[stdlib] Improve set isDisjoint by iterating on smaller set #39263

LucianoPAlmeida · 2021-09-11T17:46:30Z

Just a small improvement in Set.isDisjoint by iterating on smaller set.

LucianoPAlmeida · 2021-09-11T17:46:47Z

@swift-ci Please smoke benchmark

swift-ci · 2021-09-11T19:16:11Z

Performance (x86_64): -O

Regression	OLD	NEW	DELTA	RATIO
DictionaryOfAnyHashableStrings_insert	3402	5614	+65.0%	0.61x (?)
Set.isDisjoint.Int25	268	360	+34.3%	0.74x (?)
Set.isDisjoint.Int100	262	348	+32.8%	0.75x
Set.isDisjoint.Int50	268	351	+31.0%	0.76x (?)
Set.isDisjoint.Box25	355	448	+26.2%	0.79x (?)
StringFromLongWholeSubstring	4	5	+25.0%	0.80x
UTF8Decode_InitDecoding_ascii	240	283	+17.9%	0.85x (?)
UTF8Decode_InitFromCustom_contiguous_ascii	243	283	+16.5%	0.86x (?)
DictionaryRemove	3180	3510	+10.4%	0.91x (?)
CStringLongAscii	222	242	+9.0%	0.92x (?)
NSStringConversion.MutableCopy.Rebridge.Medium	618	668	+8.1%	0.93x (?)

Improvement	OLD	NEW	DELTA	RATIO
Set.isDisjoint.Box.Empty	158	90	-43.0%	1.76x
Set.isDisjoint.Int.Empty	142	81	-43.0%	1.75x
FlattenListFlatMap	6655	4327	-35.0%	1.54x (?)
RemoveWhereFilterString	327	296	-9.5%	1.10x
EqualSubstringSubstring	43	39	-9.3%	1.10x (?)
ArrayInClass	2025	1855	-8.4%	1.09x (?)
DistinctClassFieldAccesses	417	382	-8.4%	1.09x
DictionaryOfAnyHashableStrings_lookup	4512	4152	-8.0%	1.09x (?)
Array2D	7808	7216	-7.6%	1.08x (?)
LessSubstringSubstring	42	39	-7.1%	1.08x (?)
EqualStringSubstring	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

Performance (x86_64): -Osize

Regression	OLD	NEW	DELTA	RATIO
Set.isDisjoint.Int100	268	345	+28.7%	0.78x
StringFromLongWholeSubstring	4	5	+25.0%	0.80x

Improvement	OLD	NEW	DELTA	RATIO
Set.isDisjoint.Int.Empty	144	81	-43.7%	1.78x
Set.isDisjoint.Box.Empty	155	90	-41.9%	1.72x
Set.isDisjoint.Box25	509	451	-11.4%	1.13x (?)
LessSubstringSubstring	43	39	-9.3%	1.10x (?)
DictionaryOfAnyHashableStrings_lookup	4536	4128	-9.0%	1.10x (?)
DistinctClassFieldAccesses	402	368	-8.5%	1.09x
ArrayInClass	2005	1840	-8.2%	1.09x (?)
RandomShuffleLCG2	448	416	-7.1%	1.08x (?)
EqualSubstringSubstring	42	39	-7.1%	1.08x (?)
EqualStringSubstring	42	39	-7.1%	1.08x (?)
EqualSubstringString	42	39	-7.1%	1.08x (?)
LessSubstringSubstringGenericComparable	42	39	-7.1%	1.08x (?)
Array2D	7520	6992	-7.0%	1.08x (?)
EqualSubstringSubstringGenericEquatable	43	40	-7.0%	1.07x (?)

