[mandatory-combine] Add a canonicalization impl and prepare to use mandatory-combine to test ownership RAUW code #34968

gottesmm · 2020-12-04T20:30:53Z

This PR contains a few changes I need to make to mandatory combiner so I can use it to test my ownership RAUW code. Specifically:

I added a canonicalize instruction run to Mandatory Combiner.
I added support in Mandatory Combiner so that if we are running on a function with opts, we eliminate trivially dead instructions. The original design of Mandatory Combiner was thinking about -Onone where we are going to tear down the module right after we run Mandatory Combiner (so it doesn't make sense to spend time deleting instructions). At -O we don't have that issue so we should clean up the IR.
I made it so that in OSSA we treat as trivially dead instructions destroy_value, end_borrow with operands that have OwnershipKind::None. This can happen during SILOptimizatios as a result of offsetting ARC operation insertion combined with non-payloaded or trivial payloaded cases of non-trivial enums.

gottesmm · 2020-12-04T20:31:04Z

@swift-ci test

gottesmm · 2020-12-04T20:31:21Z

preset=buildbot,tools=RA,stdlib=RD,test=non_executable
@swift-ci Please test with preset macOS

swift-ci · 2020-12-04T21:50:03Z

Build failed
Swift Test Linux Platform
Git Sha - 915c7a29797e40fdadf9babc01e083a26254758d

swift-ci · 2020-12-04T22:26:37Z

Build failed
Swift Test OS X Platform
Git Sha - 915c7a29797e40fdadf9babc01e083a26254758d

gottesmm · 2020-12-04T22:53:31Z

@swift-ci benchmark

gottesmm · 2020-12-04T23:52:11Z

The failure of macOS platform non-executable is known: https://ci.swift.org/job/oss-swift_tools-RA_stdlib-DA_test-device-non_executable/289/console

gottesmm · 2020-12-05T00:20:47Z

@swift-ci benchmark

gottesmm · 2020-12-05T00:21:09Z

@swift-ci test OS X platform

gottesmm · 2020-12-05T00:30:26Z

Looked at the Linux runtime failures. What is happening here is that they are testing non-uniqueness by just creating a local variable and we are cleaning it up = p. I am going to change the test to store into a global. That will be preserved and ensure non-uniqueness.

…on while performing mandatory combining.

…y when compiling with opts. When originally made, the thought behind mandatory combine was that at -Onone, we are at the end of the pipeline so we are going to be destroying our module anyways, so there is no reason to delete dead instructions. We are also running it at -O, -Osize, so why not eliminate these instructions when we prepare for sil optimizations! So I put in some conditional code that causes us to delete trivial instructions as we set up our worklist and if one pops off the worklist during traversal. The reason why I am doing this is that I realized that I do not want to use SILGenCleanup to test InstSimplify code. In fact, I do not want any of my RAUW code to be emitted by SILGenCleanup since I don't want to disturb the codegen of the early passes so I don't have to update a bunch of tests = (. So instead, I am going to use mandatory combine for this purpose and doing this cleans up the tests a little bit. It also will just eliminate a bunch of early unnecessary code as well.

…ipKind::None are trivially dead, so make isInstructionTriviallyDead return true for those cases!

swift-ci · 2020-12-05T01:03:51Z

Performance: -O

Regression	OLD	NEW	DELTA	RATIO
LessSubstringSubstring	38	42	+10.5%	0.90x (?)
EqualSubstringSubstring	38	42	+10.5%	0.90x
EqualStringSubstring	39	43	+10.3%	0.91x (?)
EqualSubstringSubstringGenericEquatable	39	43	+10.3%	0.91x
SubstringComparable	12	13	+8.3%	0.92x
ObjectiveCBridgeStubToNSStringRef	113	122	+8.0%	0.93x (?)
StringWalk	2560	2760	+7.8%	0.93x
FloatingPointPrinting_Float_description_uniform	5200	5600	+7.7%	0.93x (?)
EqualSubstringString	39	42	+7.7%	0.93x (?)
LessSubstringSubstringGenericComparable	39	42	+7.7%	0.93x (?)

