| Summary: | It's best for the DFG to always have some guess of basic block frequency | ||||||
|---|---|---|---|---|---|---|---|
| Product: | WebKit | Reporter: | Filip Pizlo <fpizlo> | ||||
| Component: | JavaScriptCore | Assignee: | Filip Pizlo <fpizlo> | ||||
| Status: | RESOLVED FIXED | ||||||
| Severity: | Normal | CC: | commit-queue, keith_miller, mark.lam, msaboff, saam | ||||
| Priority: | P2 | ||||||
| Version: | WebKit Nightly Build | ||||||
| Hardware: | All | ||||||
| OS: | All | ||||||
| Bug Depends on: | |||||||
| Bug Blocks: | 150279 | ||||||
| Attachments: |
|
||||||
Created attachment 265685 [details]
the patch
Looks like I didn't break anything. Benchmark report for SunSpider, LongSpider, V8Spider, Octane, Kraken, JSRegress, AsmBench, and CompressionBench on bigmac (MacPro5,1). VMs tested: "TipOfTree" at /Volumes/Data/primary/OpenSource/WebKitBuild/Release/jsc (r192514) "Frequency" at /Volumes/Data/fromMiniMe/tertiary/OpenSource/WebKitBuild/Release/jsc (r192514) Collected 6 samples per benchmark/VM, with 6 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 Frequency SunSpider: 3d-cube 8.2552+-0.2761 ? 8.2840+-0.2600 ? 3d-morph 8.0624+-0.1060 ? 8.1407+-0.1934 ? 3d-raytrace 9.1723+-0.1483 9.0389+-0.1691 might be 1.0148x faster access-binary-trees 3.1639+-0.0969 ? 3.3289+-0.1065 ? might be 1.0521x slower access-fannkuch 8.9762+-0.0995 ? 9.1875+-0.1178 ? might be 1.0235x slower access-nbody 4.4832+-0.1974 ? 4.4981+-0.1389 ? access-nsieve 4.6182+-0.1009 ? 4.6332+-0.0959 ? bitops-3bit-bits-in-byte 1.7852+-0.0663 1.7181+-0.0727 might be 1.0391x faster bitops-bits-in-byte 5.4035+-0.0782 ? 5.4107+-0.0770 ? bitops-bitwise-and 2.8831+-0.0901 ? 2.9375+-0.1121 ? might be 1.0189x slower bitops-nsieve-bits 4.5863+-0.1098 4.5193+-0.0880 might be 1.0148x faster controlflow-recursive 3.5874+-0.0719 ? 3.6277+-0.0779 ? might be 1.0112x slower crypto-aes 6.4564+-0.1544 ? 6.6393+-0.2153 ? might be 1.0283x slower crypto-md5 3.9731+-0.1205 ? 3.9788+-0.1551 ? crypto-sha1 3.5166+-0.1260 ? 3.5281+-0.0821 ? date-format-tofte 12.8392+-0.4077 ? 13.0559+-0.1864 ? might be 1.0169x slower date-format-xparb 7.3150+-0.1442 7.1263+-0.0911 might be 1.0265x faster math-cordic 4.5268+-0.1080 4.5085+-0.0759 math-partial-sums 9.8123+-0.1377 ? 9.8249+-0.1209 ? math-spectral-norm 3.1609+-0.0525 ? 3.1905+-0.0732 ? regexp-dna 9.9397+-0.2866 9.7598+-0.2963 might be 1.0184x faster string-base64 6.4617+-0.2013 ? 6.5350+-0.1473 ? might be 1.0113x slower string-fasta 9.2694+-0.0970 ? 9.3592+-0.1756 ? string-tagcloud 13.0232+-0.3141 13.0118+-0.3682 string-unpack-code 26.4421+-0.3806 ! 27.1060+-0.1460 ! definitely 1.0251x slower string-validate-input 6.7637+-0.1101 ? 6.8742+-0.1901 ? might be 1.0163x slower <arithmetic> 7.2491+-0.0282 ? 7.3009+-0.0299 ? might be 1.0071x slower TipOfTree Frequency LongSpider: 3d-cube 1168.0347+-12.1911 ? 1175.0166+-24.1645 ? 3d-morph 1865.5320+-1.6161 1864.3220+-1.5858 3d-raytrace 993.8361+-9.7404 ? 998.0582+-8.0634 ? access-binary-trees 1345.0069+-3.6972 1344.8436+-4.8203 access-fannkuch 473.7666+-18.4344 ? 477.6772+-26.1713 ? access-nbody 1009.8978+-1.6922 1008.9511+-1.3062 access-nsieve 681.7806+-5.8602 ? 683.2169+-6.2073 ? bitops-3bit-bits-in-byte 45.4134+-0.2441 ? 45.4861+-0.3360 ? bitops-bits-in-byte 320.7441+-2.7017 ? 323.5087+-2.8557 ? bitops-nsieve-bits 618.2791+-3.3224 ? 619.0898+-1.9934 ? controlflow-recursive 749.5474+-3.8963 ! 786.5164+-0.6514 ! definitely 1.0493x slower crypto-aes 893.3391+-3.1467 ? 901.0236+-5.6981 ? crypto-md5 778.0427+-2.1970 ? 780.8865+-3.5824 ? crypto-sha1 1014.7714+-15.2412 1009.3311+-6.3278 date-format-tofte 1013.4473+-18.7723 ? 1017.8041+-6.3546 ? date-format-xparb 1012.7555+-5.2011 ? 1047.9561+-69.4851 ? might be 1.0348x slower hash-map 229.4480+-1.1970 228.4441+-3.3908 math-cordic 632.0919+-2.0133 631.5754+-0.4726 math-partial-sums 1075.7268+-2.9355 ? 1076.1337+-1.6245 ? math-spectral-norm 1074.1571+-1.5023 1073.8257+-1.1795 string-base64 563.8782+-3.9463 563.5545+-1.7302 string-fasta 587.0468+-6.0297 ? 591.1757+-4.0245 ? string-tagcloud 285.8868+-1.5750 285.2466+-1.6825 <geometric> 661.1342+-1.6209 ? 664.6166+-2.5511 ? might be 1.0053x slower TipOfTree Frequency V8Spider: crypto 74.1752+-1.1729 ? 74.7887+-1.2297 ? deltablue 94.1368+-2.5731 ? 96.4892+-6.3399 ? might be 1.0250x slower earley-boyer 65.5438+-0.3772 ? 66.4294+-0.6927 ? might be 1.0135x slower raytrace 43.2033+-1.3821 42.3623+-1.5541 might be 1.0199x faster regexp 102.5854+-0.3257 ? 102.9226+-0.5307 ? richards 78.4596+-1.4547 78.0616+-0.8865 splay 53.4592+-0.4520 ? 