have changed in Span representation, because of parent fields. Is this change still applicable in a modified form?

[cjgillot]

I don't think so. I used those very unused bits to encode the presence of the parent.
I think the current encoding uses all available bits.

[nnethercote]

This change no longer works, due to other changes that have happened in the meantime.