WebKit Bugzilla
New
Browse
Log In
×
Sign in with GitHub
or
Remember my login
Create Account
·
Forgot Password
Forgotten password account recovery
RESOLVED FIXED
74390
DFG GetByVal CSE rule should match PutByValAlias
https://bugs.webkit.org/show_bug.cgi?id=74390
Summary
DFG GetByVal CSE rule should match PutByValAlias
Filip Pizlo
Reported
2011-12-13 01:32:11 PST
Patch forthcoming
Attachments
the patch
(1.29 KB, patch)
2011-12-13 01:35 PST
,
Filip Pizlo
ggaren
: review+
Details
Formatted Diff
Diff
View All
Add attachment
proposed patch, testcase, etc.
Filip Pizlo
Comment 1
2011-12-13 01:35:14 PST
Created
attachment 118977
[details]
the patch Benchmark report for SunSpider, V8, and Kraken on bigmac.local (MacPro5,1). VMs tested: "TipOfTree" at /Volumes/Data/pizlo/quinary/OpenSource/WebKitBuild/Release/jsc (
r102562
) "PutByValAliasCSE" at /Volumes/Data/pizlo/secondary/OpenSource/WebKitBuild/Release/jsc (
r102661
) Collected 45 samples per benchmark/VM, with 15 VM invocations per benchmark. Emitted a call to gc() between sample measurements. Used 1 benchmark iteration per VM invocation for warm-up. Used the jsc-specific preciseTime() function to get microsecond-level timing. Reporting benchmark execution times with 95% confidence intervals in milliseconds. TipOfTree PutByValAliasCSE SunSpider: 3d-cube 7.3350+-0.0104 7.3167+-0.0181 3d-morph 8.3525+-0.0426 ? 8.3698+-0.0513 ? 3d-raytrace 7.7018+-0.0274 7.6631+-0.0304 access-binary-trees 1.6007+-0.0044 1.6007+-0.0030 access-fannkuch 7.5335+-0.0071 ^ 7.3790+-0.0062 ^ definitely 1.0209x faster access-nbody 3.9326+-0.0040 3.9278+-0.0045 access-nsieve 3.1982+-0.0225 3.1941+-0.0222 bitops-3bit-bits-in-byte 1.2578+-0.0079 1.2503+-0.0075 bitops-bits-in-byte 5.0874+-0.0159 5.0637+-0.0149 bitops-bitwise-and 3.2892+-0.0039 ? 3.2920+-0.0050 ? bitops-nsieve-bits 5.6519+-0.0167 5.6448+-0.0178 controlflow-recursive 2.3535+-0.0106 2.3443+-0.0085 crypto-aes 7.2043+-0.0200 ! 7.3627+-0.0180 ! definitely 1.0220x slower crypto-md5 2.4642+-0.0101 2.4617+-0.0104 crypto-sha1 2.1909+-0.0147 2.1791+-0.0143 date-format-tofte 10.6563+-0.0566 ? 10.7304+-0.0707 ? date-format-xparb 9.8916+-0.0487 9.7766+-0.0888 might be 1.0118x faster math-cordic 7.1762+-0.0174 ? 7.1842+-0.0174 ? math-partial-sums 10.4064+-0.0120 ? 10.4218+-0.0157 ? math-spectral-norm 2.6019+-0.0051 ? 2.6055+-0.0064 ? regexp-dna 8.8459+-0.0253 ? 8.8789+-0.0257 ? string-base64 4.2221+-0.0069 4.2181+-0.0071 string-fasta 7.1979+-0.0079 ? 