154572 – [JSC] Improve Select of Doubles based on Double condition

RESOLVED FIXED Bug 154572

[JSC] Improve Select of Doubles based on Double condition

https://bugs.webkit.org/show_bug.cgi?id=154572

Summary [JSC] Improve Select of Doubles based on Double condition

Benjamin Poulain

Reported 2016-02-22 18:35:58 PST

[JSC] Improve Select of Doubles based on Double condition

Attachments
Patch (53.89 KB, patch) 2016-02-22 18:44 PST, Benjamin Poulain	no flags	Details Formatted Diff Diff
View All Add attachment proposed patch, testcase, etc.

Benjamin Poulain

Comment 1 2016-02-22 18:44:15 PST

Created attachment 271979 [details] Patch

Benjamin Poulain

Comment 2 2016-02-22 18:44:36 PST

ARM64 results: Conf#1 Conf#2 SunSpider: 3d-cube 10.4678+-0.1776 10.4234+-0.0884 3d-morph 8.2765+-0.0665 ? 8.4501+-0.1354 ? might be 1.0210x slower 3d-raytrace 9.6306+-0.4978 ? 10.1917+-0.4832 ? might be 1.0583x slower access-binary-trees 4.3589+-0.0500 4.2659+-0.1246 might be 1.0218x faster access-fannkuch 11.6872+-0.9625 ? 11.7824+-1.1008 ? access-nbody 4.8167+-0.1337 4.7069+-0.0507 might be 1.0233x faster access-nsieve 3.5576+-0.2040 3.5423+-0.0670 bitops-3bit-bits-in-byte 1.8202+-0.1677 ? 1.8905+-0.2034 ? might be 1.0386x slower bitops-bits-in-byte 4.3246+-0.1038 4.3028+-0.0341 bitops-bitwise-and 3.2191+-0.0285 ? 3.2289+-0.0516 ? bitops-nsieve-bits 5.8892+-0.0526 ? 5.9472+-0.0683 ? controlflow-recursive 3.8726+-0.3588 3.5500+-0.3637 might be 1.0909x faster crypto-aes 6.5471+-0.1118 ? 6.5686+-0.1597 ? crypto-md5 4.3004+-0.3622 4.1703+-0.2402 might be 1.0312x faster crypto-sha1 3.8150+-0.1209 3.7923+-0.0568 date-format-tofte 11.1752+-0.1305 11.1364+-0.1037 date-format-xparb 7.5310+-0.4663 7.4216+-0.1289 might be 1.0147x faster math-cordic 5.4047+-0.0895 5.3905+-0.0984 math-partial-sums 11.6647+-0.1398 ? 11.7150+-0.1698 ? math-spectral-norm 3.2741+-0.2327 ? 3.4810+-0.2925 ? might be 1.0632x slower regexp-dna 9.7617+-0.0705 ? 9.7723+-0.1063 ? string-base64 6.5264+-0.0964 6.4979+-0.0381 string-fasta 9.8154+-0.1679 ? 9.9933+-0.1529 ? might be 1.0181x slower string-tagcloud 11.1328+-0.1967 11.0198+-0.1089 might be 1.0103x faster string-unpack-code 23.7781+-0.1873 23.6078+-0.1298 string-validate-input 6.6788+-0.1459 6.4669+-0.1227 might be 1.0328x faster <arithmetic> 7.4356+-0.0563 7.4352+-0.0629 might be 1.0001x faster Conf#1 Conf#2 Octane: encrypt 0.19860+-0.00155 ? 0.19905+-0.00139 ? decrypt 3.71906+-0.00576 ? 3.72646+-0.02533 ? deltablue x2 0.17446+-0.00461 0.17203+-0.00338 might be 1.0141x faster earley 0.42984+-0.01644 0.41700+-0.00675 might be 1.0308x faster boyer 7.96326+-0.37931 7.65866+-0.18503 might be 1.0398x faster navier-stokes x2 7.18257+-0.00512 ? 7.22201+-0.06471 ? raytrace x2 1.29717+-0.01003 1.28950+-0.00698 richards x2 0.11467+-0.00100 0.11424+-0.00093 splay x2 0.74971+-0.01588 ^ 0.72347+-0.00926 ^ definitely 1.0363x faster regexp x2 36.87020+-0.23798 36.68169+-0.09856 pdfjs x2 56.57145+-0.63483 ? 56.76323+-0.61822 ? mandreel x2 72.11990+-0.76653 71.65942+-1.14959 gbemu x2 45.95830+-3.68361 45.76195+-4.04929 closure 0.65412+-0.00613 ? 0.66838+-0.02829 ? might be 1.0218x slower jquery 9.68256+-0.04040 ? 9.68319+-0.01521 ? box2d x2 16.84773+-1.92125 15.97744+-0.13870 might be 1.0545x faster zlib x2 679.09902+-8.46445 ? 684.17057+-2.34361 ? typescript x2 1142.23555+-12.20514 ? 1150.15195+-6.53485 ? <geometric> 8.18626+-0.05766 8.12227+-0.04894 might be 1.0079x faster Conf#1 Conf#2 Kraken: ai-astar 155.956+-0.867 155.788+-0.440 audio-beat-detection 58.734+-0.314 ? 58.992+-1.435 ? audio-dft 147.326+-5.702 142.967+-5.213 might be 1.0305x faster audio-fft 41.914+-0.093 41.891+-0.106 audio-oscillator 58.072+-0.486 ? 58.177+-0.838 ? imaging-darkroom 69.077+-0.105 ? 69.194+-0.120 ? imaging-desaturate 78.536+-0.889 ? 78.565+-0.645 ? imaging-gaussian-blur 106.425+-1.281 ? 107.502+-0.876 ? might be 1.0101x slower json-parse-financial 48.121+-0.129 ? 48.355+-0.798 ? json-stringify-tinderbox 27.058+-0.410 ? 27.198+-0.250 ? stanford-crypto-aes 58.242+-0.182 ! 58.820+-0.387 ! definitely 1.0099x slower stanford-crypto-ccm 47.393+-1.320 46.965+-0.834 stanford-crypto-pbkdf2 139.804+-0.623 ! 141.403+-0.799 ! definitely 1.0114x slower stanford-crypto-sha256-iterative 51.459+-0.142 ? 51.518+-0.220 ? <arithmetic> 77.723+-0.467 77.667+-0.399 might be 1.0007x faster Conf#1 Conf#2 AsmBench: bigfib.cpp 673.9739+-3.9344 672.6627+-29.9566 cray.c 561.9127+-1.1886 ? 561.9176+-0.7776 ? dry.c 473.7460+-3.9622 473.5770+-4.6664 FloatMM.c 793.6519+-1.1961 792.6264+-0.3668 gcc-loops.cpp 4599.4011+-4.4631 ? 4608.1185+-19.9831 ? n-body.c 1773.3995+-2.5473 1749.8401+-40.9771 might be 1.0135x faster Quicksort.c 569.5708+-3.3983 567.4153+-3.5734 stepanov_container.cpp 5640.3250+-80.4516 5602.4305+-15.8436 Towers.c 282.4936+-0.7417 ? 282.8061+-0.6463 ? <geometric> 1005.8344+-2.2890 1002.8776+-4.2186 might be 1.0029x faster Conf#1 Conf#2 Geomean of preferred means: <scaled-result> 46.7043+-0.1008 46.5694+-0.0792 might be 1.0029x faster