Code size: -Osize

Performance (x86_64): -Onone

Regression	OLD	NEW	DELTA	RATIO
StringBuilderWithLongSubstring	3830	4440	+15.9%	0.86x (?)
Data.append.Sequence.809B.Count.RE	36919	42389	+14.8%	0.87x (?)
DataAppendSequence	3672800	4216300	+14.8%	0.87x (?)
Data.init.Sequence.809B.Count.RE	36932	42328	+14.6%	0.87x (?)
Data.init.Sequence.64kB.Count.RE.I	29795	34067	+14.3%	0.87x (?)
Data.append.Sequence.809B.Count.RE.I	36906	42167	+14.3%	0.88x (?)
Data.init.Sequence.809B.Count.RE.I	36981	42116	+13.9%	0.88x (?)
Data.append.Sequence.64kB.Count.RE.I	29688	33798	+13.8%	0.88x (?)
Data.append.Sequence.64kB.Count.RE	29745	33836	+13.8%	0.88x (?)
ArrayAppendRepeatCol	365960	415540	+13.5%	0.88x (?)
Data.init.Sequence.64kB.Count.RE	29794	33748	+13.3%	0.88x (?)
Breadcrumbs.MutatedIdxToUTF16.ASCII	11	12	+9.1%	0.92x (?)
Set.isDisjoint.Seq.Empty.Int	415	449	+8.2%	0.92x (?)
ErrorHandling	3880	4190	+8.0%	0.93x (?)

Improvement	OLD	NEW	DELTA	RATIO
Set.isDisjoint.Box25	2452	1239	-49.5%	1.98x
Set.isDisjoint.Int.Empty	734	392	-46.6%	1.87x
Set.isDisjoint.Box.Empty	778	426	-45.2%	1.83x
Set.isDisjoint.Int25	1121	699	-37.6%	1.60x
Set.isDisjoint.Int50	1079	699	-35.2%	1.54x
Set.isDisjoint.Empty.Int	576	381	-33.9%	1.51x
Set.isDisjoint.Empty.Box	615	428	-30.4%	1.44x
Set.isDisjoint.Int100	895	699	-21.9%	1.28x
Set.isDisjoint.Int0	817	737	-9.8%	1.11x (?)
DictionaryOfAnyHashableStrings_lookup	5376	4872	-9.4%	1.10x (?)
NSError	666	616	-7.5%	1.08x (?)
String.replaceSubrange.String	27	25	-7.4%	1.08x (?)
Dictionary2	1670	1555	-6.9%	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

LucianoPAlmeida · 2021-09-11T19:30:22Z

@swift-ci Please benchmark

swift-ci · 2021-09-11T20:43:25Z

Performance (x86_64): -O

Regression	OLD	NEW	DELTA	RATIO
DictionaryOfAnyHashableStrings_insert	2142	3738	+74.5%	0.57x
Set.isDisjoint.Int50	199	278	+39.7%	0.72x (?)
Set.isDisjoint.Int25	201	278	+38.3%	0.72x (?)
Set.isDisjoint.Int100	199	274	+37.7%	0.73x
UTF8Decode_InitDecoding	125	157	+25.6%	0.80x
Set.isDisjoint.Box25	284	356	+25.4%	0.80x (?)
UTF8Decode_InitFromCustom_contiguous	124	155	+25.0%	0.80x
UTF8Decode_InitFromCustom_noncontiguous	318	352	+10.7%	0.90x
SequenceAlgosUnfoldSequence	790	850	+7.6%	0.93x (?)

Improvement	OLD	NEW	DELTA	RATIO
Set.isDisjoint.Box.Empty	96	51	-46.9%	1.88x
Set.isDisjoint.Int.Empty	86	46	-46.5%	1.87x
FlattenListFlatMap	3962	2389	-39.7%	1.66x (?)
EqualSubstringSubstring	30	22	-26.7%	1.36x
LessSubstringSubstring	29	22	-24.1%	1.32x
EqualStringSubstring	29	22	-24.1%	1.32x
EqualSubstringSubstringGenericEquatable	29	22	-24.1%	1.32x
EqualSubstringString	29	22	-24.1%	1.32x
LessSubstringSubstringGenericComparable	29	22	-24.1%	1.32x
ArrayLiteral2	78	70	-10.3%	1.11x
StringComparison_longSharedPrefix	357	321	-10.1%	1.11x (?)
Set.isSubset.Seq.Empty.Int	83	76	-8.4%	1.09x (?)
SortStringsUnicode	2180	2000	-8.3%	1.09x (?)
Set.isSubset.Int.Empty	49	45	-8.2%	1.09x (?)

