[perf experiment] Only emit gep not gepi

[saethlin]

r? @ghost

Re-creating #120483, trying to measure the perf impact of gep vs gepi. That previous PR was with LLVM 17 and it found a runtime perf improvement, so in part this PR is assessing if that was fixed in LLVM 18.

Codegen tests will fail.

[saethlin]

@bors try @rust-timer queue

[bors]

⌛ Trying commit dcfa484 with merge e8fece1...

[bors]

☀️ Try build successful - checks-actions
Build commit: e8fece1 (e8fece15519f67c5712e05195a58d2c7e0116089)

[rust-timer]

Finished benchmarking commit (e8fece1): 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-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.

	mean	range	count
Regressions ❌ (primary)	3.0%	[0.6%, 6.8%]	237
Regressions ❌ (secondary)	3.0%	[0.6%, 8.9%]	226
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	3.0%	[0.6%, 6.8%]	237

Max RSS (memory usage)

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

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.

	mean	range	count
Regressions ❌ (primary)	1.8%	[0.9%, 3.1%]	83
Regressions ❌ (secondary)	5.5%	[1.4%, 15.4%]	24
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	1.8%	[0.9%, 3.1%]	83

Binary size

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.

	mean	range	count
Regressions ❌ (primary)	0.3%	[0.1%, 1.0%]	20
Regressions ❌ (secondary)	0.2%	[0.0%, 1.3%]	37
Improvements ✅ (primary)	-0.2%	[-0.2%, -0.2%]	4
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.2%	[-0.2%, 1.0%]	24

Bootstrap: 631.732s -> 640.84s (1.44%)
Artifact size: 305.05 MiB -> 305.32 MiB (0.09%)

[saethlin]

@bors try @rust-timer queue

[bors]

⌛ Trying commit 1743c5b with merge 926c405...

[bors]

☀️ Try build successful - checks-actions
Build commit: 926c405 (926c405ce7a22495a945112b3f5d38b99c4af7f4)

[rust-timer]

Finished benchmarking commit (926c405): 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-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.

	mean	range	count
Regressions ❌ (primary)	2.7%	[0.6%, 5.3%]	250
Regressions ❌ (secondary)	2.8%	[0.5%, 10.4%]	236
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	2.7%	[0.6%, 5.3%]	250

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.

	mean	range	count
Regressions ❌ (primary)	4.0%	[4.0%, 4.0%]	1
Regressions ❌ (secondary)	5.0%	[5.0%, 5.0%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	4.0%	[4.0%, 4.0%]	1

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.

	mean	range	count
Regressions ❌ (primary)	1.4%	[0.7%, 2.8%]	75
Regressions ❌ (secondary)	3.5%	[1.0%, 6.5%]	11
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	1.4%	[0.7%, 2.8%]	75

Binary size

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.

	mean	range	count
Regressions ❌ (primary)	0.4%	[0.1%, 1.0%]	23
Regressions ❌ (secondary)	1.0%	[0.0%, 1.4%]	39
Improvements ✅ (primary)	-0.3%	[-0.3%, -0.2%]	5
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.3%	[-0.3%, 1.0%]	28

Bootstrap: 673.458s -> 684.253s (1.60%)
Artifact size: 315.85 MiB -> 318.64 MiB (0.88%)

[saethlin]

Wonderful, it looks like the LLVM bug has been fixed.