Improvement	OLD	NEW	DELTA	RATIO
Data.init.Sequence.64kB.Count.I	58	39	-32.8%	1.49x
Data.init.Sequence.64kB.Count	58	39	-32.8%	1.49x
Data.init.Sequence.2047B.Count.I	102	73	-28.4%	1.40x
Data.init.Sequence.2049B.Count.I	102	73	-28.4%	1.40x
Data.init.Sequence.809B.Count	95	73	-23.2%	1.30x
Data.init.Sequence.809B.Count.I	95	73	-23.2%	1.30x
Data.init.Sequence.511B.Count.I	112	88	-21.4%	1.27x (?)
Data.init.Sequence.513B.Count.I	112	91	-18.7%	1.23x
String.data.Medium	128	108	-15.6%	1.19x (?)
String.data.Small	48	43	-10.4%	1.12x (?)
String.data.LargeUnicode	128	116	-9.4%	1.10x (?)
String.data.Empty	44	41	-6.8%	1.07x (?)
CharIndexing_russian_unicodeScalars_Backwards	5480	5120	-6.6%	1.07x (?)

Code size: -O

Regression	OLD	NEW	DELTA	RATIO
SortStrings.o	24538	84416	+244.0%	0.29x
FloatingPointParsing.o	17986	32592	+81.2%	0.55x
SuperChars.o	1569	1985	+26.5%	0.79x
StringWalk.o	33975	39455	+16.1%	0.86x
StringMatch.o	4186	4554	+8.8%	0.92x
StringEdits.o	9985	10753	+7.7%	0.93x
StringReplaceSubrange.o	4843	5189	+7.1%	0.93x
CString.o	6327	6722	+6.2%	0.94x
UTF8Decode.o	23411	24737	+5.7%	0.95x
CharacterProperties.o	23323	24377	+4.5%	0.96x
WordCount.o	37169	38306	+3.1%	0.97x
Substring.o	18304	18709	+2.2%	0.98x
LuhnAlgoEager.o	12582	12805	+1.8%	0.98x
LuhnAlgoLazy.o	12582	12805	+1.8%	0.98x
RC4.o	3479	3535	+1.6%	0.98x
DiffingMyers.o	6725	6829	+1.5%	0.98x
Diffing.o	8379	8483	+1.2%	0.99x
Breadcrumbs.o	42591	43091	+1.2%	0.99x

Performance: -Osize

Regression	OLD	NEW	DELTA	RATIO
EqualSubstringSubstring	38	42	+10.5%	0.90x (?)
EqualStringSubstring	38	42	+10.5%	0.90x (?)
EqualSubstringSubstringGenericEquatable	38	42	+10.5%	0.90x
AngryPhonebook.Armenian.Small	807	875	+8.4%	0.92x (?)
ObjectiveCBridgeToNSDictionary	15000	16250	+8.3%	0.92x (?)
AngryPhonebook.Strasse.Small	948	1027	+8.3%	0.92x (?)
LessSubstringSubstring	39	42	+7.7%	0.93x (?)
EqualSubstringString	39	42	+7.7%	0.93x (?)
LessSubstringSubstringGenericComparable	39	42	+7.7%	0.93x (?)

Improvement	OLD	NEW	DELTA	RATIO
FlattenListLoop	2506	1628	-35.0%	1.54x (?)
CharIteration_tweet_unicodeScalars	7080	4960	-29.9%	1.43x
FlattenListFlatMap	6649	4956	-25.5%	1.34x (?)
CharIteration_ascii_unicodeScalars	3400	2600	-23.5%	1.31x
CharIteration_utf16_unicodeScalars	5320	4760	-10.5%	1.12x (?)
DataCreateSmall	3050	2750	-9.8%	1.11x (?)
CharIteration_punctuated_unicodeScalars	840	760	-9.5%	1.11x
CharIteration_chinese_unicodeScalars	3640	3360	-7.7%	1.08x (?)
StrComplexWalk	5410	5000	-7.6%	1.08x (?)
CharIteration_russian_unicodeScalars	4280	3960	-7.5%	1.08x (?)
StringMatch	7500	7000	-6.7%	1.07x (?)