53.9432+-0.6842 ? <geometric> 70.3186+-0.5699 ? 70.6486+-0.9741 ? might be 1.0047x slower TipOfTree Frequency Octane: encrypt 0.26371+-0.00710 ^ 0.25493+-0.00130 ^ definitely 1.0344x faster decrypt 4.77881+-0.00690 ? 4.78716+-0.00786 ? deltablue x2 0.21822+-0.00116 ? 0.22095+-0.00158 ? might be 1.0125x slower earley 0.51106+-0.00173 0.51029+-0.00617 boyer 8.31045+-0.07589 8.22905+-0.04754 navier-stokes x2 6.32991+-0.01177 6.32397+-0.00369 raytrace x2 1.47195+-0.00806 ? 1.47638+-0.00756 ? richards x2 0.15245+-0.00118 0.15201+-0.00087 splay x2 0.53307+-0.00254 ? 0.53379+-0.00373 ? regexp x2 38.28769+-0.14871 ? 38.33249+-0.11896 ? pdfjs x2 60.26650+-0.49908 60.15724+-0.34679 mandreel x2 66.23484+-1.10373 ? 66.78193+-1.11426 ? gbemu x2 55.09957+-1.23098 ? 55.68769+-1.28158 ? might be 1.0107x slower closure 0.94113+-0.00263 ? 0.94568+-0.00239 ? jquery 12.91019+-0.06343 ? 13.00293+-0.05667 ? box2d x2 16.40449+-0.10838 16.33493+-0.14011 zlib x2 572.64501+-20.31403 ? 582.50724+-4.82843 ? might be 1.0172x slower typescript x2 1134.73979+-5.65880 ? 1140.29716+-4.50410 ? <geometric> 8.67937+-0.02589 ? 8.69890+-0.02009 ? might be 1.0023x slower TipOfTree Frequency Kraken: ai-astar 217.665+-15.310 212.015+-1.017 might be 1.0266x faster audio-beat-detection 79.294+-0.787 79.111+-0.508 audio-dft 127.961+-1.334 ? 128.615+-1.129 ? audio-fft 58.320+-0.216 ? 58.447+-0.187 ? audio-oscillator 93.070+-0.222 92.951+-0.478 imaging-darkroom 94.864+-0.293 ? 95.085+-0.275 ? imaging-desaturate 90.086+-0.584 89.739+-0.202 imaging-gaussian-blur 154.588+-2.042 154.556+-1.404 json-parse-financial 68.341+-0.505 ! 69.346+-0.497 ! definitely 1.0147x slower json-stringify-tinderbox 40.844+-0.190 40.791+-0.189 stanford-crypto-aes 63.431+-0.580 ? 63.509+-0.646 ? stanford-crypto-ccm 62.886+-1.810 ^ 60.479+-0.516 ^ definitely 1.0398x faster stanford-crypto-pbkdf2 145.869+-0.630 145.718+-0.926 stanford-crypto-sha256-iterative 57.971+-0.438 ? 57.990+-0.558 ? <arithmetic> 96.799+-1.023 96.311+-0.290 might be 1.0051x faster TipOfTree Frequency JSRegress: abc-forward-loop-equal 57.0983+-0.2538 ? 57.3503+-0.2090 ? abc-postfix-backward-loop 56.4307+-0.2428 56.3666+-0.3390 abc-simple-backward-loop 56.1778+-0.2163 56.1366+-0.3775 abc-simple-forward-loop 56.2855+-0.2317 ? 56.3087+-0.2990 ? abc-skippy-loop 39.1727+-0.5026 ? 39.1855+-0.3193 ? abs-boolean 3.9651+-0.1179 3.9112+-0.1225 might be 1.0138x faster adapt-to-double-divide 18.3339+-0.4632 ? 18.5698+-0.6657 ? might be 1.0129x slower aliased-arguments-getbyval 1.7688+-0.1182 1.6714+-0.1238 might be 1.0582x faster allocate-big-object 3.8809+-0.1258 3.8010+-0.2487 might be 1.0210x faster arguments-named-and-reflective 15.7146+-0.4914 ? 15.8170+-0.2779 ? arguments-out-of-bounds 15.4960+-0.7647 15.4152+-0.3545 arguments-strict-mode 13.5990+-0.4249 ? 14.0136+-0.4564 ? might be 1.0305x slower arguments 12.4825+-0.2911 ? 12.5939+-0.4013 ? arity-mismatch-inlining 1.3049+-0.1009 ? 1.3665+-0.0427 ? might be 1.0472x slower array-access-polymorphic-structure 11.3781+-0.1995 11.2727+-0.2174 array-nonarray-polymorhpic-access 41.0453+-1.1471 40.2336+-0.9435 might be 1.0202x faster array-prototype-every 121.2574+-1.0143 ? 123.1488+-1.3473 ? might be 1.0156x slower array-prototype-forEach 120.3295+-1.7792 ? 122.2520+-1.6872 ? might be 1.0160x slower array-prototype-map 135.0972+-2.6303 134.1022+-0.9705 array-prototype-reduce 113.1060+-1.7755 ? 114.2478+-1.0232 ? might be 1.0101x slower array-prototype-reduceRight 113.0359+-1.2512 ? 113.9913+-1.0491 ? array-prototype-some 121.9730+-1.4999 ? 123.6344+-1.2052 ? might be 1.0136x slower array-splice-contiguous 36.9843+-0.3626 36.5884+-0.1229 might be 1.0108x faster array-with-double-add 5.8028+-0.0909 ? 5.8685+-0.0551 ? might be 1.0113x slower array-with-double-increment 4.4583+-0.1673 ? 4.5825+-0.0616 ? might be 1.0279x slower array-with-double-mul-add 7.5203+-0.1324 ? 7.6060+-0.0655 ? might be 1.0114x slower array-with-double-sum 4.4840+-0.1434 4.4698+-0.0817 array-with-int32-add-sub 10.3043+-0.1270 ? 10.3253+-0.1200 ? array-with-int32-or-double-sum 4.5514+-0.1102 4.4947+-0.0473 might be 1.0126x faster ArrayBuffer-DataView-alloc-large-long-lived 44.9657+-0.8675 ! 47.3111+-0.6038 ! definitely 1.0522x slower ArrayBuffer-DataView-alloc-long-lived 18.3693+-0.2542 ? 18.7700+-0.2688 ? might be 1.0218x slower ArrayBuffer-Int32Array-byteOffset 5.6102+-0.1639 5.4830+-0.1433 might be 1.0232x faster ArrayBuffer-Int8Array-alloc-large-long-lived 46.4411+-0.9064 ? 47.2448+-0.8976 ? might be 1.0173x slower ArrayBuffer-Int8Array-alloc-long-lived-buffer 30.4200+-1.0734 30.0104+-0.3538 might be 1.0136x faster ArrayBuffer-Int8Array-alloc-long-lived 16.8656+-0.1217 ! 