Enable DCE on OSSA

[meg-gupta]

• Support destroy_value/copy_value/begin_borrow/end_borrow
• Support adding reverse dependencies for SILArgument
• Insert lifetime end instructions (destroy_value/end_borrow) while deleting instructions which would have ended the lifetime of its operands

[meg-gupta]

@swift-ci test

[meg-gupta]

@swift-ci test source compatibility

[meg-gupta]

@swift-ci benchmark

[swift-ci]

Performance: -O

Regression	OLD	NEW	DELTA	RATIO
StringFromLongWholeSubstring	4	5	+25.0%	0.80x
UTF8Decode_InitFromBytes_ascii_as_ascii	445	520	+16.9%	0.86x (?)
ArrayInitFromSlice	458	517	+12.9%	0.89x (?)
LessSubstringSubstringGenericComparable	39	43	+10.3%	0.91x
DataSetCountSmall	131	142	+8.4%	0.92x (?)
EqualSubstringString	39	42	+7.7%	0.93x
Improvement	OLD	NEW	DELTA	RATIO
StringWalk	2560	1280	-50.0%	2.00x
String.data.LargeUnicode	126	100	-20.6%	1.26x (?)
HashTest	1580	1260	-20.3%	1.25x (?)
StrComplexWalk	2840	2310	-18.7%	1.23x
CSVParsingAlt2	1716	1474	-14.1%	1.16x
RemoveWhereFilterString	342	300	-12.3%	1.14x
String.replaceSubrange.Substring.Small	84	75	-10.7%	1.12x (?)
String.replaceSubrange.String.Small	67	60	-10.4%	1.12x (?)
CharIteration_russian_unicodeScalars	3840	3440	-10.4%	1.12x (?)
StringBuilderLong	1460	1310	-10.3%	1.11x (?)
StringBuilderWithLongSubstring	1730	1590	-8.1%	1.09x (?)
ObjectiveCBridgeStringHash	136	125	-8.1%	1.09x (?)
CSVParsing.Scalar	201	185	-8.0%	1.09x (?)
Data.hash.Empty	77	71	-7.8%	1.08x (?)
RandomShuffleLCG2	480	448	-6.7%	1.07x (?)

Code size: -O

Regression	OLD	NEW	DELTA	RATIO
StrComplexWalk.o	2578	2642	+2.5%	0.98x

Performance: -Osize

Regression	OLD	NEW	DELTA	RATIO
StringFromLongWholeSubstring	4	5	+25.0%	0.80x
ArrayInitFromSlice	458	525	+14.6%	0.87x (?)
PrefixWhileAnySequenceLazy	1580	1792	+13.4%	0.88x (?)
CharIteration_korean_unicodeScalars	3880	4360	+12.4%	0.89x (?)
CharIteration_japanese_unicodeScalars	5280	5840	+10.6%	0.90x (?)
LessSubstringSubstring	39	43	+10.3%	0.91x
LessSubstringSubstringGenericComparable	39	43	+10.3%	0.91x
CharIteration_chinese_unicodeScalars	3160	3440	+8.9%	0.92x
DataSetCountSmall	134	145	+8.2%	0.92x
EqualSubstringSubstring	39	42	+7.7%	0.93x (?)
EqualStringSubstring	39	42	+7.7%	0.93x (?)
EqualSubstringSubstringGenericEquatable	39	42	+7.7%	0.93x
EqualSubstringString	39	42	+7.7%	0.93x
Improvement	OLD	NEW	DELTA	RATIO
StringWalk	6080	2160	-64.5%	2.81x
FlattenListLoop	2520	1630	-35.3%	1.55x (?)
HashTest	1660	1320	-20.5%	1.26x
CSVParsingAlt2	1826	1540	-15.7%	1.19x
String.replaceSubrange.Substring.Small	84	75	-10.7%	1.12x (?)
FloatingPointPrinting_Float_description_small	5616	5076	-9.6%	1.11x (?)
String.replaceSubrange.String.Small	66	60	-9.1%	1.10x (?)
ObjectiveCBridgeFromNSSetAnyObjectToStringForced	91000	83500	-8.2%	1.09x (?)
ObjectiveCBridgeStringHash	136	125	-8.1%	1.09x (?)
FlattenListFlatMap	6614	6087	-8.0%	1.09x (?)
DictionaryBridgeToObjC_Access	916	845	-7.8%	1.08x (?)
Array2D	7504	6928	-7.7%	1.08x (?)
Data.hash.Empty	80	74	-7.5%	1.08x (?)
StrComplexWalk	5170	4790	-7.4%	1.08x
RandomShuffleLCG2	448	416	-7.1%	1.08x
AngryPhonebook.Cyrillic.Small	830	773	-6.9%	1.07x (?)
UTF8Decode_InitDecoding_ascii	279	260	-6.8%	1.07x (?)

