[Webkit-unassigned] [Bug 75342] Adjust spill order in DFG
bugzilla-daemon at webkit.org
bugzilla-daemon at webkit.org
Thu Dec 29 04:52:23 PST 2011
https://bugs.webkit.org/show_bug.cgi?id=75342
--- Comment #2 from Yuqiang Xian <yuqiang.xian at intel.com> 2011-12-29 04:52:23 PST ---
Performance result on 32bit (Core i7 Nehalem Linux): Some cases get obvious improvement while some get slowdown (needs further investigation). Overall almost neutral.
VMs tested:
"ToT" at /home/yxian/WebKit_orig/WebKitBuild/Release/Source/JavaScriptCore/shell/jsc_efl
"SpillOrder" at /mnt/supplement/WebKit/WebKitBuild/Release_efl/Source/JavaScriptCore/shell/jsc_efl
Collected 12 samples per benchmark/VM, with 4 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.
ToT SpillOrder
SunSpider:
3d-cube 7.1117+-0.0836 ! 7.4183+-0.0941 ! definitely 1.0431x slower
3d-morph 10.6417+-0.0857 10.6191+-0.1000
3d-raytrace 9.5057+-0.0722 ? 9.5482+-0.0750 ?
access-binary-trees 1.8021+-0.0360 ? 1.8418+-0.0472 ? might be 1.0220x slower
access-fannkuch 10.7467+-0.0628 ! 10.8864+-0.0622 ! definitely 1.0130x slower
access-nbody 5.4244+-0.0447 ^ 4.3335+-0.0586 ^ definitely 1.2517x faster
access-nsieve 3.7283+-0.0538 3.6961+-0.0432
bitops-3bit-bits-in-byte 1.1868+-0.0384 1.1696+-0.0299 might be 1.0147x faster
bitops-bits-in-byte 4.6701+-0.0406 ? 4.6898+-0.0489 ?
bitops-bitwise-and 4.2241+-0.0366 4.2229+-0.0515
bitops-nsieve-bits 6.8010+-0.1106 6.7109+-0.1390 might be 1.0134x faster
controlflow-recursive 2.7756+-0.0994 2.7524+-0.0462
crypto-aes 9.8318+-0.1618 9.7819+-0.1195
crypto-md5 3.0158+-0.0298 ? 3.0516+-0.0669 ? might be 1.0118x slower
crypto-sha1 2.6072+-0.0896 2.5776+-0.0408 might be 1.0115x faster
date-format-tofte 11.6659+-0.1436 ? 11.7558+-0.1014 ?
date-format-xparb 11.3672+-0.1453 11.1826+-0.1065 might be 1.0165x faster
math-cordic 9.1636+-0.0433 ? 9.1973+-0.0679 ?
math-partial-sums 13.8785+-0.0619 13.8386+-0.0674
math-spectral-norm 2.5840+-0.0408 ? 2.6044+-0.0486 ?
regexp-dna 9.0942+-0.0692 ? 9.1036+-0.0740 ?
string-base64 5.3639+-0.0457 5.3459+-0.0590
string-fasta 9.4544+-0.0840 ? 9.4748+-0.0962 ?
string-tagcloud 15.3900+-0.0881 ? 15.5524+-0.1076 ? might be 1.0106x slower
string-unpack-code 24.4761+-0.1017 ! 24.7848+-0.2017 ! definitely 1.0126x slower
string-validate-input 7.3702+-0.0667 ? 7.4342+-0.0782 ?
<arithmetic> * 7.8416+-0.0215 7.8298+-0.0235 might be 1.0015x faster
<geometric> 6.2691+-0.0241 6.2310+-0.0176 might be 1.0061x faster
<harmonic> 4.7724+-0.0372 4.7361+-0.0222 might be 1.0077x faster
ToT SpillOrder
V8:
crypto 97.1258+-1.0538 ? 99.0318+-1.2908 ? might be 1.0196x slower
deltablue 167.9481+-0.6349 ? 168.0176+-0.8081 ?
earley-boyer 107.8401+-0.1878 ! 108.4812+-0.4099 ! definitely 1.0059x slower
raytrace 53.0712+-0.4534 52.6155+-0.5008
regexp 129.2446+-1.0419 128.8020+-0.5767
richards 173.4236+-0.3544 ^ 171.5025+-1.5316 ^ definitely 1.0112x faster
splay 125.3138+-0.4902 124.7199+-0.5666
<arithmetic> 121.9953+-0.3170 121.8815+-0.3005 might be 1.0009x faster
<geometric> * 114.7610+-0.3722 114.7226+-0.2784 might be 1.0003x faster
<harmonic> 106.2956+-0.4520 106.2297+-0.3151 might be 1.0006x faster
ToT SpillOrder
Kraken:
ai-astar 785.4906+-1.0815 ? 786.9144+-1.8163 ?
audio-beat-detection 248.5505+-1.2685 ! 256.1680+-0.2308 ! definitely 1.0306x slower
audio-dft 355.2405+-1.9364 ? 358.1129+-2.4869 ?
audio-fft 155.0614+-0.0526 ! 163.8682+-0.6396 ! definitely 1.0568x slower
audio-oscillator 335.0814+-1.3393 ! 341.2124+-1.6898 ! definitely 1.0183x slower
imaging-darkroom 380.4842+-9.7790 ? 382.8186+-10.1730 ?
imaging-desaturate 317.8581+-1.0630 317.4988+-0.4798
imaging-gaussian-blur 596.2950+-1.4765 ^ 537.7586+-0.9815 ^ definitely 1.1089x faster
json-parse-financial 69.5950+-0.4176 ? 70.1253+-0.5553 ?
json-stringify-tinderbox 105.5944+-0.8867 104.1428+-1.3432 might be 1.0139x faster
stanford-crypto-aes 132.9845+-0.3843 ? 133.6331+-0.3833 ?
stanford-crypto-ccm 127.3089+-0.5082 ? 127.8093+-0.4991 ?
stanford-crypto-pbkdf2 281.0584+-0.6893 ! 283.0713+-0.8316 ! definitely 1.0072x slower
stanford-crypto-sha256-iterative 108.2076+-0.2986 108.1264+-0.3654
<arithmetic> * 285.6293+-0.8992 ^ 283.6614+-0.8614 ^ definitely 1.0069x faster
<geometric> 227.2966+-0.6312 ? 227.6839+-0.7165 ? might be 1.0017x slower
<harmonic> 181.7799+-0.4415 ? 182.8114+-0.6295 ? might be 1.0057x slower
ToT SpillOrder
All benchmarks:
<arithmetic> 107.5885+-0.3070 ^ 106.9788+-0.2664 ^ definitely 1.0057x faster
<geometric> 28.1674+-0.0808 28.0854+-0.0606 might be 1.0029x faster
<harmonic> 8.4065+-0.0638 8.3448+-0.0381 might be 1.0074x faster
ToT SpillOrder
Geomean of preferred means:
<scaled-result> 63.5818+-0.1712 63.3964+-0.1339 might be 1.0029x faster
--
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