17.6127+-0.1921 ! definitely 1.0443x slower ArrayBuffer-Int8Array-alloc 14.3457+-0.2404 ? 14.7969+-0.2428 ? might be 1.0315x slower arrowfunction-call 15.5462+-0.3329 ? 15.5569+-0.1598 ? asmjs_bool_bug 9.7297+-0.1208 9.7226+-0.0847 assign-custom-setter-polymorphic 4.3818+-0.1728 4.3228+-0.1048 might be 1.0137x faster assign-custom-setter 5.9789+-0.1805 ? 6.0598+-0.2984 ? might be 1.0135x slower basic-set 11.1920+-0.2630 ? 11.3861+-0.1673 ? might be 1.0173x slower big-int-mul 5.8753+-0.0962 5.8583+-0.0958 boolean-test 4.5658+-0.0864 ? 4.5758+-0.1044 ? branch-fold 4.8533+-0.0537 4.7940+-0.0251 might be 1.0124x faster branch-on-string-as-boolean 25.6907+-0.5471 ? 25.8632+-0.7428 ? by-val-generic 3.8268+-0.0660 3.7692+-0.1106 might be 1.0153x faster call-spread-apply 39.9024+-0.4669 ! 42.6313+-0.5600 ! definitely 1.0684x slower call-spread-call 31.7376+-0.3114 ! 35.6080+-0.7018 ! definitely 1.1219x slower captured-assignments 0.7252+-0.0182 ? 0.7817+-0.0870 ? might be 1.0779x slower cast-int-to-double 8.6973+-0.0862 8.6458+-0.0668 cell-argument 8.5935+-0.1722 ? 8.8319+-0.3781 ? might be 1.0277x slower cfg-simplify 3.9362+-0.0950 3.8846+-0.0246 might be 1.0133x faster chain-getter-access 11.8465+-0.3948 ? 11.9659+-0.4620 ? might be 1.0101x slower cmpeq-obj-to-obj-other 14.6276+-0.6998 ? 14.9112+-0.5574 ? might be 1.0194x slower constant-test 8.1577+-0.0795 8.1075+-0.1030 create-lots-of-functions 16.4913+-0.1914 ? 16.5217+-0.2581 ? cse-new-array-buffer 3.4190+-0.1208 ? 3.5041+-0.1442 ? might be 1.0249x slower cse-new-array 3.5027+-0.1389 3.4843+-0.1432 custom-setter-getter-as-put-get-by-id 0.9612+-0.0206 0.8764+-0.1163 might be 1.0967x faster DataView-custom-properties 52.5960+-0.9036 ! 54.1355+-0.5778 ! definitely 1.0293x slower delay-tear-off-arguments-strictmode 19.7168+-0.2547 ? 19.9641+-0.5668 ? might be 1.0125x slower deltablue-varargs 277.7861+-1.7032 ? 278.2091+-1.2912 ? destructuring-arguments 247.8931+-1.6977 ? 248.3360+-1.2859 ? destructuring-parameters-overridden-by-function 0.8606+-0.0495 ^ 0.7229+-0.0539 ^ definitely 1.1905x faster destructuring-swap 7.8481+-0.0810 ? 7.8485+-0.0642 ? direct-arguments-getbyval 1.6495+-0.1329 ? 1.7591+-0.1796 ? might be 1.0665x slower div-boolean-double 5.6467+-0.0846 5.6053+-0.1048 div-boolean 9.8783+-0.1199 ? 9.9780+-0.0794 ? might be 1.0101x slower double-get-by-val-out-of-bounds 6.4038+-0.1216 6.3774+-0.0829 double-pollution-getbyval 9.8056+-0.0631 9.7885+-0.0649 double-pollution-putbyoffset 5.5431+-0.0314 ? 5.5819+-0.1753 ? double-real-use 40.3544+-0.0845 ? 40.4128+-0.1090 ? double-to-int32-typed-array-no-inline 3.1081+-0.0794 ? 3.1814+-0.1159 ? might be 1.0236x slower double-to-int32-typed-array 2.9550+-0.1040 ? 2.9802+-0.0825 ? double-to-uint32-typed-array-no-inline 3.3137+-0.0899 ? 3.3534+-0.1214 ? might be 1.0120x slower double-to-uint32-typed-array 2.9344+-0.1152 2.9108+-0.1193 elidable-new-object-dag 54.8728+-0.6164 54.7115+-1.2883 elidable-new-object-roflcopter 55.2279+-0.7517 ? 55.3599+-0.5616 ? elidable-new-object-then-call 56.3457+-1.2811 55.9964+-1.1966 elidable-new-object-tree 65.0552+-1.5554 64.8909+-0.5423 empty-string-plus-int 7.5897+-0.1114 ! 7.7972+-0.0953 ! definitely 1.0273x slower emscripten-cube2hash 49.3171+-0.1771 ? 49.8178+-0.9033 ? might be 1.0102x slower exit-length-on-plain-object 23.6861+-0.7966 ? 24.4648+-0.7077 ? might be 1.0329x slower external-arguments-getbyval 1.7907+-0.1235 1.7424+-0.1007 might be 1.0277x faster external-arguments-putbyval 3.4541+-0.0722 ? 3.5121+-0.1115 ? might be 1.0168x slower fixed-typed-array-storage-var-index 1.6320+-0.0570 ? 1.6402+-0.0690 ? fixed-typed-array-storage 1.3944+-0.0940 1.3942+-0.1237 Float32Array-matrix-mult 6.7694+-0.2178 ? 6.8583+-0.1313 ? might be 1.0131x slower Float32Array-to-Float64Array-set 79.6664+-0.6509 ? 81.4068+-1.5160 ? might be 1.0218x slower Float64Array-alloc-long-lived 95.5214+-0.2796 ? 95.8938+-1.1558 ? Float64Array-to-Int16Array-set 89.5785+-0.8263 89.5246+-1.2445 fold-double-to-int 19.2267+-0.0926 ? 19.2496+-0.2315 ? fold-get-by-id-to-multi-get-by-offset-rare-int 12.2999+-0.1347 ? 12.5632+-0.3575 ? might be 1.0214x slower fold-get-by-id-to-multi-get-by-offset 10.6818+-0.1570 10.5540+-0.2904 might be 1.0121x faster fold-multi-get-by-offset-to-get-by-offset 9.1057+-0.4509 ? 9.3018+-0.0904 ? might be 1.0215x slower fold-multi-get-by-offset-to-poly-get-by-offset 9.6507+-0.4378 ? 9.8204+-0.0715 ? might be 1.0176x slower fold-multi-put-by-offset-to-poly-put-by-offset 10.9620+-0.5526 ? 