Code size: -O

Performance (x86_64): -Osize

Regression	OLD	NEW	DELTA	RATIO
Set.isDisjoint.Int100	207	280	+35.3%	0.74x
UTF8Decode_InitDecoding	123	159	+29.3%	0.77x
UTF8Decode_InitFromCustom_contiguous	125	156	+24.8%	0.80x
Set.isDisjoint.Int50	243	280	+15.2%	0.87x
Set.isDisjoint.Int25	245	280	+14.3%	0.88x
UTF8Decode_InitFromCustom_noncontiguous	246	275	+11.8%	0.89x
SortSortedStrings	45	49	+8.9%	0.92x (?)

Improvement	OLD	NEW	DELTA	RATIO
Set.isDisjoint.Int.Empty	91	51	-44.0%	1.78x
Set.isDisjoint.Box.Empty	96	55	-42.7%	1.75x
EqualSubstringSubstring	30	22	-26.7%	1.36x
LessSubstringSubstringGenericComparable	30	22	-26.7%	1.36x
LessSubstringSubstring	29	22	-24.1%	1.32x
EqualStringSubstring	29	22	-24.1%	1.32x (?)
EqualSubstringSubstringGenericEquatable	29	22	-24.1%	1.32x
EqualSubstringString	29	22	-24.1%	1.32x (?)
StringComparison_longSharedPrefix	357	321	-10.1%	1.11x (?)
Set.isDisjoint.Box25	395	361	-8.6%	1.09x (?)
SortStringsUnicode	2215	2025	-8.6%	1.09x (?)
Set.isDisjoint.Empty.Int	87	81	-6.9%	1.07x (?)

Code size: -Osize

Performance (x86_64): -Onone

Regression	OLD	NEW	DELTA	RATIO
UTF8Decode_InitDecoding	132	165	+25.0%	0.80x (?)
UTF8Decode_InitFromCustom_contiguous	134	167	+24.6%	0.80x

Improvement	OLD	NEW	DELTA	RATIO
Set.isDisjoint.Int.Empty	459	228	-50.3%	2.01x
Set.isDisjoint.Box.Empty	481	241	-49.9%	2.00x
Set.isDisjoint.Box25	1687	868	-48.5%	1.94x
Set.isDisjoint.Int25	731	473	-35.3%	1.55x
Set.isDisjoint.Empty.Int	359	238	-33.7%	1.51x
Set.isDisjoint.Int50	707	473	-33.1%	1.49x
Set.isDisjoint.Empty.Box	375	252	-32.8%	1.49x
Set.isDisjoint.Int100	594	473	-20.4%	1.26x
EqualSubstringSubstring	59	51	-13.6%	1.16x (?)
EqualSubstringSubstringGenericEquatable	57	50	-12.3%	1.14x (?)
LessSubstringSubstring	58	51	-12.1%	1.14x (?)
EqualSubstringString	58	51	-12.1%	1.14x (?)
LessSubstringSubstringGenericComparable	56	50	-10.7%	1.12x (?)
EqualStringSubstring	57	51	-10.5%	1.12x (?)
Set.isDisjoint.Int0	564	524	-7.1%	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 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: 64 GB

lorentey

This is good change, thank you! We should keep the sequence-based code paths intact (see note).

stdlib/public/core/Set.swift

LucianoPAlmeida · 2021-09-11T22:14:13Z

@swift-ci Please benchmark

swift-ci · 2021-09-11T22:46:07Z

Performance (x86_64): -O

Regression	OLD	NEW	DELTA	RATIO
DictionaryOfAnyHashableStrings_insert	2100	3752	+78.7%	0.56x
Set.isDisjoint.Int50	197	277	+40.6%	0.71x (?)
Set.isDisjoint.Int100	197	277	+40.6%	0.71x
Set.isDisjoint.Int25	201	274	+36.3%	0.73x (?)
Set.isDisjoint.Box25	285	359	+26.0%	0.79x (?)
StringRemoveDupes	180	199	+10.6%	0.90x (?)

Improvement	OLD	NEW	DELTA	RATIO
Set.isDisjoint.Box.Empty	96	54	-43.7%	1.78x
Set.isDisjoint.Int.Empty	84	49	-41.7%	1.71x
FlattenListFlatMap	3204	2684	-16.2%	1.19x (?)