Code size: -Osize

Regression	OLD	NEW	DELTA	RATIO
SortStrings.o	27045	87467	+223.4%	0.31x
FloatingPointParsing.o	17088	32058	+87.6%	0.53x
SuperChars.o	1462	1857	+27.0%	0.79x
StringWalk.o	32900	38056	+15.7%	0.86x
StringMatch.o	4140	4492	+8.5%	0.92x
StringReplaceSubrange.o	4730	5055	+6.9%	0.94x
StringEdits.o	10048	10717	+6.7%	0.94x
CString.o	6042	6439	+6.6%	0.94x
UTF8Decode.o	22231	23401	+5.3%	0.95x
CharacterProperties.o	22495	23591	+4.9%	0.95x
WordCount.o	36001	37014	+2.8%	0.97x
DiffingMyers.o	6520	6652	+2.0%	0.98x
Substring.o	17673	18023	+2.0%	0.98x
LuhnAlgoEager.o	14913	15116	+1.4%	0.99x
LuhnAlgoLazy.o	14913	15116	+1.4%	0.99x
Breadcrumbs.o	35977	36449	+1.3%	0.99x
DataBenchmarks.o	50594	51176	+1.2%	0.99x

Performance: -Onone

Regression	OLD	NEW	DELTA	RATIO
NSError	853	936	+9.7%	0.91x (?)
DataCreateEmpty	1850	2000	+8.1%	0.93x (?)
ErrorHandling	4370	4710	+7.8%	0.93x (?)
DictionaryBridgeToObjC_Access	1089	1172	+7.6%	0.93x (?)

Improvement	OLD	NEW	DELTA	RATIO
ClassArrayGetter2	5080	4630	-8.9%	1.10x (?)
Integrate	4099	3771	-8.0%	1.09x (?)
StringComparison_abnormal	1060	980	-7.5%	1.08x (?)
ArrayOfPOD	1133	1049	-7.4%	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

gottesmm · 2020-12-05T01:21:32Z

@swift-ci benchmark

swift-ci · 2020-12-05T02:00:23Z

Performance: -O

Regression	OLD	NEW	DELTA	RATIO
LessSubstringSubstring	34	38	+11.8%	0.89x (?)
EqualSubstringSubstring	34	38	+11.8%	0.89x
EqualStringSubstring	35	39	+11.4%	0.90x (?)
EqualSubstringSubstringGenericEquatable	35	39	+11.4%	0.90x
AngryPhonebook.Strasse.Small	868	946	+9.0%	0.92x (?)
StringWalk	2280	2480	+8.8%	0.92x
EqualSubstringString	35	38	+8.6%	0.92x
LessSubstringSubstringGenericComparable	35	38	+8.6%	0.92x
DictionaryBridgeToObjC_Access	741	799	+7.8%	0.93x (?)
CharIteration_punctuated_unicodeScalars_Backwards	1040	1120	+7.7%	0.93x

Improvement	OLD	NEW	DELTA	RATIO
Data.init.Sequence.64kB.Count.I	52	35	-32.7%	1.49x
Data.init.Sequence.64kB.Count	52	35	-32.7%	1.49x
Data.init.Sequence.2047B.Count.I	91	65	-28.6%	1.40x
Data.init.Sequence.2049B.Count.I	91	65	-28.6%	1.40x
Data.init.Sequence.809B.Count	86	65	-24.4%	1.32x
Data.init.Sequence.809B.Count.I	86	65	-24.4%	1.32x
Data.init.Sequence.511B.Count.I	99	79	-20.2%	1.25x (?)
Data.init.Sequence.513B.Count.I	101	81	-19.8%	1.25x (?)
String.data.Small	44	39	-11.4%	1.13x (?)
String.data.Empty	41	37	-9.8%	1.11x (?)
ParseFloat.Float.Exp	12	11	-8.3%	1.09x (?)
CharIndexing_punctuated_unicodeScalars_Backwards	1120	1040	-7.1%	1.08x (?)