11.1475+-0.4818 ? might be 1.0169x slower fold-multi-put-by-offset-to-put-by-offset 10.6500+-0.3414 ? 10.6687+-0.3862 ? fold-multi-put-by-offset-to-replace-or-transition-put-by-offset 15.7686+-0.7384 ? 16.1527+-0.8078 ? might be 1.0244x slower fold-put-by-id-to-multi-put-by-offset 12.5382+-0.5663 ? 12.6263+-0.6438 ? fold-put-by-val-with-string-to-multi-put-by-offset 12.7654+-0.4801 12.7620+-0.3853 fold-put-by-val-with-symbol-to-multi-put-by-offset 12.6740+-0.5141 ? 12.7136+-0.5907 ? fold-put-structure 8.5662+-0.1402 ? 8.6390+-0.0951 ? for-of-iterate-array-entries 16.2968+-0.2299 ? 16.4652+-0.5820 ? might be 1.0103x slower for-of-iterate-array-keys 5.4843+-0.2253 5.2854+-0.0851 might be 1.0376x faster for-of-iterate-array-values 5.1651+-0.1945 ? 5.2494+-0.2495 ? might be 1.0163x slower fround 21.9996+-0.3972 21.7922+-0.6335 ftl-library-inlining-dataview 92.7352+-0.6036 ? 93.2347+-2.6847 ? ftl-library-inlining 104.3940+-18.5195 ? 110.2295+-15.3709 ? might be 1.0559x slower ftl-object-sub 43.4895+-0.2921 ? 44.1049+-0.3424 ? might be 1.0141x slower function-call 15.6743+-0.4775 ? 15.8781+-0.3365 ? might be 1.0130x slower function-dot-apply 3.1146+-0.0489 ? 3.1611+-0.0721 ? might be 1.0149x slower function-test 4.4167+-0.0895 ? 4.4204+-0.1407 ? function-with-eval 127.6180+-0.6862 127.5048+-0.8736 gcse-poly-get-less-obvious 33.7295+-0.3657 33.6196+-0.4753 gcse-poly-get 34.4694+-0.2421 34.4655+-0.1281 gcse 6.8016+-0.0993 6.7325+-0.1271 might be 1.0103x faster get-by-id-bimorphic-check-structure-elimination-simple 3.2883+-0.1008 ? 3.3011+-0.0757 ? get-by-id-bimorphic-check-structure-elimination 8.1949+-0.1870 ! 9.2222+-0.0553 ! definitely 1.1254x slower get-by-id-chain-from-try-block 3.5213+-0.1665 ? 3.6216+-0.0574 ? might be 1.0285x slower get-by-id-check-structure-elimination 7.6797+-0.0633 ? 7.6962+-0.0779 ? get-by-id-proto-or-self 23.8837+-0.9044 ? 24.9967+-2.3163 ? might be 1.0466x slower get-by-id-quadmorphic-check-structure-elimination-simple 4.0430+-0.1053 ? 4.1126+-0.0652 ? might be 1.0172x slower get-by-id-self-or-proto 24.3988+-2.7313 ? 26.4021+-2.0253 ? might be 1.0821x slower get-by-val-out-of-bounds 6.3533+-0.0724 6.2799+-0.1426 might be 1.0117x faster get-by-val-with-string-bimorphic-check-structure-elimination-simple 3.9530+-0.0691 ? 4.0201+-0.0700 ? might be 1.0170x slower get-by-val-with-string-bimorphic-check-structure-elimination 10.7635+-0.1282 10.7199+-0.1371 get-by-val-with-string-chain-from-try-block 3.6471+-0.0892 3.6293+-0.1191 get-by-val-with-string-check-structure-elimination 9.4814+-0.2151 ? 9.5975+-0.0868 ? might be 1.0122x slower get-by-val-with-string-proto-or-self 25.9869+-2.2157 24.8047+-1.4677 might be 1.0477x faster get-by-val-with-string-quadmorphic-check-structure-elimination-simple 4.9911+-0.0981 4.9811+-0.0621 get-by-val-with-string-self-or-proto 24.5103+-1.5050 ? 25.2354+-2.0932 ? might be 1.0296x slower get-by-val-with-symbol-bimorphic-check-structure-elimination-simple 4.9439+-0.0817 ^ 4.6611+-0.0829 ^ definitely 1.0607x faster get-by-val-with-symbol-bimorphic-check-structure-elimination 20.0223+-0.3784 ^ 19.3363+-0.0848 ^ definitely 1.0355x faster get-by-val-with-symbol-chain-from-try-block 3.7142+-0.1475 3.7054+-0.0861 get-by-val-with-symbol-check-structure-elimination 18.4132+-0.1418 ? 18.4937+-0.2357 ? get-by-val-with-symbol-proto-or-self 23.6595+-0.6481 ? 25.4265+-1.9439 ? might be 1.0747x slower get-by-val-with-symbol-quadmorphic-check-structure-elimination-simple 6.0933+-0.0745 6.0926+-0.0780 get-by-val-with-symbol-self-or-proto 24.8387+-1.4891 ? 24.9672+-1.5565 ? get_callee_monomorphic 3.7059+-0.1258 ? 3.7793+-0.1871 ? might be 1.0198x slower get_callee_polymorphic 4.7537+-0.1145 ? 4.9031+-0.1258 ? might be 1.0314x slower getter-no-activation 5.9128+-0.0726 5.8993+-0.0625 getter-prototype 12.8278+-0.3666 ? 12.9014+-0.4076 ? getter-richards-try-catch 1749.6302+-75.4854 ? 1787.0940+-38.6182 ? might be 1.0214x slower getter-richards 139.5913+-2.5437 ? 142.0457+-2.4650 ? might be 1.0176x slower getter 7.3012+-0.0608 ? 7.3964+-0.0607 ? might be 1.0130x slower global-object-access-with-mutating-structure 8.4238+-0.0605 8.3014+-0.0936 might be 1.0147x faster global-var-const-infer-fire-from-opt 1.2204+-0.1059 1.2137+-0.1500 global-var-const-infer 1.1641+-0.1487 ? 1.2026+-0.1259 ? might be 1.0330x slower hard-overflow-check-equal 45.7783+-0.4128 ? 45.7867+-0.2912 ? hard-overflow-check 45.5964+-0.2521 45.5181+-0.2560 HashMap-put-get-iterate-keys 34.0174+-0.5459 ? 