Code size: -O

Performance (x86_64): -Osize

Regression	OLD	NEW	DELTA	RATIO
Set.isDisjoint.Int100	208	280	+34.6%	0.74x
FlattenListFlatMap	3328	3893	+17.0%	0.85x (?)
Set.isDisjoint.Int50	243	279	+14.8%	0.87x
Set.isDisjoint.Int25	245	280	+14.3%	0.88x
FlattenListLoop	1387	1567	+13.0%	0.89x (?)

Improvement	OLD	NEW	DELTA	RATIO
Set.isDisjoint.Int.Empty	91	51	-44.0%	1.78x
Set.isDisjoint.Box.Empty	97	56	-42.3%	1.73x
Set.isDisjoint.Box25	400	361	-9.7%	1.11x
Data.hash.Empty	52	47	-9.6%	1.11x (?)

Code size: -Osize

Performance (x86_64): -Onone

Improvement	OLD	NEW	DELTA	RATIO
Set.isDisjoint.Int.Empty	459	233	-49.2%	1.97x
Set.isDisjoint.Box25	1689	864	-48.8%	1.95x
Set.isDisjoint.Box.Empty	481	251	-47.8%	1.92x
Set.isDisjoint.Int25	732	476	-35.0%	1.54x
Set.isDisjoint.Empty.Int	360	241	-33.1%	1.49x
Set.isDisjoint.Empty.Box	378	254	-32.8%	1.49x
Set.isDisjoint.Int50	707	476	-32.7%	1.49x
Set.isDisjoint.Int100	594	476	-19.9%	1.25x
Set.isDisjoint.Int0	563	524	-6.9%	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 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: 64 GB

LucianoPAlmeida · 2021-09-14T20:29:52Z

@swift-ci Please benchmark

swift-ci · 2021-09-14T22:00:40Z

Performance (x86_64): -O

Regression	OLD	NEW	DELTA	RATIO
DictionaryOfAnyHashableStrings_insert	2898	5012	+72.9%	0.58x
Set.isDisjoint.Int25	241	315	+30.7%	0.77x (?)
Set.isDisjoint.Int100	237	305	+28.7%	0.78x
Set.isDisjoint.Box25	325	401	+23.4%	0.81x (?)
FlattenListLoop	1454	1790	+23.1%	0.81x (?)
DictionaryKeysContainsNative	19	23	+21.1%	0.83x (?)
String.data.Medium	91	109	+19.8%	0.83x (?)
String.data.Small	37	42	+13.5%	0.88x (?)
String.data.LargeUnicode	93	102	+9.7%	0.91x (?)
DropLastAnyCollectionLazy	25071	27068	+8.0%	0.93x (?)

Improvement	OLD	NEW	DELTA	RATIO
Set.isDisjoint.Smaller.Box0	758	284	-62.5%	2.67x
Set.isDisjoint.Smaller.Int0	362	149	-58.8%	2.43x
Set.isDisjoint.Box.Empty	147	80	-45.6%	1.84x
Set.isDisjoint.Int.Empty	126	74	-41.3%	1.70x
DictionaryOfAnyHashableStrings_lookup	4032	3720	-7.7%	1.08x (?)

Code size: -O

Performance (x86_64): -Osize

Regression	OLD	NEW	DELTA	RATIO
Set.isDisjoint.Int100	236	306	+29.7%	0.77x

Improvement	OLD	NEW	DELTA	RATIO
Set.isDisjoint.Smaller.Box0	744	298	-59.9%	2.50x
Set.isDisjoint.Smaller.Int0	369	149	-59.6%	2.48x
Set.isDisjoint.Box.Empty	152	82	-46.1%	1.85x
Set.isDisjoint.Int.Empty	129	71	-45.0%	1.82x
Set.isDisjoint.Box25	458	414	-9.6%	1.11x (?)
DictionaryOfAnyHashableStrings_lookup	4056	3720	-8.3%	1.09x (?)

