[Webkit-unassigned] [Bug 75342] Adjust spill order in DFG
bugzilla-daemon at webkit.org
bugzilla-daemon at webkit.org
Thu Dec 29 04:59:52 PST 2011
https://bugs.webkit.org/show_bug.cgi?id=75342
--- Comment #3 from Yuqiang Xian <yuqiang.xian at intel.com> 2011-12-29 04:59:52 PST ---
Performance result on 64bit (Mac OS): Overall almost neutral while a obvious slowdown (8%) on imaging-gaussian-blur is observed (on 32bit there's a 11% improvement for this case), which may be further investigated in the future.
VMs tested:
"ToT" at /webkit/WebKitBuild-orig/Release/jsc
"SpillOrder" at /webkit/WebKitBuild-64/Release/jsc
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 6.5219+-0.0772 ? 6.5384+-0.0815 ?
3d-morph 8.0982+-0.0638 ? 8.1566+-0.0522 ?
3d-raytrace 8.3841+-0.8049 7.8268+-0.1195 might be 1.0712x faster
access-binary-trees 1.7863+-0.0935 1.7486+-0.0512 might be 1.0215x faster
access-fannkuch 7.3589+-0.0960 7.3325+-0.0762
access-nbody 3.8475+-0.0496 ? 3.8839+-0.0597 ?
access-nsieve 3.1162+-0.0518 ? 3.1720+-0.0791 ? might be 1.0179x slower
bitops-3bit-bits-in-byte 1.2023+-0.0080 ? 1.2062+-0.0125 ?
bitops-bits-in-byte 4.8450+-0.0564 ? 4.8767+-0.0639 ?
bitops-bitwise-and 3.1650+-0.0216 3.1500+-0.0052
bitops-nsieve-bits 5.4344+-0.0415 ? 5.4718+-0.0413 ?
controlflow-recursive 2.2807+-0.0339 2.2514+-0.0101 might be 1.0130x faster
crypto-aes 8.5308+-0.1246 8.4498+-0.1046
crypto-md5 2.4437+-0.0741 2.4285+-0.0325
crypto-sha1 2.2442+-0.0703 ? 2.2763+-0.1707 ? might be 1.0143x slower
date-format-tofte 11.0530+-0.0630 ? 11.1376+-0.0764 ?
date-format-xparb 9.5936+-0.0409 ^ 9.4306+-0.0888 ^ definitely 1.0173x faster
math-cordic 7.0268+-0.0950 6.9927+-0.0693
math-partial-sums 10.3018+-0.0443 10.2964+-0.0644
math-spectral-norm 2.5789+-0.0124 2.5767+-0.0079
regexp-dna 9.0382+-0.1357 ? 9.1914+-0.1509 ? might be 1.0169x slower
string-base64 4.6544+-0.0123 ? 4.6783+-0.0162 ?
string-fasta 7.8880+-0.1092 ? 7.8938+-0.1069 ?
string-tagcloud 12.8570+-0.1142 ? 12.9316+-0.0944 ?
string-unpack-code 21.8200+-0.2621 21.6093+-0.1502
string-validate-input 6.0590+-0.0585 ? 6.0633+-0.2153 ?
<arithmetic> * 6.6204+-0.0354 6.5989+-0.0214 might be 1.0033x faster
<geometric> 5.3390+-0.0330 5.3287+-0.0287 might be 1.0019x faster
<harmonic> 4.1997+-0.0354 4.1934+-0.0359 might be 1.0015x faster
ToT SpillOrder
V8:
crypto 75.4973+-0.2389 75.0776+-0.3059
deltablue 161.0954+-0.4027 ! 163.9173+-0.4714 ! definitely 1.0175x slower
earley-boyer 99.6115+-1.5928 ? 100.1018+-1.4401 ?
raytrace 50.3873+-0.2048 ^ 49.8641+-0.0921 ^ definitely 1.0105x faster
regexp 121.3851+-0.1344 ! 122.4113+-0.3352 ! definitely 1.0085x slower
richards 135.4591+-0.2543 134.3140+-1.1460
splay 97.6467+-0.9025 96.1546+-1.4374 might be 1.0155x faster
<arithmetic> 105.8689+-0.2327 ? 105.9773+-0.2850 ? might be 1.0010x slower
<geometric> * 99.7080+-0.2394 99.5747+-0.2911 might be 1.0013x faster
<harmonic> 93.0755+-0.2284 92.7282+-0.2595 might be 1.0037x faster
ToT SpillOrder
Kraken:
ai-astar 837.8626+-49.6049 796.1270+-0.4181 might be 1.0524x faster
audio-beat-detection 186.1555+-0.9161 ? 187.6734+-1.2043 ?
audio-dft 446.1301+-1.5645 ? 453.0984+-8.1987 ? might be 1.0156x slower
audio-fft 122.8065+-7.3403 118.5203+-3.1245 might be 1.0362x faster
audio-oscillator 256.0711+-1.2247 255.9696+-1.2486
imaging-darkroom 290.8491+-5.3218 ? 291.6089+-5.3305 ?
imaging-desaturate 220.6728+-0.0440 ? 220.7556+-0.0591 ?
imaging-gaussian-blur 497.9859+-0.9100 ! 540.6406+-0.0741 ! definitely 1.0857x slower
json-parse-financial 62.6418+-1.6843 61.9968+-0.0362 might be 1.0104x faster
json-stringify-tinderbox 75.5961+-0.1585 75.2695+-0.1742
stanford-crypto-aes 111.9879+-0.3878 111.9130+-0.2919
stanford-crypto-ccm 109.2299+-0.8334 ? 109.3411+-0.8627 ?
stanford-crypto-pbkdf2 213.3721+-0.1806 ! 214.8904+-0.2552 ! definitely 1.0071x slower
stanford-crypto-sha256-iterative 104.4605+-0.1416 ? 104.5802+-0.1995 ?
<arithmetic> * 252.5587+-3.6387 ? 253.0275+-0.8441 ? might be 1.0019x slower
<geometric> 191.5593+-1.3249 ? 191.8617+-0.5199 ? might be 1.0016x slower
<harmonic> 151.4256+-1.0019 151.0671+-0.3450 might be 1.0024x faster
ToT SpillOrder
All benchmarks:
<arithmetic> 94.6603+-1.1013 ? 94.8042+-0.2741 ? might be 1.0015x slower
<geometric> 23.9847+-0.1147 23.9658+-0.0879 might be 1.0008x faster
<harmonic> 7.3915+-0.0609 7.3801+-0.0618 might be 1.0015x faster
ToT SpillOrder
Geomean of preferred means:
<scaled-result> 55.0351+-0.3552 54.9871+-0.1370 might be 1.0009x 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