34.2996+-0.5372 ? HashMap-put-get-iterate 34.6268+-0.5229 34.2650+-0.4785 might be 1.0106x faster HashMap-string-put-get-iterate 36.8429+-0.8913 36.7206+-0.7599 hoist-make-rope 15.6480+-1.6355 15.5031+-0.0625 hoist-poly-check-structure-effectful-loop 6.6705+-0.0896 ? 6.6854+-0.1166 ? hoist-poly-check-structure 4.5125+-0.0868 ? 4.5402+-0.1049 ? imul-double-only 10.3415+-0.1922 ? 10.5275+-0.1114 ? might be 1.0180x slower imul-int-only 12.7636+-0.5724 11.8916+-0.4341 might be 1.0733x faster imul-mixed 10.4492+-0.5651 10.0872+-0.6026 might be 1.0359x faster in-four-cases 28.3348+-0.1295 ? 28.5244+-0.2241 ? in-one-case-false 16.3925+-0.1248 ? 16.4686+-0.0830 ? in-one-case-true 16.3980+-0.1330 16.3694+-0.1101 in-two-cases 17.0072+-0.2031 ? 17.0748+-0.0974 ? indexed-properties-in-objects 4.0548+-0.0933 ? 4.1419+-0.0787 ? might be 1.0215x slower infer-closure-const-then-mov-no-inline 5.4362+-0.0724 5.4341+-0.0762 infer-closure-const-then-mov 22.9956+-0.2372 ! 23.3829+-0.0877 ! definitely 1.0168x slower infer-closure-const-then-put-to-scope-no-inline 16.2022+-0.1610 16.0630+-0.2222 infer-closure-const-then-put-to-scope 27.9715+-0.2309 ? 28.0444+-0.2400 ? infer-closure-const-then-reenter-no-inline 71.1827+-0.3376 ? 71.2020+-0.2826 ? infer-closure-const-then-reenter 32.8709+-0.1836 32.8066+-0.2669 infer-constant-global-property 4.9312+-0.1268 4.8401+-0.0889 might be 1.0188x faster infer-constant-property 3.3669+-0.0871 ? 3.3763+-0.1018 ? infer-one-time-closure-ten-vars 12.7410+-0.2000 12.7057+-0.3127 infer-one-time-closure-two-vars 11.9638+-0.2842 ? 12.2467+-0.4121 ? might be 1.0236x slower infer-one-time-closure 11.8714+-0.2557 ? 12.0137+-0.1061 ? might be 1.0120x slower infer-one-time-deep-closure 19.2411+-0.3103 ? 19.3161+-0.4689 ? inline-arguments-access 6.0874+-0.0499 6.0605+-0.1347 inline-arguments-aliased-access 6.1289+-0.1361 6.0913+-0.1208 inline-arguments-local-escape 6.2095+-0.0951 ? 6.2177+-0.2467 ? inline-get-scoped-var 5.8250+-0.0756 ? 5.8311+-0.0754 ? inlined-put-by-id-transition 15.4196+-0.2735 ? 15.7765+-0.3028 ? might be 1.0231x slower inlined-put-by-val-with-string-transition 67.8275+-1.3571 ? 70.5634+-2.3386 ? might be 1.0403x slower inlined-put-by-val-with-symbol-transition 69.0780+-0.7154 ! 70.6784+-0.7863 ! definitely 1.0232x slower int-or-other-abs-then-get-by-val 6.7251+-0.1049 6.7104+-0.1391 int-or-other-abs-zero-then-get-by-val 27.7333+-0.1094 ? 27.9215+-0.2033 ? int-or-other-add-then-get-by-val 6.5792+-0.0823 6.5244+-0.1013 int-or-other-add 8.2566+-0.0650 ? 8.3608+-0.0499 ? might be 1.0126x slower int-or-other-div-then-get-by-val 5.1211+-0.1494 ? 5.2274+-0.0865 ? might be 1.0208x slower int-or-other-max-then-get-by-val 5.1191+-0.0427 5.1122+-0.0481 int-or-other-min-then-get-by-val 5.1471+-0.1064 ? 5.2186+-0.0881 ? might be 1.0139x slower int-or-other-mod-then-get-by-val 5.0021+-0.0860 4.9881+-0.0513 int-or-other-mul-then-get-by-val 5.2786+-0.0745 5.2708+-0.0833 int-or-other-neg-then-get-by-val 5.8306+-0.0946 ? 5.9118+-0.0840 ? might be 1.0139x slower int-or-other-neg-zero-then-get-by-val 28.2672+-0.1215 28.2529+-0.1326 int-or-other-sub-then-get-by-val 6.5256+-0.0645 6.4811+-0.0907 int-or-other-sub 5.4032+-0.1023 5.4016+-0.0630 int-overflow-local 6.1253+-0.0293 ? 6.1509+-0.0732 ? Int16Array-alloc-long-lived 66.6381+-0.5658 66.3757+-0.4909 Int16Array-bubble-sort-with-byteLength 39.5272+-0.1838 ? 39.5333+-0.1507 ? Int16Array-bubble-sort 40.6455+-0.1613 ? 40.7747+-0.2388 ? Int16Array-load-int-mul 2.1605+-0.0485 2.1537+-0.0784 Int16Array-to-Int32Array-set 71.9015+-1.3447 70.3369+-0.6989 might be 1.0222x faster Int32Array-alloc-large 32.7819+-0.5169 32.4024+-0.5608 might be 1.0117x faster Int32Array-alloc-long-lived 75.4653+-0.4022 75.0424+-0.5180 Int32Array-alloc 4.6605+-0.0696 ? 4.6726+-0.0306 ? Int32Array-Int8Array-view-alloc 9.4331+-0.1680 ? 9.5736+-0.1477 ? might be 1.0149x slower int52-spill 7.0711+-0.5007 ? 7.1612+-0.1852 ? might be 1.0127x slower Int8Array-alloc-long-lived 59.4503+-0.5762 59.0886+-0.7708 Int8Array-load-with-byteLength 4.8617+-0.0920 4.8241+-0.0869 Int8Array-load 4.8970+-0.0621 4.8305+-0.0734 might be 1.0138x faster integer-divide 13.9936+-0.0962 13.9883+-0.1202 integer-modulo 2.7795+-0.0957 2.7738+-0.1158 is-boolean-fold-tricky 5.7353+-0.0843 ? 5.7359+-0.0749 ? is-boolean-fold 4.1177+-0.0878 4.0837+-0.1390 is-function-fold-tricky-internal-function 15.8111+-0.1496 ^ 14.8999+-0.3564 ^ definitely 1.0612x faster is-function-fold-tricky 5.8927+-0.0528 ? 5.8937+-0.0741 ? is-function-fold 4.1699+-0.0709 ? 4.1995+-0.0728 ? is-number-fold-tricky 5.8431+-0.1509 5.7969+-0.0721 is-number-fold 4.0403+-0.0915 ? 