Code size: -Osize

Performance (x86_64): -Onone

Regression	OLD	NEW	DELTA	RATIO
ArrayAppendRepeatCol	324090	373940	+15.4%	0.87x (?)
Data.append.Sequence.809B.Count.RE	32971	37621	+14.1%	0.88x (?)
Data.append.Sequence.64kB.Count.RE	26531	30263	+14.1%	0.88x (?)
Data.init.Sequence.64kB.Count.RE.I	26659	30359	+13.9%	0.88x (?)
Data.append.Sequence.64kB.Count.RE.I	26647	30298	+13.7%	0.88x (?)
Data.append.Sequence.809B.Count.RE.I	33310	37682	+13.1%	0.88x (?)
DataAppendSequence	`3330200`	3766800	+13.1%	0.88x (?)
Data.init.Sequence.809B.Count.RE	33143	37390	+12.8%	0.89x (?)
Data.init.Sequence.64kB.Count.RE	26753	30122	+12.6%	0.89x (?)
Data.init.Sequence.809B.Count.RE.I	33172	37266	+12.3%	0.89x (?)

Improvement	OLD	NEW	DELTA	RATIO
Set.isDisjoint.Smaller.Int0	936	322	-65.6%	2.91x
Set.isDisjoint.Smaller.Box0	3059	1305	-57.3%	2.34x
Set.isDisjoint.Box25	2228	1111	-50.1%	2.01x
Set.isDisjoint.Int.Empty	661	347	-47.5%	1.90x
Set.isDisjoint.Box.Empty	692	376	-45.7%	1.84x
Set.isDisjoint.Int25	1009	618	-38.8%	1.63x
Set.isDisjoint.Int50	973	618	-36.5%	1.57x
Set.isDisjoint.Empty.Int	528	356	-32.6%	1.48x
Set.isDisjoint.Empty.Box	548	377	-31.2%	1.45x
Set.isDisjoint.Int100	809	618	-23.6%	1.31x
Set.isDisjoint.Int0	732	652	-10.9%	1.12x (?)
SIMDReduce.Int8x16.Initializer	609589	555485	-8.9%	1.10x (?)
Dictionary2	1495	1380	-7.7%	1.08x (?)
String.data.Empty	73	68	-6.8%	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: 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

LucianoPAlmeida · 2021-09-15T00:05:02Z

Improvement | OLD | NEW | DELTA | RATIO
Set.isDisjoint.Smaller.Box0 | 758 | 284 | -62.5% | 2.67x
Set.isDisjoint.Smaller.Int0 | 362 | 149 | -58.8% | 2.43x
Set.isDisjoint.Box.Empty | 147 | 80 | -45.6% | 1.84x
Set.isDisjoint.Int.Empty | 126 | 74 | -41.3% | 1.70x
DictionaryOfAnyHashableStrings_lookup | 4032 | 3720 | -7.7% | 1.08x (?)

Now the new benchmarks can show the improvements =]

lorentey

Good work!

(I know this has been a long process; but if you have a bit of extra time, it might be worth taking a very quick look at the code generated for, say, Set<Int>.isDisjoint(with:) -- just enough to make sure it looks sensible. Sometimes there can be unpleasant surprises. 🙈

FWIW, it feels like this version of isDisjoint expects less of the compiler than the version in #21300, but you never know...)

lorentey · 2021-09-15T04:15:25Z

@swift-ci test

swift-ci · 2021-09-15T05:16:31Z

Build failed
Swift Test Linux Platform
Git Sha - 00cb1c5

LucianoPAlmeida · 2021-09-15T09:38:07Z

@swift-ci Please test Linux Platform

LucianoPAlmeida · 2021-09-15T09:38:23Z

@swift-ci Please test Windows Platform

LucianoPAlmeida · 2021-09-15T11:16:07Z

Yeah, by looking into this compiler explorer(link here) is interesting to see that the version in the previous

# previous
output.f1(s1: Swift.Set<Swift.Int>, s2: Swift.Set<Swift.Int>) -> Swift.Bool:
        mov     rax, rdi
        mov     rcx, qword ptr [rdi + 16]
        cmp     rcx, qword ptr [rsi + 16]
        jae     .LBB6_2
        mov     rdi, rax
        jmp     (generic specialization <Swift.Int, Swift.Set<Swift.Int>> of (extension in output):Swift.Set._isDisjoint1<A where A == A1.Element, A1: Swift.Sequence>(with: A1) -> Swift.Bool)
.LBB6_2:
        mov     rdi, rsi
        mov     rsi, rax
        jmp     (generic specialization <Swift.Int, Swift.Set<Swift.Int>> of (extension in output):Swift.Set._isDisjoint1<A where A == A1.Element, A1: Swift.Sequence>(with: A1) -> Swift.Bool)

