Stop making const None in GVN

[scottmcm]

Having ALLOCs for these seems unhelpful, since it's extra stuff to keep track of, makes it harder for other passes to know what's in it (reversing a tag constant to a VariantIdx isn't trivial), and even in codegen it's arguably harder to get a None from a valtree than from seeing "oh, aggregate nothing with this VariantIdx".

Making constants for the discriminant(_n) is useful for putting in switchInts, and still happens with this PR. It's just for the full enum that we don't make consts (except for might-as-well-be-an-integer enums like Ordering, or ones that are actually ZSTs like const Option::<Infallible>::None).

This also adds a couple tests I've been using as I look at how we're doing enums in MIR. They're not really improved by this change materially, but they help motivate why I'm trying to tweak how the enums come out.

[rustbot]

r? @fmease

rustbot has assigned @fmease.
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

[rustbot]

Some changes occurred to MIR optimizations

cc @rust-lang/wg-mir-opt

[fmease]

r? compiler

[scottmcm]

@bors try @rust-timer queue

[bors]

⌛ Trying commit 41f6339 with merge e18ec78...

[bors]

☀️ Try build successful - checks-actions
Build commit: e18ec78 (e18ec7895190171d8837347d5500d58a4166c464)

[rust-timer]

Finished benchmarking commit (e18ec78): comparison URL.

Overall result: ❌✅ regressions and improvements - 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%, 0.7%]	9
Regressions ❌ (secondary)	0.3%	[0.2%, 0.7%]	16
Improvements ✅ (primary)	-0.3%	[-0.4%, -0.3%]	2
Improvements ✅ (secondary)	-0.3%	[-0.4%, -0.1%]	4
All ❌✅ (primary)	0.2%	[-0.4%, 0.7%]	11

Max RSS (memory usage)

Results (primary -0.2%, secondary 3.0%)

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)	2.7%	[0.9%, 5.2%]	4
Regressions ❌ (secondary)	5.1%	[2.1%, 7.5%]	11
Improvements ✅ (primary)	-4.0%	[-6.9%, -2.0%]	3
Improvements ✅ (secondary)	-4.5%	[-6.4%, -3.2%]	3
All ❌✅ (primary)	-0.2%	[-6.9%, 5.2%]	7

Cycles

Results (primary 5.5%, secondary 2.8%)

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)	5.5%	[1.3%, 12.7%]	17
Regressions ❌ (secondary)	2.8%	[1.5%, 4.8%]	7
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	5.5%	[1.3%, 12.7%]	17

Binary size

Results (primary -0.0%, secondary -0.1%)

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)	0.2%	[0.0%, 0.8%]	28
Regressions ❌ (secondary)	0.1%	[0.0%, 0.2%]	7
Improvements ✅ (primary)	-0.1%	[-0.9%, -0.0%]	51
Improvements ✅ (secondary)	-0.1%	[-0.4%, -0.0%]	21
All ❌✅ (primary)	-0.0%	[-0.9%, 0.8%]	79

Bootstrap: 774.866s -> 772.816s (-0.26%)
Artifact size: 365.06 MiB -> 365.10 MiB (0.01%)

[jieyouxu]

r? mir-opt

[scottmcm]

nvm, looks like this is bad on its own and will need other things before it can be considered.