4.0958+-0.0802 ? might be 1.0137x slower is-object-or-null-fold-functions 4.1721+-0.0965 ? 4.2415+-0.0932 ? might be 1.0166x slower is-object-or-null-fold-less-tricky 5.7939+-0.0966 ? 5.8828+-0.1163 ? might be 1.0153x slower is-object-or-null-fold-tricky 7.4220+-0.1519 ? 7.4763+-0.0723 ? is-object-or-null-fold 4.1857+-0.1096 4.1496+-0.1063 is-object-or-null-trickier-function 5.9767+-0.0931 ? 5.9780+-0.0472 ? is-object-or-null-trickier-internal-function 15.5227+-0.1056 15.4780+-0.3312 is-object-or-null-tricky-function 5.8551+-0.0826 ? 5.9474+-0.1443 ? might be 1.0158x slower is-object-or-null-tricky-internal-function 11.5699+-0.5058 11.4434+-0.1067 might be 1.0111x faster is-string-fold-tricky 5.7255+-0.0679 ? 5.8018+-0.1048 ? might be 1.0133x slower is-string-fold 4.0684+-0.1091 4.0566+-0.1007 is-undefined-fold-tricky 4.7871+-0.0862 ? 4.8720+-0.1076 ? might be 1.0177x slower is-undefined-fold 4.0771+-0.0784 ? 4.0839+-0.0778 ? JSONP-negative-0 0.4478+-0.0413 ? 0.4587+-0.0574 ? might be 1.0243x slower large-int-captured 7.0445+-0.1316 6.7862+-0.1566 might be 1.0381x faster large-int-neg 21.6218+-0.2117 21.4717+-0.1588 large-int 18.9154+-0.2272 ? 19.1070+-0.2924 ? might be 1.0101x slower load-varargs-elimination 30.6027+-1.0911 ? 30.8670+-0.4962 ? logical-not-weird-types 5.3874+-0.1765 5.2458+-0.0911 might be 1.0270x faster logical-not 6.5876+-0.0702 ? 6.6953+-0.1807 ? might be 1.0163x slower lots-of-fields 17.3774+-0.2392 ? 17.6038+-0.2433 ? might be 1.0130x slower make-indexed-storage 4.5338+-0.1115 4.3968+-0.2308 might be 1.0312x faster make-rope-cse 6.4963+-0.1345 ? 6.5212+-0.1271 ? marsaglia-larger-ints 53.4628+-0.6801 52.5704+-0.6183 might be 1.0170x faster marsaglia-osr-entry 28.4028+-0.4762 ? 28.5441+-0.2763 ? math-with-out-of-bounds-array-values 34.1890+-0.7777 ? 35.8037+-2.0009 ? might be 1.0472x slower max-boolean 3.3975+-0.1106 3.3753+-0.0606 method-on-number 23.6970+-0.7063 ? 24.1791+-0.1689 ? might be 1.0203x slower min-boolean 3.3794+-0.1060 ? 3.4165+-0.0669 ? might be 1.0110x slower minus-boolean-double 4.2959+-0.0920 4.2818+-0.0637 minus-boolean 3.3947+-0.0765 3.3875+-0.0862 misc-strict-eq 50.1497+-2.5427 ? 50.2730+-2.1787 ? mod-boolean-double 11.8858+-0.1480 ? 11.9358+-0.0610 ? mod-boolean 9.2092+-0.0377 9.0761+-0.1944 might be 1.0147x faster mul-boolean-double 4.9768+-0.0547 4.9539+-0.0669 mul-boolean 3.6112+-0.1161 3.5927+-0.0631 neg-boolean 4.3506+-0.0603 4.3073+-0.0757 might be 1.0101x faster negative-zero-divide 0.6000+-0.0642 ? 0.6083+-0.0826 ? might be 1.0138x slower negative-zero-modulo 0.5892+-0.0772 0.5708+-0.0722 might be 1.0322x faster negative-zero-negate 0.6181+-0.0642 0.5745+-0.0707 might be 1.0759x faster nested-function-parsing 73.1155+-0.1388 73.0922+-0.4471 new-array-buffer-dead 144.3565+-0.9545 ? 144.6703+-0.8298 ? new-array-buffer-push 9.9153+-0.2110 9.8831+-0.2906 new-array-dead 23.1180+-1.5206 ? 24.8457+-0.9605 ? might be 1.0747x slower new-array-push 5.5283+-0.1711 5.5090+-0.1935 no-inline-constructor 50.1815+-0.6416 49.9788+-0.4621 number-test 4.4834+-0.1296 ? 4.5343+-0.1010 ? might be 1.0114x slower object-closure-call 7.5328+-0.1646 ? 7.6873+-0.0626 ? might be 1.0205x slower object-get-own-property-symbols-on-large-array 5.3242+-0.3472 5.2199+-0.2904 might be 1.0200x faster object-test 4.3305+-0.1341 ? 4.4122+-0.1232 ? might be 1.0189x slower obvious-sink-pathology-taken 166.9730+-1.6237 165.3880+-0.5837 obvious-sink-pathology 52.7712+-1.5720 52.3824+-0.9857 obviously-elidable-new-object 49.8206+-1.6329 49.2730+-3.0416 might be 1.0111x faster plus-boolean-arith 3.3374+-0.0666 3.3171+-0.0755 plus-boolean-double 4.3700+-0.0846 4.3111+-0.0543 might be 1.0136x faster plus-boolean 3.2889+-0.0707 3.2817+-0.0807 poly-chain-access-different-prototypes-simple 3.9633+-0.0904 3.9441+-0.0452 poly-chain-access-different-prototypes 3.4301+-0.0988 3.4136+-0.0680 poly-chain-access-simpler 3.9430+-0.0833 3.9246+-0.0845 poly-chain-access 3.3568+-0.0745 ? 3.3639+-0.0640 ? poly-stricteq 83.6080+-2.0937 82.1537+-0.1467 might be 1.0177x faster polymorphic-array-call 2.0964+-0.2685 1.9938+-0.2174 might be 1.0514x faster polymorphic-get-by-id 4.2185+-0.0454 ? 4.3202+-0.3527 ? might be 1.0241x slower polymorphic-put-by-id 54.6951+-3.2458 53.5030+-6.6871 might be 1.0223x faster polymorphic-put-by-val-with-string 44.5520+-0.6760 44.2771+-0.8165 polymorphic-put-by-val-with-symbol 43.9569+-0.2389 ? 44.6213+-0.6032 ? might be 1.0151x slower polymorphic-structure 25.1838+-0.0744 ? 25.1941+-0.1499 ? polyvariant-monomorphic-get-by-id 11.6264+-0.0383 ? 11.6609+-0.0698 ? proto-getter-access 11.7205+-0.4204 ? 