[stdlib] use @_transparent in an appropriate spot

[glessard]

This function was inconsistently @inlinable, as opposed to its peers which are @_transparent. As a consequence, the detailed diagnostics surrounding Optional with == and != differ strangely based on minute details.

Resolves rdar://46444561

[glessard]

@swift-ci please test

[glessard]

@swift-ci please benchmark

[glessard]

Performance (x86_64): -O

Regression	OLD	NEW	DELTA	RATIO
LessSubstringSubstring	23	29	+26.1%	0.79x
EqualSubstringSubstringGenericEquatable	23	29	+26.1%	0.79x (?)
EqualSubstringString	23	29	+26.1%	0.79x
LessSubstringSubstringGenericComparable	23	29	+26.1%	0.79x (?)
EqualSubstringSubstring	24	30	+25.0%	0.80x (?)
EqualStringSubstring	24	30	+25.0%	0.80x
FlattenListFlatMap	3826	4417	+15.4%	0.87x (?)
DataCreateEmptyArray	1350	1550	+14.8%	0.87x (?)
StringComparison_longSharedPrefix	293	331	+13.0%	0.89x (?)
FindString.Loop1.Substring	272	301	+10.7%	0.90x (?)
SortStringsUnicode	1945	2105	+8.2%	0.92x (?)
Improvement	OLD	NEW	DELTA	RATIO
Dictionary4	217	153	-29.5%	1.42x
UTF8Decode_InitFromCustom_contiguous	159	125	-21.4%	1.27x (?)
UTF8Decode_InitDecoding	159	127	-20.1%	1.25x (?)
Dictionary4OfObjects	259	207	-20.1%	1.25x (?)
UTF8Decode_InitFromCustom_noncontiguous	276	239	-13.4%	1.15x (?)
NormalizedIterator_emoji	396	360	-9.1%	1.10x (?)
DistinctClassFieldAccesses	250	229	-8.4%	1.09x (?)
Set.isStrictSubset.Box0	84	77	-8.3%	1.09x (?)
ArrayInClass	1200	1110	-7.5%	1.08x (?)
UnicodeStringFromCodable	202	187	-7.4%	1.08x (?)
String.replaceSubrange.String.Small	41	38	-7.3%	1.08x (?)
RemoveWhereQuadraticString	213	198	-7.0%	1.08x (?)
DictionaryKeysContainsNative	15	14	-6.7%	1.07x (?)

Code size: -O

Performance (x86_64): -Osize

Regression	OLD	NEW	DELTA	RATIO
EqualSubstringSubstring	23	30	+30.4%	0.77x
LessSubstringSubstring	23	29	+26.1%	0.79x (?)
EqualStringSubstring	23	29	+26.1%	0.79x
EqualSubstringSubstringGenericEquatable	23	29	+26.1%	0.79x (?)
EqualSubstringString	23	29	+26.1%	0.79x (?)
LessSubstringSubstringGenericComparable	23	29	+26.1%	0.79x (?)
DataCreateEmptyArray	1400	1600	+14.3%	0.88x (?)
FlattenListFlatMap	3663	4181	+14.1%	0.88x (?)
FindString.Loop1.Substring	264	300	+13.6%	0.88x (?)
StringAdder	221	250	+13.1%	0.88x (?)
StringComparison_longSharedPrefix	294	332	+12.9%	0.89x (?)
StringWalk	1360	1520	+11.8%	0.89x (?)
StringHasSuffixAscii	1350	1480	+9.6%	0.91x (?)
StringInterpolationSmall	1080	1180	+9.3%	0.92x (?)
SortStringsUnicode	1970	2145	+8.9%	0.92x (?)
Improvement	OLD	NEW	DELTA	RATIO
UTF8Decode_InitDecoding	159	125	-21.4%	1.27x (?)
UTF8Decode_InitFromCustom_contiguous	160	127	-20.6%	1.26x (?)
RemoveWhereFilterInts	28	23	-17.9%	1.22x (?)
UTF8Decode_InitFromCustom_noncontiguous	278	243	-12.6%	1.14x (?)
Array2D	4640	4176	-10.0%	1.11x (?)
Set.isDisjoint.Empty.Box	90	82	-8.9%	1.10x (?)
ArrayInClass	1240	1130	-8.9%	1.10x (?)
Set.isSuperset.Seq.Int0	125	114	-8.8%	1.10x (?)
Set.isStrictSubset.Int0	82	75	-8.5%	1.09x (?)
Set.isStrictSubset.Box0	84	77	-8.3%	1.09x (?)
Set.isDisjoint.Seq.Empty.Box	90	83	-7.8%	1.08x (?)
String.replaceSubrange.String.Small	41	38	-7.3%	1.08x (?)

Code size: -Osize

Performance (x86_64): -Onone

Regression	OLD	NEW	DELTA	RATIO
EqualSubstringString	28	35	+25.0%	0.80x (?)
EqualSubstringSubstringGenericEquatable	26	32	+23.1%	0.81x (?)
LessSubstringSubstringGenericComparable	26	32	+23.1%	0.81x
LessSubstringSubstring	28	34	+21.4%	0.82x (?)
EqualSubstringSubstring	28	34	+21.4%	0.82x (?)
EqualStringSubstring	29	35	+20.7%	0.83x (?)
DictionaryCompactMapValuesOfNilValue	10050	10850	+8.0%	0.93x (?)
Improvement	OLD	NEW	DELTA	RATIO
SubstringTrimmingASCIIWhitespace	659	469	-28.8%	1.41x
UTF8Decode_InitDecoding	169	132	-21.9%	1.28x (?)
UTF8Decode_InitFromCustom_contiguous	169	133	-21.3%	1.27x
StringMatch	28900	23400	-19.0%	1.24x (?)
StackPromo	34800	29100	-16.4%	1.20x
FindString.Rec3.Substring	567	482	-15.0%	1.18x (?)
FindString.Rec3.String	580	498	-14.1%	1.16x (?)
String.replaceSubrange.String.Small	44	40	-9.1%	1.10x (?)
ObjectiveCBridgeStubDateAccess	2168	1975	-8.9%	1.10x (?)
ObjectiveCBridgeStubNSDateRefAccess	2207	2028	-8.1%	1.09x (?)

Code size: -swiftlibs

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 mini
  Model Identifier: Macmini8,1
  Processor Name: 6-Core Intel Core i7
  Processor Speed: 3.2 GHz
  Number of Processors: 1
  Total Number of Cores: 6
  L2 Cache (per Core): 256 KB
  L3 Cache: 12 MB
  Memory: 32 GB

[lorentey]

Looks reasonable to me! Another thing we can try is to switch to using @inlinable @inline(__always) in all of these overloads.

[karwa]

FYI #36309

I don't remember the details any more, but it's potentially worth looking at map and flatMap as well.

[eeckstein]

This change lgtm!
The benchmark results are probably just noise.