WebKit Commit Bot

Comment 3 2016-02-22 18:46:35 PST

Attachment 271979 [details] did not pass style-queue: ERROR: Source/JavaScriptCore/b3/testb3.cpp:9691: More than one command on the same line [whitespace/newline] [4] ERROR: Source/JavaScriptCore/b3/testb3.cpp:9729: More than one command on the same line [whitespace/newline] [4] ERROR: Source/JavaScriptCore/b3/testb3.cpp:10021: More than one command on the same line [whitespace/newline] [4] ERROR: Source/JavaScriptCore/b3/testb3.cpp:10022: More than one command on the same line [whitespace/newline] [4] ERROR: Source/JavaScriptCore/b3/testb3.cpp:10023: More than one command on the same line [whitespace/newline] [4] ERROR: Source/JavaScriptCore/b3/testb3.cpp:10024: More than one command on the same line [whitespace/newline] [4] ERROR: Source/JavaScriptCore/b3/testb3.cpp:10025: More than one command on the same line [whitespace/newline] [4] ERROR: Source/JavaScriptCore/b3/testb3.cpp:10026: More than one command on the same line [whitespace/newline] [4] ERROR: Source/JavaScriptCore/b3/testb3.cpp:10027: More than one command on the same line [whitespace/newline] [4] ERROR: Source/JavaScriptCore/b3/testb3.cpp:10243: More than one command on the same line [whitespace/newline] [4] ERROR: Source/JavaScriptCore/b3/testb3.cpp:10244: More than one command on the same line [whitespace/newline] [4] ERROR: Source/JavaScriptCore/b3/testb3.cpp:10245: More than one command on the same line [whitespace/newline] [4] ERROR: Source/JavaScriptCore/b3/testb3.cpp:10246: More than one command on the same line [whitespace/newline] [4] ERROR: Source/JavaScriptCore/b3/testb3.cpp:10247: More than one command on the same line [whitespace/newline] [4] ERROR: Source/JavaScriptCore/b3/testb3.cpp:10248: More than one command on the same line [whitespace/newline] [4] ERROR: Source/JavaScriptCore/b3/testb3.cpp:10249: More than one command on the same line [whitespace/newline] [4] Total errors found: 16 in 7 files If any of these errors are false positives, please file a bug against check-webkit-style.

WebKit Commit Bot

Comment 4 2016-03-02 21:18:54 PST

Comment on attachment 271979 [details] Patch Clearing flags on attachment: 271979 Committed r197490: <http://trac.webkit.org/changeset/197490>

WebKit Commit Bot

Comment 5 2016-03-02 21:18:58 PST

All reviewed patches have been landed. Closing bug.

Note You need to log in before you can comment on or make changes to this bug.

Status RESOLVED

Resolution FIXED

Priority P2

Severity Normal

Classification Unclassified

Version WebKit Nightly Build

Hardware Unspecified

OS Unspecified

Product WebKit

Component New Bugs

Assignee

Benjamin Poulain

Reported

2016-02-22 18:35 PST

Modified

2016-03-02 21:18 PST History

CC List

6 users Show

URL

Keywords

Depends on

Blocks