11.8282+-0.2868 ? prototype-access-with-mutating-prototype 7.9210+-0.2569 ? 8.2675+-0.4219 ? might be 1.0438x slower put-by-id-replace-and-transition 12.4801+-0.1428 ? 12.6898+-0.1724 ? might be 1.0168x slower put-by-id-slightly-polymorphic 3.6768+-0.0837 3.6603+-0.0893 put-by-id 18.4882+-0.2465 ? 18.5260+-0.2936 ? put-by-val-direct 0.6434+-0.0788 ? 0.6848+-0.0501 ? might be 1.0644x slower put-by-val-large-index-blank-indexing-type 8.6146+-0.2073 ? 8.8374+-0.1631 ? might be 1.0259x slower put-by-val-machine-int 3.7202+-0.1348 ? 3.7581+-0.1606 ? might be 1.0102x slower put-by-val-with-string-replace-and-transition 19.2592+-0.2911 19.2561+-0.1744 put-by-val-with-string-slightly-polymorphic 4.8575+-0.1319 4.7395+-0.0961 might be 1.0249x faster put-by-val-with-string 19.3541+-0.5955 ? 19.3998+-0.3106 ? put-by-val-with-symbol-replace-and-transition 20.7117+-0.2157 ? 20.8037+-0.3222 ? put-by-val-with-symbol-slightly-polymorphic 5.0675+-0.1013 5.0565+-0.0882 put-by-val-with-symbol 19.3127+-0.4178 ? 19.6142+-0.2714 ? might be 1.0156x slower rare-osr-exit-on-local 17.5236+-0.3287 ? 17.6647+-0.2309 ? raytrace-with-empty-try-catch 9.3656+-0.2319 ? 9.3791+-0.0742 ? raytrace-with-try-catch 16.8841+-0.1882 ! 17.2826+-0.1138 ! definitely 1.0236x slower register-pressure-from-osr 25.1679+-0.1008 ? 25.1740+-0.1573 ? repeat-multi-get-by-offset 28.4865+-0.2664 ? 28.6497+-0.3563 ? richards-empty-try-catch 89.3465+-1.2759 ? 90.4603+-1.5702 ? might be 1.0125x slower richards-try-catch 386.8007+-3.3703 ? 390.6119+-6.9005 ? setter-prototype 13.3212+-0.7881 ? 13.9446+-1.0998 ? might be 1.0468x slower setter 6.8726+-0.0455 ? 6.9880+-0.0908 ? might be 1.0168x slower simple-activation-demo 35.2362+-0.1845 ^ 34.8442+-0.1786 ^ definitely 1.0112x faster simple-getter-access 15.4832+-0.1581 ? 15.5696+-0.2999 ? simple-poly-call-nested 10.7813+-0.3056 ? 10.9715+-0.0732 ? might be 1.0176x slower simple-poly-call 1.8656+-0.0693 ? 1.8928+-0.1042 ? might be 1.0145x slower sin-boolean 27.3680+-0.6626 25.2616+-2.1553 might be 1.0834x faster singleton-scope 84.1790+-0.1416 ? 84.2393+-0.2853 ? sink-function 14.4853+-1.3394 ? 14.9600+-1.3678 ? might be 1.0328x slower sink-huge-activation 21.7392+-0.3921 ? 22.2689+-0.5066 ? might be 1.0244x slower sinkable-new-object-dag 87.6818+-4.4613 ? 90.0717+-2.0848 ? might be 1.0273x slower sinkable-new-object-taken 67.6926+-3.2551 67.5728+-2.9152 sinkable-new-object 53.7617+-1.5398 53.4140+-0.4930 slow-array-profile-convergence 4.2314+-0.1709 4.1525+-0.2136 might be 1.0190x faster slow-convergence 3.8775+-0.0617 ! 4.1629+-0.1238 ! definitely 1.0736x slower slow-ternaries 33.2450+-0.7067 33.0503+-0.2491 sorting-benchmark 26.3014+-0.6654 ? 26.3295+-0.1906 ? sparse-conditional 1.7994+-0.0965 1.7603+-0.0315 might be 1.0222x faster splice-to-remove 19.7700+-0.1933 19.7144+-0.2010 string-char-code-at 20.8072+-0.0563 20.6416+-0.1420 string-concat-object 3.2555+-0.0681 ? 3.2596+-0.0537 ? string-concat-pair-object 3.2552+-0.1644 3.1650+-0.0664 might be 1.0285x faster string-concat-pair-simple 17.0635+-0.2728 17.0041+-0.2824 string-concat-simple 17.4845+-0.3216 17.2687+-0.0966 might be 1.0125x faster string-cons-repeat 11.1936+-0.1384 ? 11.3032+-0.1006 ? string-cons-tower 10.8926+-0.0377 10.8531+-0.1282 string-equality 24.8263+-0.1569 24.7982+-0.1104 string-get-by-val-big-char 10.5704+-0.1915 10.4684+-0.4285 string-get-by-val-out-of-bounds-insane 5.1607+-0.1438 ! 5.5033+-0.1097 ! definitely 1.0664x slower string-get-by-val-out-of-bounds 6.7981+-0.0867 ? 6.8143+-0.1119 ? string-get-by-val 4.8422+-0.0558 ? 4.8600+-0.0642 ? string-hash 2.9015+-0.0523 2.8294+-0.1025 might be 1.0255x faster string-long-ident-equality 20.0990+-0.1817 ? 20.1384+-0.1333 ? string-out-of-bounds 16.3010+-0.1650 16.2238+-0.3899 string-repeat-arith 42.7288+-0.4380 42.3068+-0.1965 string-rope-with-object 26.2848+-0.4786 26.1248+-0.3179 string-sub 57.1286+-0.8624 56.2855+-0.7947 might be 1.0150x faster string-test 4.5101+-0.1435 4.4410+-0.0506 might be 1.0156x faster string-var-equality 60.6885+-2.3217 ? 60.8813+-2.3835 ? structure-hoist-over-transitions 3.5516+-0.0955 ? 3.5581+-0.1204 ? substring-concat-weird 57.4040+-0.3022 ? 57.4198+-0.5872 ? substring-concat 62.2273+-0.1579 62.1440+-0.3078 substring 68.5945+-0.2007 ? 68.8549+-0.2347 ? switch-char-constant 3.4122+-0.0946 ? 3.4912+-0.0675 ? might be 1.0232x slower switch-char 8.1829+-0.0737 ? 8.2107+-0.0466 ? switch-constant 11.4731+-0.1806 ? 11.4901+-0.3727 ? switch-string-basic-big-var 32.7152+-0.2694 32.4896+-0.3740 switch-string-basic-big 28.3410+-1.9483 ? 