Code size: -O

Regression	OLD	NEW	DELTA	RATIO
SortStrings.o	24538	84416	+244.0%	0.29x
FloatingPointParsing.o	17986	32592	+81.2%	0.55x
SuperChars.o	1569	1985	+26.5%	0.79x
StringWalk.o	33975	39455	+16.1%	0.86x
StringMatch.o	4186	4554	+8.8%	0.92x
StringEdits.o	9985	10753	+7.7%	0.93x
StringReplaceSubrange.o	4843	5189	+7.1%	0.93x
CString.o	6327	6722	+6.2%	0.94x
UTF8Decode.o	23411	24737	+5.7%	0.95x
CharacterProperties.o	23323	24377	+4.5%	0.96x
WordCount.o	37169	38306	+3.1%	0.97x
Substring.o	18304	18709	+2.2%	0.98x
LuhnAlgoEager.o	12582	12805	+1.8%	0.98x
LuhnAlgoLazy.o	12582	12805	+1.8%	0.98x
RC4.o	3479	3535	+1.6%	0.98x
DiffingMyers.o	6725	6829	+1.5%	0.98x
Diffing.o	8379	8483	+1.2%	0.99x
Breadcrumbs.o	42591	43091	+1.2%	0.99x

Performance: -Osize

Regression	OLD	NEW	DELTA	RATIO
EqualSubstringSubstring	34	38	+11.8%	0.89x (?)
EqualStringSubstring	34	38	+11.8%	0.89x (?)
EqualSubstringSubstringGenericEquatable	34	38	+11.8%	0.89x (?)
LessSubstringSubstring	35	38	+8.6%	0.92x (?)
EqualSubstringString	35	38	+8.6%	0.92x
LessSubstringSubstringGenericComparable	35	38	+8.6%	0.92x (?)

Improvement	OLD	NEW	DELTA	RATIO
CharIteration_tweet_unicodeScalars	6360	4440	-30.2%	1.43x
CharIteration_ascii_unicodeScalars	3040	2320	-23.7%	1.31x
String.data.LargeUnicode	123	102	-17.1%	1.21x (?)
CharIteration_utf16_unicodeScalars	4760	4240	-10.9%	1.12x (?)
CharIteration_punctuated_unicodeScalars	760	680	-10.5%	1.12x
ObjectiveCBridgeFromNSDictionaryAnyObjectForced	8850	7950	-10.2%	1.11x (?)
DataCreateSmall	2740	2470	-9.9%	1.11x (?)
String.data.Medium	110	101	-8.2%	1.09x (?)
CharIteration_chinese_unicodeScalars	3240	3000	-7.4%	1.08x (?)
CharIteration_russian_unicodeScalars	3840	3560	-7.3%	1.08x (?)
CharIteration_japanese_unicodeScalars	5480	5120	-6.6%	1.07x (?)

Code size: -Osize

Regression	OLD	NEW	DELTA	RATIO
SortStrings.o	27045	87467	+223.4%	0.31x
FloatingPointParsing.o	17088	32058	+87.6%	0.53x
SuperChars.o	1462	1857	+27.0%	0.79x
StringWalk.o	32900	38056	+15.7%	0.86x
StringMatch.o	4140	4492	+8.5%	0.92x
StringReplaceSubrange.o	4730	5055	+6.9%	0.94x
StringEdits.o	10048	10717	+6.7%	0.94x
CString.o	6042	6439	+6.6%	0.94x
UTF8Decode.o	22231	23401	+5.3%	0.95x
CharacterProperties.o	22495	23591	+4.9%	0.95x
WordCount.o	36001	37014	+2.8%	0.97x
DiffingMyers.o	6520	6652	+2.0%	0.98x
Substring.o	17673	18023	+2.0%	0.98x
LuhnAlgoEager.o	14913	15116	+1.4%	0.99x
LuhnAlgoLazy.o	14913	15116	+1.4%	0.99x
Breadcrumbs.o	35977	36449	+1.3%	0.99x
DataBenchmarks.o	50594	51176	+1.2%	0.99x

Performance: -Onone

Regression	OLD	NEW	DELTA	RATIO
UTF8Decode_InitFromBytes_ascii_as_ascii	419	488	+16.5%	0.86x (?)
DataCreateEmpty	1660	1790	+7.8%	0.93x

