RESOLVED FIXED 75199
The ArgumentCount field in the CallFrame should have its tag left blank for other uses
https://bugs.webkit.org/show_bug.cgi?id=75199
Summary The ArgumentCount field in the CallFrame should have its tag left blank for o...
Filip Pizlo
Reported 2011-12-24 14:35:21 PST
Patch forthcoming.
Attachments
the patch (12.73 KB, patch)
2011-12-24 14:41 PST, Filip Pizlo
oliver: review+
Filip Pizlo
Comment 1 2011-12-24 14:35:29 PST
This looks neutral. Benchmark report for SunSpider, V8, and Kraken on bigmac (MacPro5,1). VMs tested: "TipOfTree" at /Volumes/Data/pizlo/quinary/OpenSource/WebKitBuild/Release/jsc (r103662) "ArgCountPayload" at /Volumes/Data/pizlo/OpenSource/WebKitBuild/Release/jsc (r103662) 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. TipOfTree ArgCountPayload SunSpider: 3d-cube 5.8577+-0.0457 ! 6.1934+-0.0206 ! definitely 1.0573x slower 3d-morph 8.8313+-0.1031 ^ 8.4641+-0.1097 ^ definitely 1.0434x faster 3d-raytrace 7.6188+-0.0820 7.5398+-0.0692 might be 1.0105x faster access-binary-trees 1.5984+-0.0099 ? 1.6007+-0.0094 ? access-fannkuch 7.5037+-0.0104 ! 7.5936+-0.0556 ! definitely 1.0120x slower access-nbody 3.9462+-0.0071 ? 3.9515+-0.0097 ? access-nsieve 3.1351+-0.0515 3.1251+-0.0432 bitops-3bit-bits-in-byte 1.2467+-0.0142 ? 1.2561+-0.0157 ? bitops-bits-in-byte 5.1008+-0.0512 5.0160+-0.0468 might be 1.0169x faster bitops-bitwise-and 3.2987+-0.0206 3.2855+-0.0047 bitops-nsieve-bits 5.6916+-0.0538 5.6454+-0.0493 controlflow-recursive 2.3411+-0.0128 2.3335+-0.0101 crypto-aes 7.4948+-0.0339 ? 7.5048+-0.0352 ? crypto-md5 2.4716+-0.0193 ? 2.4782+-0.0201 ? crypto-sha1 2.2424+-0.0214 ? 2.2488+-0.0240 ? date-format-tofte 10.4520+-0.0772 ? 10.5273+-0.1198 ? date-format-xparb 9.8497+-0.0642 ! 10.4410+-0.1977 ! definitely 1.0600x slower math-cordic 7.2277+-0.0559 ? 7.3120+-0.1018 ? might be 1.0117x slower math-partial-sums 10.4219+-0.0092 ? 10.4440+-0.0387 ? math-spectral-norm 2.6396+-0.0052 ? 2.6407+-0.0100 ? regexp-dna 8.8854+-0.0963 ? 8.9600+-0.0721 ? string-base64 4.3374+-0.0303 4.2874+-0.0239 might be 1.0117x faster string-fasta 7.2840+-0.0705 7.2219+-0.0147 string-tagcloud 12.4754+-0.0461 12.3926+-0.0597 string-unpack-code 20.3865+-0.0978 ? 20.4062+-0.0888 ? string-validate-input 5.6296+-0.0324 5.6070+-0.0285 <arithmetic> * 6.4603+-0.0212 ? 6.4799+-0.0159 ? might be 1.0030x slower <geometric> 5.2604+-0.0202 ? 5.2714+-0.0161 ? might be 1.0021x slower <harmonic> 4.1714+-0.0204 ? 4.1782+-0.0181 ? might be 1.0016x slower TipOfTree ArgCountPayload V8: crypto 78.0675+-0.3137 77.8171+-0.3602 deltablue 171.4637+-0.5150 ^ 167.6141+-0.3777 ^ definitely 1.0230x faster earley-boyer 99.2320+-1.2425 ? 99.3733+-1.2785 ? raytrace 52.1713+-0.4113 ^ 51.4091+-0.2221 ^ definitely 1.0148x faster regexp 120.9003+-0.5036 ! 122.1827+-0.3873 ! definitely 1.0106x slower richards 138.0955+-0.2675 ? 138.9790+-0.9531 ? splay 88.9065+-0.6555 ! 92.0309+-2.2569 ! definitely 1.0351x slower <arithmetic> 106.9767+-0.2359 ? 107.0580+-0.4154 ? might be 1.0008x slower <geometric> * 100.3999+-0.2618 ? 100.5686+-0.4412 ? might be 1.0017x slower <harmonic> 93.7660+-0.2931 ? 93.8409+-0.4177 ? might be 1.0008x slower TipOfTree ArgCountPayload Kraken: ai-astar 818.8230+-10.7242 ? 821.6242+-12.1917 ? audio-beat-detection 191.3774+-0.3651 ? 191.4837+-0.4960 ? audio-dft 274.9830+-2.6713 ? 275.4426+-2.8746 ? audio-fft 119.6430+-1.0735 119.6197+-0.6903 audio-oscillator 283.8881+-3.1189 280.9616+-1.9885 might be 1.0104x faster imaging-darkroom 301.1023+-6.3739 299.9262+-6.4462 imaging-desaturate 230.0989+-0.1701 ? 230.1786+-0.1203 ? imaging-gaussian-blur 519.3550+-0.2594 ? 520.1332+-2.0884 ? json-parse-financial 72.1991+-0.3767 ^ 70.7429+-0.2089 ^ definitely 1.0206x faster json-stringify-tinderbox 81.6776+-0.1959 81.6460+-0.2903 stanford-crypto-aes 118.9614+-0.3827 ^ 118.1380+-0.4393 ^ definitely 1.0070x faster stanford-crypto-ccm 115.5464+-0.7556 ^ 113.8675+-0.8451 ^ definitely 1.0147x faster stanford-crypto-pbkdf2 229.9377+-0.6722 ? 230.4209+-0.7746 ? stanford-crypto-sha256-iterative 95.0573+-0.1615 95.0260+-0.2973 <arithmetic> * 246.6179+-1.0310 246.3722+-1.1658 might be 1.0010x faster <geometric> 192.1019+-0.4651 191.4509+-0.5324 might be 1.0034x faster <harmonic> 156.2700+-0.2787 ^ 155.3966+-0.2632 ^ definitely 1.0056x faster TipOfTree ArgCountPayload All benchmarks: <arithmetic> 92.9671+-0.3387 92.9169+-0.3713 might be 1.0005x faster <geometric> 23.8337+-0.0701 ? 23.8431+-0.0644 ? might be 1.0004x slower <harmonic> 7.3471+-0.0353 ? 7.3582+-0.0314 ? might be 1.0015x slower TipOfTree ArgCountPayload Geomean of preferred means: <scaled-result> 54.2836+-0.1565 ? 54.3504+-0.1545 ? might be 1.0012x slower
Filip Pizlo
Comment 2 2011-12-24 14:36:09 PST
Filip Pizlo
Comment 3 2011-12-24 14:41:16 PST
Created attachment 120504 [details] the patch
Filip Pizlo
Comment 4 2011-12-25 18:54:04 PST
Note You need to log in before you can comment on or make changes to this bug.