Code size: -Osize

Improvement	OLD	NEW	DELTA	RATIO
DiffingMyers.o	6395	6330	-1.0%	1.01x

Performance: -Onone

Regression	OLD	NEW	DELTA	RATIO
DataReplaceSmallBuffer	5300	6800	+28.3%	0.78x (?)
EqualSubstringSubstringGenericEquatable	117	145	+23.9%	0.81x
LessSubstringSubstringGenericComparable	117	145	+23.9%	0.81x
StringRemoveDupes	748	925	+23.7%	0.81x
EqualSubstringSubstring	120	148	+23.3%	0.81x
EqualSubstringString	123	151	+22.8%	0.81x
CharIndexing_japanese_unicodeScalars_Backwards	411560	502480	+22.1%	0.82x (?)
EqualStringSubstring	123	150	+22.0%	0.82x
LessSubstringSubstring	122	148	+21.3%	0.82x
CharIndexing_ascii_unicodeScalars_Backwards	337560	408200	+20.9%	0.83x (?)
CharIndexing_russian_unicodeScalars_Backwards	288000	345120	+19.8%	0.83x (?)
CharIndexing_punctuatedJapanese_unicodeScalars_Backwards	60440	72280	+19.6%	0.84x (?)
CharIndexing_tweet_unicodeScalars_Backwards	680080	810720	+19.2%	0.84x (?)
DictionaryRemove	27110	32290	+19.1%	0.84x (?)
CharIndexing_chinese_unicodeScalars_Backwards	262280	312040	+19.0%	0.84x (?)
CharIndexing_punctuated_unicodeScalars_Backwards	77000	91520	+18.9%	0.84x (?)
DictionarySwap	6440	7604	+18.1%	0.85x (?)
CharIndexing_korean_unicodeScalars_Backwards	341840	402600	+17.8%	0.85x (?)
UTF8Decode_InitFromBytes_ascii_as_ascii	467	550	+17.8%	0.85x (?)
BitCount	50857	59606	+17.2%	0.85x (?)
StringInterpolationSmall	4600	5390	+17.2%	0.85x (?)
TwoSum	3748	4368	+16.5%	0.86x (?)
ErrorHandling	4250	4770	+12.2%	0.89x (?)
DataReplaceMedium	6400	7100	+10.9%	0.90x (?)
DataToStringSmall	5950	6600	+10.9%	0.90x (?)
Histogram	7534	8109	+7.6%	0.93x (?)
Improvement	OLD	NEW	DELTA	RATIO
UTF8Decode_InitFromData_ascii_as_ascii	760	660	-13.2%	1.15x (?)
Set.isStrictSuperset.Seq.Empty.Int	1624	1444	-11.1%	1.12x (?)
String.replaceSubrange.Substring.Small	90	82	-8.9%	1.10x (?)
String.replaceSubrange.String.Small	72	66	-8.3%	1.09x (?)
ObjectiveCBridgeStringHash	136	125	-8.1%	1.09x (?)
ObjectiveCBridgeStubDateMutation	1444	1345	-6.9%	1.07x (?)
UTF8Decode_InitDecoding_ascii_as_ascii	327	305	-6.7%	1.07x (?)

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

[eeckstein]

LGTM!

[meg-gupta]

Thanks for the review @eeckstein

[meg-gupta]

@swift-ci test and merge