Improvement	OLD	NEW	DELTA	RATIO
StringToDataSmall	4200	3650	-13.1%	1.15x (?)
StringToDataEmpty	4050	3550	-12.3%	1.14x (?)
NSStringConversion.Mutable	3575	3210	-10.2%	1.11x (?)
ClassArrayGetter2	4590	4150	-9.6%	1.11x
StringComparison_abnormal	960	880	-8.3%	1.09x (?)
Integrate	3679	3381	-8.1%	1.09x (?)
UTF8Decode_InitFromCustom_contiguous_ascii_as_ascii	433	400	-7.6%	1.08x (?)
ArrayOfPOD	996	921	-7.5%	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: 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

swift-ci · 2020-12-05T02:15:03Z

Build failed
Swift Test OS X Platform
Git Sha - 915c7a29797e40fdadf9babc01e083a26254758d

gottesmm · 2020-12-05T04:40:40Z

@swift-ci test

gottesmm · 2020-12-05T04:41:54Z

If I can't figure out quickly where the divergence in these results are coming from, I am just going to enable this thing behind an option and then loop back around. I need this for testing purposes I don't care if in the short term this changes do not run in the pipeline.

swift-ci · 2020-12-05T06:11:43Z

Build failed
Swift Test Linux Platform
Git Sha - c6525fd

swift-ci · 2020-12-05T06:39:20Z

Build failed
Swift Test OS X Platform
Git Sha - c6525fd

gottesmm · 2020-12-06T00:54:27Z

Changed my mind. Going to do this behind a flag for now so I can do my testing and not change code-gen.

gottesmm · 2020-12-06T03:42:39Z

I am going to put this behind an option and use this for tests. In the future I am going to loop back around on this and see what is going on.

…g so I can use it for writing OSSA RAUW tests. I am going to loop back around and enable this all the time, but this is seeming to disturb codegen more than expected, so I am going to enable it everywhere later (since it is not my primary goal).

gottesmm · 2020-12-06T03:58:01Z

@swift-ci benchmark

zoecarver · 2020-12-06T04:03:03Z

lib/SILOptimizer/Mandatory/MandatoryCombine.cpp

@@ -156,6 +214,12 @@ void MandatoryCombiner::addReachableCodeToWorklist(SILFunction &function) {
      ++iterator;

      if (isInstructionTriviallyDead(instruction)) {


We're already "doing the work" in Onone. We should either just erase the known-dead instruction, or not call isInstructionTriviallyDead at all (preferably the later).

We need to know to skip it. We are keeping the size of the worklist down.

Fair enough.

zoecarver · 2020-12-06T04:04:43Z

test/SILOptimizer/mandatory_combiner.sil

@@ -75,6 +75,12 @@ sil @myint_from_myint_and_proto : $@convention(thin) (MyInt, @guaranteed { var P

 sil @myint_from_proto_and_myint : $@convention(thin) (@guaranteed { var Proto }, MyInt) -> MyInt

+// Optional support
+enum FakeOptional<T> {


Looks like this declaration isn't actually used in the test...?

It would be good to test a Swift program that uses this type, though. That's where we've run into problems with this optimization in the past.

This is just misc boiler plate tests that I add in most tests. Its useful to have.

zoecarver · 2020-12-06T04:10:36Z

lib/SILOptimizer/Mandatory/MandatoryCombine.cpp

@@ -156,6 +214,12 @@ void MandatoryCombiner::addReachableCodeToWorklist(SILFunction &function) {
      ++iterator;

      if (isInstructionTriviallyDead(instruction)) {
+        if (EnableCanonicalizationAndTrivialDCE) {


Could these be flattened into one if?

This is something that is meant to be deleted soon. I think it is fine given that this is temporary. I need this for testing purposes for something else.

zoecarver · 2020-12-06T04:12:51Z

@gottesmm just FYI I made a similar change to remove dead instructions in #30429. It was then reverted (in #31077) because of a bug with enums. I have a fix for that enum bug in #31779 and plan to re-land that patch once the enum thing is fixed. This is behind a flag so it should be fine, but maybe still good to run a full source compatibility test.

swift-ci · 2020-12-06T04:31:59Z

Performance: -O

Regression	OLD	NEW	DELTA	RATIO
Data.append.Sequence.809B.Count	92	102	+10.9%	0.90x (?)

Improvement	OLD	NEW	DELTA	RATIO
ObjectiveCBridgeStubToNSStringRef	122	114	-6.6%	1.07x (?)

Code size: -O

Performance: -Osize

Regression	OLD	NEW	DELTA	RATIO
FlattenListFlatMap	6044	6775	+12.1%	0.89x (?)