29.1737+-0.3171 ? might be 1.0294x slower switch-string-basic-var 32.1380+-0.7536 ? 32.4493+-0.2491 ? switch-string-basic 22.9763+-0.4111 ? 23.6620+-0.9364 ? might be 1.0298x slower switch-string-big-length-tower-var 28.3169+-0.5712 ? 28.5033+-0.2486 ? switch-string-length-tower-var 21.2496+-0.2364 ? 21.2511+-0.1600 ? switch-string-length-tower 15.5093+-0.1384 ? 15.5658+-0.2037 ? switch-string-short 15.5840+-0.1930 ? 15.7295+-0.3400 ? switch 18.2242+-0.2254 17.9987+-0.2816 might be 1.0125x faster symbol-tostringtag 4.9302+-0.0928 4.8675+-0.0600 might be 1.0129x faster tear-off-arguments-simple 4.6292+-0.1216 4.5649+-0.1787 might be 1.0141x faster tear-off-arguments 6.4578+-0.2642 ? 6.4879+-0.3139 ? temporal-structure 16.5257+-0.1780 ? 16.5586+-0.1056 ? to-int32-boolean 21.1988+-0.1618 ? 26.3770+-13.6216 ? might be 1.2443x slower try-catch-get-by-val-cloned-arguments 14.9428+-0.1063 ? 14.9723+-0.0942 ? try-catch-get-by-val-direct-arguments 3.3309+-0.1137 3.3014+-0.1194 try-catch-get-by-val-scoped-arguments 7.0446+-0.1097 ? 7.0958+-0.1089 ? typed-array-get-set-by-val-profiling 37.5870+-0.9462 ? 39.6179+-2.9118 ? might be 1.0540x slower undefined-property-access 439.3280+-0.5299 438.6209+-0.7152 undefined-test 4.6315+-0.1149 ? 4.6982+-0.0783 ? might be 1.0144x slower unprofiled-licm 16.4383+-0.1698 ? 16.6578+-0.2684 ? might be 1.0134x slower v8-raytrace-with-empty-try-catch 90.4266+-0.9780 90.2095+-0.4018 v8-raytrace-with-try-catch 110.6370+-1.1344 ! 112.5999+-0.7846 ! definitely 1.0177x slower varargs-call 18.8903+-0.3807 18.8372+-0.4026 varargs-construct-inline 37.0325+-0.2108 ? 37.0343+-0.3272 ? varargs-construct 28.7948+-0.2164 ? 28.7979+-0.2762 ? varargs-inline 12.7139+-0.2457 12.6523+-0.2690 varargs-strict-mode 14.5151+-0.1396 ? 14.6667+-0.2298 ? might be 1.0104x slower varargs 14.3920+-0.1949 ? 14.5308+-0.1105 ? weird-inlining-const-prop 3.2130+-0.0788 ? 3.2983+-0.1285 ? might be 1.0265x slower <geometric> 12.3044+-0.0070 ! 12.3556+-0.0237 ! definitely 1.0042x slower TipOfTree Frequency AsmBench: bigfib.cpp 684.8822+-10.9716 ? 690.7452+-2.0853 ? cray.c 627.1713+-7.5147 622.0252+-2.9838 dry.c 646.8788+-17.8046 635.8453+-17.2113 might be 1.0174x faster FloatMM.c 915.1520+-0.5706 ? 915.1570+-0.3972 ? gcc-loops.cpp 5893.5286+-9.7096 5893.0811+-15.7535 n-body.c 1674.8983+-0.7053 ? 1675.1260+-0.9040 ? Quicksort.c 576.1181+-2.3870 564.4794+-9.5388 might be 1.0206x faster stepanov_container.cpp 4856.4435+-101.1813 4823.5172+-26.7501 Towers.c 399.8617+-0.5545 ? 402.3874+-2.4832 ? <geometric> 1121.4956+-4.3812 1116.8261+-5.3731 might be 1.0042x faster TipOfTree Frequency CompressionBench: huffman 81.6700+-0.2910 81.3780+-0.5075 arithmetic-simple 435.9021+-0.8128 ? 436.6113+-0.0977 ? arithmetic-precise 366.2816+-0.9817 ? 366.8237+-3.9121 ? arithmetic-complex-precise 366.0766+-3.4060 ? 366.4592+-3.1309 ? arithmetic-precise-order-0 450.7275+-1.1283 450.3326+-1.3063 arithmetic-precise-order-1 421.2679+-1.5494 ? 422.3979+-2.3416 ? arithmetic-precise-order-2 492.6092+-3.4004 492.3882+-2.4296 arithmetic-simple-order-1 500.1158+-0.8141 499.9897+-0.9909 arithmetic-simple-order-2 566.5458+-1.3574 565.6265+-1.5757 lz-string 429.7642+-6.6971 ^ 416.6518+-3.2410 ^ definitely 1.0315x faster <geometric> 374.5668+-1.0647 373.4095+-0.2755 might be 1.0031x faster TipOfTree Frequency Geomean of preferred means: <scaled-result> 78.6449+-0.1891 ? 78.7542+-0.1754 ? might be 1.0014x slower Comment on attachment 265685 [details]
the patch
r=me
Landed in http://trac.webkit.org/changeset/192529 |
Previously we used PNaN whenever we weren't sure about the execution frequency. This was partly to ensure that the execution frequencies "made sense". I think that this is the wrong approach, and it's definitely not profitable when you consider how B3 will use the frequencies. We should define frequency by what effects it has on optimization and correctness. I propose: - Frequency has no effect on correctness. It should be OK to pass any bag of double bits, including NaN, as the frequency. Using NaN could cause the compiler to make silly and unprofitable decisions, but it shouldn't break anything. - Frequency should be used for register allocation and layout. Higher frequency means that we want to prioritize that basic block. Some optimizations that frequency will affect: Register allocation: temporaries used in a block with higher frequency will get higher priority. Layout: when a block branches to multiple blocks, the block with highest frequency is likely to be scheduled closest, so that in the best case it will be the fall-through block.