[Webkit-unassigned] [Bug 66363] The executable allocator makes it difficult to free individual chunks of executable memory
bugzilla-daemon at webkit.org
bugzilla-daemon at webkit.org
Tue Aug 16 23:48:28 PDT 2011
https://bugs.webkit.org/show_bug.cgi?id=66363
--- Comment #2 from Filip Pizlo <fpizlo at apple.com> 2011-08-16 23:48:28 PST ---
Current performance impact of a best-fit allocator:
Benchmark report for SunSpider, V8, and Kraken.
VMs tested:
"TipOfTree" at /Volumes/Data/pizlo/quinary/OpenSource/WebKitBuild/Release/jsc
"BestFitExecAlloc" at /Volumes/Data/pizlo/senary/OpenSource/WebKitBuild/Release/jsc
Collected 45 samples per benchmark/VM, with 15 VM invocations per benchmark. 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 BestFitExecAlloc
SunSpider:
3d-cube 7.5848+-0.0895 ? 7.6980+-0.0913 ? might be 1.0149x slower
3d-morph 7.4992+-0.0902 7.4299+-0.0678
3d-raytrace 7.5167+-0.0967 ? 7.6565+-0.0865 ? might be 1.0186x slower
access-binary-trees 2.1908+-0.0406 2.1906+-0.0265
access-fannkuch 11.5275+-0.1206 11.4809+-0.1055
access-nbody 4.2683+-0.0597 4.2403+-0.0361
access-nsieve 2.4786+-0.0511 2.4438+-0.0305 might be 1.0142x faster
bitops-3bit-bits-in-byte 1.6919+-0.0255 ! 1.7734+-0.0259 ! definitely 1.0481x slower
bitops-bits-in-byte 4.1634+-0.0680 ^ 4.0591+-0.0289 ^ definitely 1.0257x faster
bitops-bitwise-and 3.5814+-0.0290 ^ 3.4740+-0.0246 ^ definitely 1.0309x faster
bitops-nsieve-bits 5.3693+-0.0350 5.3506+-0.0326
controlflow-recursive 2.0457+-0.0223 ^ 1.9956+-0.0266 ^ definitely 1.0251x faster
crypto-aes 6.5512+-0.0948 6.4834+-0.0812 might be 1.0105x faster
crypto-md5 2.7238+-0.0306 2.7201+-0.0311
crypto-sha1 2.1835+-0.0309 ! 2.2627+-0.0397 ! definitely 1.0363x slower
date-format-tofte 9.7838+-0.1275 ? 9.9886+-0.1203 ? might be 1.0209x slower
date-format-xparb 8.5392+-0.1422 ? 8.7528+-0.1075 ? might be 1.0250x slower
math-cordic 6.2146+-0.0660 6.1621+-0.0641
math-partial-sums 7.5369+-0.0822 7.5212+-0.0997
math-spectral-norm 2.4702+-0.0358 ? 2.4720+-0.0232 ?
regexp-dna 10.2566+-0.0865 ! 10.7704+-0.0813 ! definitely 1.0501x slower
string-base64 6.0445+-0.1184 5.9681+-0.0809 might be 1.0128x faster
string-fasta 7.2674+-0.0775 ? 7.2963+-0.0770 ?
string-tagcloud 13.3241+-0.1118 13.3222+-0.1283
string-unpack-code 18.3815+-0.1928 ? 18.5709+-0.1796 ? might be 1.0103x slower
string-validate-input 6.9840+-0.1006 ? 6.9874+-0.1120 ?
<arithmetic> 6.4684+-0.0138 ! 6.5027+-0.0165 ! definitely 1.0053x slower
<geometric> 5.3575+-0.0113 ? 5.3737+-0.0134 ?
<harmonic> 4.3735+-0.0149 ? 4.3878+-0.0149 ?
TipOfTree BestFitExecAlloc
V8:
crypto 87.9887+-0.2026 ^ 87.4986+-0.2128 ^ definitely 1.0056x faster
deltablue 261.7402+-0.7046 ! 264.2305+-0.9393 ! definitely 1.0095x slower
earley-boyer 99.9972+-0.2294 99.6741+-0.1972
raytrace 76.4790+-0.3443 ^ 75.8491+-0.1981 ^ definitely 1.0083x faster
regexp 107.5038+-0.2056 ^ 105.7876+-0.1971 ^ definitely 1.0162x faster
richards 248.0936+-0.3812 ^ 238.6467+-0.3379 ^ definitely 1.0396x faster
splay 108.0297+-0.3112 107.4892+-0.3354
<arithmetic> 141.4046+-0.1320 ^ 139.8823+-0.1787 ^ definitely 1.0109x faster
<geometric> 126.1123+-0.1141 ^ 124.9021+-0.1382 ^ definitely 1.0097x faster
<harmonic> 115.0262+-0.1278 ^ 114.0441+-0.1267 ^ definitely 1.0086x faster
TipOfTree BestFitExecAlloc
Kraken:
ai-astar 1081.4993+-3.1740 ? 1083.8513+-3.7182 ?
audio-beat-detection 453.1965+-1.4023 ^ 447.1013+-0.6723 ^ definitely 1.0136x faster
audio-dft 411.1923+-3.7095 ? 412.7745+-2.2890 ?
audio-fft 356.4072+-1.5136 354.6045+-0.9395
audio-oscillator 378.4706+-0.4562 ! 380.7842+-1.1168 ! definitely 1.0061x slower
imaging-darkroom 590.5360+-2.1499 ^ 565.1859+-1.3840 ^ definitely 1.0449x faster
imaging-desaturate 586.5267+-1.2839 ? 586.9413+-1.2455 ?
imaging-gaussian-blur 1711.9912+-5.0103 ^ 1697.7598+-1.8843 ^ definitely 1.0084x faster
json-parse-financial 48.0579+-0.1204 ^ 47.4060+-0.1959 ^ definitely 1.0138x faster
json-stringify-tinderbox 62.0471+-0.3300 ^ 60.8928+-0.1510 ^ definitely 1.0190x faster
stanford-crypto-aes 140.7944+-0.2706 ! 141.8431+-0.4629 ! definitely 1.0074x slower
stanford-crypto-ccm 110.3081+-0.5687 110.2512+-0.3872
stanford-crypto-pbkdf2 367.4630+-1.4301 365.7738+-2.9701
stanford-crypto-sha256-iterative 138.3388+-0.4035 ? 138.6526+-0.2152 ?
<arithmetic> 459.7735+-0.5107 ^ 456.7016+-0.4367 ^ definitely 1.0067x faster
<geometric> 295.0588+-0.3691 ^ 293.2588+-0.3174 ^ definitely 1.0061x faster
<harmonic> 179.3775+-0.3080 ^ 177.9309+-0.2398 ^ definitely 1.0081x faster
TipOfTree BestFitExecAlloc
All benchmarks:
<arithmetic> 161.5923+-0.1607 ^ 160.4695+-0.1357 ^ definitely 1.0070x faster
<geometric> 28.3039+-0.0368 28.2591+-0.0404
<harmonic> 7.7255+-0.0258 ? 7.7486+-0.0256 ?
--
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.
More information about the webkit-unassigned
mailing list