Code size: -Osize

Performance: -Onone

Improvement	OLD	NEW	DELTA	RATIO
DataCreateSmallArray	46150	38900	-15.7%	1.19x (?)

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

gottesmm · 2020-12-06T05:09:33Z

@zoecarver I think I may have fixed that issue with my ArgumentScope change.

gottesmm · 2020-12-06T05:11:08Z

@swift-ci smoke test

zoecarver · 2020-12-06T05:22:11Z

@zoecarver I think I may have fixed that issue with my ArgumentScope change.

Thanks. I'll see if the tests pass on main.

gottesmm requested a review from meg-gupta December 4, 2020 20:30

gottesmm added 3 commits December 4, 2020 16:47

[mandatory-combine] Just like SILCombine, perform SIL canonicializati…

8f1b353

…on while performing mandatory combining.

[ownership] end_borrow/destroy_value with an operand that has Ownersh…

c6525fd

…ipKind::None are trivially dead, so make isInstructionTriviallyDead return true for those cases!

gottesmm force-pushed the pr-50e8ac8f699f7e3ef7096b4b56510ffc897669e5 branch from 915c7a2 to c6525fd Compare December 5, 2020 00:47

zoecarver reviewed Dec 6, 2020

View reviewed changes

gottesmm merged commit 27ece6f into swiftlang:main Dec 6, 2020

gottesmm deleted the pr-50e8ac8f699f7e3ef7096b4b56510ffc897669e5 branch December 6, 2020 22:43

		@@ -156,6 +214,12 @@ void MandatoryCombiner::addReachableCodeToWorklist(SILFunction &function) {
		++iterator;

		if (isInstructionTriviallyDead(instruction)) {

[mandatory-combine] Add a canonicalization impl and prepare to use mandatory-combine to test ownership RAUW code #34968

[mandatory-combine] Add a canonicalization impl and prepare to use mandatory-combine to test ownership RAUW code #34968

Uh oh!

Conversation

gottesmm commented Dec 4, 2020

Uh oh!

gottesmm commented Dec 4, 2020

Uh oh!

gottesmm commented Dec 4, 2020

Uh oh!

swift-ci commented Dec 4, 2020

Uh oh!

swift-ci commented Dec 4, 2020

Uh oh!

gottesmm commented Dec 4, 2020

Uh oh!

gottesmm commented Dec 4, 2020

Uh oh!

gottesmm commented Dec 5, 2020

Uh oh!

gottesmm commented Dec 5, 2020

Uh oh!

gottesmm commented Dec 5, 2020

Uh oh!

swift-ci commented Dec 5, 2020

Performance: -O

Code size: -O

Performance: -Osize

Code size: -Osize

Performance: -Onone

Code size: -swiftlibs

Uh oh!

gottesmm commented Dec 5, 2020

Uh oh!

swift-ci commented Dec 5, 2020

Performance: -O

Code size: -O

Performance: -Osize

Code size: -Osize

Performance: -Onone

Code size: -swiftlibs

Uh oh!

swift-ci commented Dec 5, 2020

Uh oh!

gottesmm commented Dec 5, 2020

Uh oh!

gottesmm commented Dec 5, 2020

Uh oh!

swift-ci commented Dec 5, 2020

Uh oh!

swift-ci commented Dec 5, 2020

Uh oh!

gottesmm commented Dec 6, 2020

Uh oh!

gottesmm commented Dec 6, 2020

Uh oh!

gottesmm commented Dec 6, 2020

Uh oh!

zoecarver Dec 6, 2020

Choose a reason for hiding this comment

Uh oh!

gottesmm Dec 6, 2020

Choose a reason for hiding this comment

Uh oh!

zoecarver Dec 6, 2020

Choose a reason for hiding this comment

Uh oh!

zoecarver Dec 6, 2020

Choose a reason for hiding this comment

Uh oh!

gottesmm Dec 6, 2020

Choose a reason for hiding this comment

Uh oh!

zoecarver Dec 6, 2020

Choose a reason for hiding this comment

Uh oh!

gottesmm Dec 6, 2020

Choose a reason for hiding this comment

Uh oh!

zoecarver commented Dec 6, 2020