Simplify {Float,Double,Float80}.init(_: String)

[tbkka]

The original version scanned the entire input string for whitespace and
non-ASCII characters. Both are unnecessary: the C routines we're building on
already stop at non-ASCII characters or non-leading whitespace. So we need only
check the first character for whitespace and verify that all characters are
consumed.

This demonstrates significant improvements for -Onone benchmarks:
about 4k smaller RSS and 3x speedups. I'm not seeing any local
changes to the -O benchmarks, however.

Resolves rdar://33045403

Future: We could further improve this by pushing these checks down into the runtime
stubs and simplifying the ABI so those stubs just returne a success/failure boolean.
In particular, that would avoid inlining these checks into every client.

[tbkka]

@swift-ci Please smoke test Linux

[tbkka]

@swift-ci Please smoke test macOS

[compnerd]

@swift-ci please benchmark

[swift-ci]

Performance: -O

Regression	OLD	NEW	DELTA	RATIO
ObjectiveCBridgeStringHash	67	78	+16.4%	0.86x
Improvement	OLD	NEW	DELTA	RATIO
ParseFloat.Double.Exp	18	12	-33.3%	1.50x
ParseFloat.Float.Exp	17	13	-23.5%	1.31x
ParseFloat.Float.Uniform	40	32	-20.0%	1.25x
ParseFloat.Double.Uniform	59	50	-15.3%	1.18x (?)
ParseFloat.Float80.Exp	54	47	-13.0%	1.15x (?)
EqualSubstringSubstring	42	39	-7.1%	1.08x (?)
LessSubstringSubstring	42	39	-7.1%	1.08x (?)
EqualSubstringSubstringGenericEquatable	42	39	-7.1%	1.08x (?)
EqualSubstringString	42	39	-7.1%	1.08x (?)
LessSubstringSubstringGenericComparable	42	39	-7.1%	1.08x (?)

Code size: -O

Improvement	OLD	NEW	DELTA	RATIO
FloatingPointParsing.o	15274	13649	-10.6%	1.12x

Performance: -Osize

Regression	OLD	NEW	DELTA	RATIO
FlattenListLoop	4451	5302	+19.1%	0.84x (?)
ObjectiveCBridgeStringHash	68	78	+14.7%	0.87x (?)
NSStringConversion.MutableCopy.UTF8	788	859	+9.0%	0.92x (?)
Improvement	OLD	NEW	DELTA	RATIO
ParseFloat.Double.Exp	27	12	-55.6%	2.25x
ParseFloat.Float.Exp	22	13	-40.9%	1.69x
ParseFloat.Float.Uniform	51	32	-37.3%	1.59x
ParseFloat.Double.Uniform	71	50	-29.6%	1.42x
ParseFloat.Float80.Exp	62	47	-24.2%	1.32x
ObjectiveCBridgeFromNSArrayAnyObjectToString	44700	41200	-7.8%	1.08x (?)
EqualSubstringString	42	39	-7.1%	1.08x (?)
EqualSubstringSubstring	43	40	-7.0%	1.07x (?)
LessSubstringSubstring	43	40	-7.0%	1.07x (?)
EqualStringSubstring	43	40	-7.0%	1.07x (?)
EqualSubstringSubstringGenericEquatable	43	40	-7.0%	1.07x
LessSubstringSubstringGenericComparable	43	40	-7.0%	1.07x

Code size: -Osize

Improvement	OLD	NEW	DELTA	RATIO
FloatingPointParsing.o	14696	13257	-9.8%	1.11x

Performance: -Onone

Regression	OLD	NEW	DELTA	RATIO
ArrayAppendLatin1	25024	30940	+23.6%	0.81x (?)
ArrayAppendAscii	24752	30464	+23.1%	0.81x (?)
ArrayAppendUTF16	24888	30566	+22.8%	0.81x (?)
ObjectiveCBridgeStringHash	67	78	+16.4%	0.86x (?)
ArrayPlusEqualSingleElementCollection	214508	245387	+14.4%	0.87x (?)
ArrayPlusEqualFiveElementCollection	171643	195101	+13.7%	0.88x (?)
Dictionary4Legacy	1275	1409	+10.5%	0.90x (?)
ArrayOfPOD	1037	1121	+8.1%	0.93x (?)
Improvement	OLD	NEW	DELTA	RATIO
ObjectiveCBridgeStubFromNSDateRef	5370	4280	-20.3%	1.25x (?)
ParseFloat.Double.Exp	49	43	-12.2%	1.14x (?)
ParseFloat.Float.Uniform	74	65	-12.2%	1.14x (?)
ParseFloat.Float80.Exp	89	79	-11.2%	1.13x (?)
ParseFloat.Double.Uniform	93	84	-9.7%	1.11x (?)
DictionaryCompactMapValuesOfCastValue	67284	61128	-9.1%	1.10x (?)
DictionaryBridgeToObjC_Access	1224	1116	-8.8%	1.10x (?)
StringBuilderWithLongSubstring	2850	2620	-8.1%	1.09x (?)
LessSubstringSubstringGenericComparable	50	46	-8.0%	1.09x (?)
EqualSubstringString	51	47	-7.8%	1.09x
StringUTF16SubstringBuilder	17910	16530	-7.7%	1.08x (?)

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 Pro
  Model Identifier: MacPro6,1
  Processor Name: 12-Core Intel Xeon E5
  Processor Speed: 2.7 GHz
  Number of Processors: 1
  Total Number of Cores: 12
  L2 Cache (per Core): 256 KB
  L3 Cache: 30 MB
  Memory: 64 GB

[stephentyrone]

Room to make this cleaner still, but good enough start to take now.