Pre-intern some const infer vars

[compiler-errors]

Cache const vars like we do for ty/region vars

[rustbot]

r? @nnethercote

rustbot has assigned @nnethercote.
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

[compiler-errors]

@bors try @rust-timer queue

[bors]

⌛ Trying commit 5771471 with merge 5221322...

[compiler-errors]

@nnethercote: You authored #107869, but presumably didn't add a fast path for const infer vars at that point b/c we still had to store the type of the const in the ty::Const (thanks @BoxyUwU for #125958 🎉), so it wasn't possible to implement at that point.

Let's see how this fares.

Some questions:

• Do you have any intuition about what values you chose for NUM_PREINTERNED_*?
• Somewhat unrelated, but was it just not worthwhile to pre-intern int and float vars (the non-fresh ones) when you were testing out that PR?

I also didn't pre-intern fresh const vars, but those may benefit from this too. Maybe a good follow-up after this tho, though.

[bors]

☀️ Try build successful - checks-actions
Build commit: 5221322 (5221322a8e78f5547d7c0b5585ff992545789beb)

[rust-timer]

Finished benchmarking commit (5221322): 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)	3.0%	[3.0%, 3.0%]	1
Regressions ❌ (secondary)	0.1%	[0.1%, 0.1%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	3.0%	[3.0%, 3.0%]	1

Max RSS (memory usage)

Results (primary -4.1%, secondary -2.2%)

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)	3.7%	[3.7%, 3.7%]	1
Improvements ✅ (primary)	-4.1%	[-4.1%, -4.1%]	1
Improvements ✅ (secondary)	-3.3%	[-5.8%, -2.2%]	5
All ❌✅ (primary)	-4.1%	[-4.1%, -4.1%]	1

Cycles

Results (primary 3.3%, secondary -4.5%)

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)	3.3%	[3.3%, 3.3%]	1
Regressions ❌ (secondary)	0.9%	[0.9%, 0.9%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-6.3%	[-7.1%, -5.3%]	3
All ❌✅ (primary)	3.3%	[3.3%, 3.3%]	1

Binary size

Results (primary 1.1%)

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)	1.1%	[1.1%, 1.1%]	1
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	1.1%	[1.1%, 1.1%]	1

Bootstrap: 776.203s -> 778.493s (0.30%)
Artifact size: 366.33 MiB -> 366.31 MiB (-0.01%)

[nnethercote]

Some questions:

* Do you have any intuition about what values you chose for `NUM_PREINTERNED_*`?

It would have been based on measured data, either trying to hit some fraction of actual occurrences without pre-interning too many, or maybe the actual icount results.

* Somewhat unrelated, but was it just not worthwhile to pre-intern int and float vars (the non-_fresh_ ones) when you were testing out that PR?

I don't remember specifically, but probably the profile data didn't point at that being useful. You'd have to try it to be certain.

I also didn't pre-intern fresh const vars, but those may benefit from this too. Maybe a good follow-up after this tho, though.

There is a very small benefit here for ctfe-stress, small enough that you have to click "show non-relevant results" to see it. I'm guessing you didn't try this because it showed up in a profile, but because you saw some types and regions were pre-interned and figured it was worth trying with consts too?

The perf results suggest this is borderline whether it's worth merging, what do you think?