# this
output.f2(s1: Swift.Set<Swift.Int>, s2: Swift.Set<Swift.Int>) -> Swift.Bool:
        mov     rax, rdi
        mov     rdi, rsi
        mov     rsi, rax
        jmp     (generic specialization <Swift.Int> of (extension in output):Swift.Set.isDisjoint2(with: Swift.Set<A>) -> Swift.Bool)

the difference is basically that for the previous version two branches calling the _isDisjoint function, which is fine but the problem would be if the optimizer decides to inline them, then this would increase code size. I experimented with that by making _isDisjoint a @inline(__always) to see if the optimizer would make it (link here) and we can see indeed how the _isDisjoint call is inlined on the two branches.

So I think it would be fine to just go with this version :)

LucianoPAlmeida requested a review from lorentey September 11, 2021 17:47

lorentey reviewed Sep 11, 2021

View reviewed changes

stdlib/public/core/Set.swift Outdated Show resolved Hide resolved

LucianoPAlmeida force-pushed the set-improvements branch from 3a75f4d to a65fd67 Compare September 11, 2021 22:13

LucianoPAlmeida requested a review from lorentey September 11, 2021 22:45

LucianoPAlmeida mentioned this pull request Sep 12, 2021

[Benchmarks][stdlib] Adding an extra benchmark for set isDisjoint for disjoint sets of different size #39265

Merged

LucianoPAlmeida force-pushed the set-improvements branch from a65fd67 to 9a7bb1a Compare September 14, 2021 01:14

[stdlib] Improve set isDisjoint by iterating on smaller set

00cb1c5

LucianoPAlmeida force-pushed the set-improvements branch from 9a7bb1a to 00cb1c5 Compare September 14, 2021 20:29

lorentey approved these changes Sep 15, 2021

View reviewed changes

LucianoPAlmeida merged commit 50a1817 into swiftlang:main Sep 15, 2021

LucianoPAlmeida deleted the set-improvements branch September 15, 2021 12:13

[stdlib] Improve set isDisjoint by iterating on smaller set #39263

[stdlib] Improve set isDisjoint by iterating on smaller set #39263

Uh oh!

Conversation

LucianoPAlmeida commented Sep 11, 2021

Uh oh!

LucianoPAlmeida commented Sep 11, 2021

Uh oh!

swift-ci commented Sep 11, 2021

Performance (x86_64): -O

Code size: -O

Performance (x86_64): -Osize

Code size: -Osize

Performance (x86_64): -Onone

Code size: -swiftlibs

Uh oh!

LucianoPAlmeida commented Sep 11, 2021

Uh oh!

swift-ci commented Sep 11, 2021

Performance (x86_64): -O

Code size: -O

Performance (x86_64): -Osize

Code size: -Osize

Performance (x86_64): -Onone

Code size: -swiftlibs

Uh oh!

lorentey left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

LucianoPAlmeida commented Sep 11, 2021

Uh oh!

swift-ci commented Sep 11, 2021

Performance (x86_64): -O

Code size: -O

Performance (x86_64): -Osize

Code size: -Osize

Performance (x86_64): -Onone

Code size: -swiftlibs

Uh oh!

LucianoPAlmeida commented Sep 14, 2021

Uh oh!

swift-ci commented Sep 14, 2021

Performance (x86_64): -O

Code size: -O

Performance (x86_64): -Osize

Code size: -Osize

Performance (x86_64): -Onone

Code size: -swiftlibs

Uh oh!

LucianoPAlmeida commented Sep 15, 2021

Uh oh!

lorentey left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

lorentey commented Sep 15, 2021

Uh oh!

swift-ci commented Sep 15, 2021

Uh oh!

LucianoPAlmeida commented Sep 15, 2021

Uh oh!

LucianoPAlmeida commented Sep 15, 2021

Uh oh!

LucianoPAlmeida commented Sep 15, 2021 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

lorentey left a comment •

edited

Loading

LucianoPAlmeida commented Sep 15, 2021 •

edited

Loading