7.2052+-0.0093 ? string-tagcloud 12.3993+-0.0296 12.3754+-0.0338 string-unpack-code 20.9027+-0.0539 ^ 20.7012+-0.0592 ^ definitely 1.0097x faster string-validate-input 5.6090+-0.0197 5.5907+-0.0230 <arithmetic> * 6.5024+-0.0081 6.4899+-0.0095 might be 1.0019x faster <geometric> 5.2827+-0.0075 5.2745+-0.0075 might be 1.0016x faster <harmonic> 4.1800+-0.0078 4.1719+-0.0074 might be 1.0019x faster TipOfTree PutByValAliasCSE V8: crypto 76.1163+-0.1000 ? 76.2169+-0.1372 ? deltablue 172.8269+-0.3326 ? 173.1246+-0.3492 ? earley-boyer 99.2811+-0.6146 ? 99.3397+-0.5229 ? raytrace 52.2787+-0.1007 ! 52.7063+-0.2284 ! definitely 1.0082x slower regexp 122.5528+-0.2666 122.4658+-0.2735 richards 139.9164+-0.3003 ^ 139.3020+-0.2854 ^ definitely 1.0044x faster splay 92.1539+-0.6990 91.2885+-0.6590 <arithmetic> 107.8752+-0.1441 107.7777+-0.1476 might be 1.0009x faster <geometric> * 101.0833+-0.1448 101.0436+-0.1663 might be 1.0004x faster <harmonic> 94.2205+-0.1363 ? 94.2814+-0.1864 ? might be 1.0006x slower TipOfTree PutByValAliasCSE Kraken: ai-astar 816.5833+-5.1826 ? 821.5976+-3.8664 ? audio-beat-detection 203.2783+-0.2957 ! 204.6801+-0.4901 ! definitely 1.0069x slower audio-dft 276.5288+-1.0650 275.3805+-1.0953 audio-fft 131.9214+-0.2135 ! 133.1070+-0.3343 ! definitely 1.0090x slower audio-oscillator 286.0799+-2.0936 284.7149+-1.9930 imaging-darkroom 341.0025+-3.3151 336.3170+-2.4769 might be 1.0139x faster imaging-desaturate 237.5525+-0.0434 237.4708+-0.0383 imaging-gaussian-blur 625.8866+-0.1034 ? 626.0054+-0.1557 ? json-parse-financial 69.7659+-0.1073 ? 69.9326+-0.0694 ? json-stringify-tinderbox 82.0705+-0.1258 ! 82.4643+-0.1323 ! definitely 1.0048x slower stanford-crypto-aes 117.6011+-0.3017 ! 118.7071+-0.6210 ! definitely 1.0094x slower stanford-crypto-ccm 113.7530+-0.2544 ? 114.0518+-0.3293 ? stanford-crypto-pbkdf2 229.8478+-0.5139 ? 230.3339+-0.5859 ? stanford-crypto-sha256-iterative 95.8354+-0.1263 95.5967+-0.1244 <arithmetic> * 259.1219+-0.3887 ? 259.3114+-0.3453 ? might be 1.0007x slower <geometric> 198.5541+-0.1832 ? 198.8088+-0.1810 ? might be 1.0013x slower <harmonic> 158.7630+-0.1260 ! 159.1449+-0.1166 ! definitely 1.0024x slower TipOfTree PutByValAliasCSE All benchmarks: <arithmetic> 96.8488+-0.1178 ? 96.8839+-0.1150 ? might be 1.0004x slower <geometric> 24.1502+-0.0247 24.1372+-0.0252 might be 1.0005x faster <harmonic> 7.3638+-0.0135 7.3502+-0.0128 might be 1.0019x faster TipOfTree PutByValAliasCSE Geomean of preferred means: <scaled-result> 55.4307+-0.0514 55.4014+-0.0609 might be 1.0005x faster
Geoffrey Garen
Comment 2
2011-12-13 13:30:04 PST
Comment on
attachment 118977
[details]
the patch r=me
Filip Pizlo
Comment 3
2011-12-13 14:12:56 PST
Landed in
http://trac.webkit.org/changeset/102707
Note
You need to
log in
before you can comment on or make changes to this bug.
Top of Page
Format For Printing
XML
Clone This Bug