[benchmark] Janitor Duty: Sweep I

[palimondo]

A little bit of benchmark cleanup.

[palimondo]

@swift-ci please benchmark

[palimondo]

@swift-ci please smoke test

[palimondo]

@eeckstein Please review 🙏

[eeckstein]

basically lgtm.
Does it make sense to rename the ArrayLiteral benchmark?

[palimondo]

@swift-ci please benchmark

[palimondo]

@swift-ci please smoke test

[swift-ci]

Performance: -O

TEST	MIN	MAX	MEAN	MAX_RSS
Added
ArrayLiteral2	184	187	185	—
ArraySetElement	437	437	437	—
Removed
ArrayLiteral	0	0	0	—

Code size: -O

TEST	OLD	NEW	DELTA	RATIO
Regression
ArrayLiteral.o	3143	3311	+5.3%	0.95x
ArraySetElement.o	1119	1135	+1.4%	0.99x

Performance: -Osize

TEST	OLD	NEW	DELTA	RATIO
Improvement
DataCreateSmall	3000	2710	-9.7%	1.11x (?)
Added
ArrayLiteral2	197	203	200	—
ArraySetElement	436	436	436	—
Removed
ArrayLiteral	197	200	198	—

Code size: -Osize

TEST	OLD	NEW	DELTA	RATIO
Regression
ArrayLiteral.o	2984	3109	+4.2%	0.96x
ArraySetElement.o	1309	1325	+1.2%	0.99x

Performance: -Onone

TEST	OLD	NEW	DELTA	RATIO
Improvement
DataAppendArray	5000	4100	-18.0%	1.22x (?)
Added
ArrayLiteral2	1484	1593	1520	—
ArraySetElement	1213	1256	1231	—
Removed
ArrayLiteral	1510	1513	1512	—

✅	Benchmark Check Report

How to read the data

The tables contain differences in performance which are larger than 8% and differences in code size which are larger than 1%.

If you see any unexpected regressions, you should consider fixing the
regressions before you merge the PR.

Noise: Sometimes the performance results (not code size!) contain false
alarms. Unexpected regressions which are marked with '(?)' are probably noise.
If you see regressions which you cannot explain you can try to run the
benchmarks again. If regressions still show up, please consult with the
performance team (@eeckstein).

Hardware Overview

  Model Name: Mac Pro
  Model Identifier: MacPro6,1
  Processor Name: 8-Core Intel Xeon E5
  Processor Speed: 3 GHz
  Number of Processors: 1
  Total Number of Cores: 8
  L2 Cache (per Core): 256 KB
  L3 Cache: 25 MB
  Memory: 64 GB

[palimondo]

@eeckstein Hmm… from the above report, I would argue that we should not rename the ArrayLiteral benchmark, as the -Osize and -Onone performance was unchanged and only the -O is being fixed by the blackHole application to go from 0 to some regular number.

[eeckstein]

ok, LGTM

[palimondo]

I’m confused now. You approved this with the changed name… Do you want me to merge this as is (ArrayLiteral2) or with original name ArrayLiteral (the version before last force push) to maintain tracking continuity in -Osize and -Onone?

[eeckstein]

Please merge it as is (with ArrayLiteral2)