pdf2htmlEX/demo/demo11.page

<div class="pd w0 h0"><div id="pfb" class="pf" data-page-no="b"><div class="pc pcb"><img class="bi x2f y263 w4 h20" alt="" src="data:image/png;base64,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"/><div class="c xb7 y264 w23 h21"><div class="t m0 xb8 h5 y265 ff17 fs10 fc0 sc0 ls0 ws0">!&quot;</div><div class="t m0 xb8 h5 y266 ff17 fs10 fc0 sc0 ls0 ws0">#&quot;</div><div class="t m0 x6a h5 y52 ff17 fs10 fc0 sc0 ls0 ws0">$!&quot;</div><div class="t m0 x6a h5 y1b0 ff17 fs10 fc0 sc0 ls0 ws0">$#&quot;</div><div class="t m0 x6a h5 y175 ff17 fs10 fc0 sc0 ls0 ws0">%!&quot;</div><div class="t m0 x6a h5 y267 ff17 fs10 fc0 sc0 ls0 ws0">%#&quot;</div></div><div class="c xb y268 w24 h22"><div class="t m0 x6a h5 y269 ff17 fs10 fc0 sc0 ls0 ws0">&amp;<span class="_ _2"></span>&apos;<span class="_ _2"></span>()*+,&quot;</div></div><div class="c xb y63 w25 h22"><div class="t m0 x6a h5 y269 ff17 fs10 fc0 sc0 ls0 ws0">-./&quot;</div></div><div class="c xb y259 w26 h22"><div class="t m0 x6a h5 y269 ff17 fs10 fc0 sc0 ls0 ws0">01&quot;</div></div><div class="t m0 x2f h5 y161 ff1 fs3 fc0 sc0 ls0 ws0">Figure<span class="_ _3"> </span>10.<span class="_ _1"> </span><span class="ff2">Speedup<span class="_ _3"> </span>vs.<span class="_ _3"> </span>a<span class="_ _3"> </span>baseline<span class="_ _3"> </span>Jav<span class="_ _2"></span>aScript<span class="_ _3"> </span>interpreter<span class="_ _3"> </span>(SpiderMonkey)<span class="_ _3"> </span>for<span class="_ _3"> </span>our<span class="_ _3"> </span>trace-based<span class="_ _3"> </span>JIT<span class="_ _3"> </span>compiler<span class="_ _2"></span>,<span class="_ _3"> </span>Apple’<span class="_ _2"></span>s<span class="_ _3"> </span>SquirrelFish<span class="_ _3"> </span>Extreme</span></div><div class="t m0 x2f h5 y162 ff2 fs3 fc0 sc0 ls0 ws0">inline<span class="_ _5"> </span>threading<span class="_ _7"> </span>interpreter<span class="_ _5"> </span>and<span class="_ _7"> </span>Google’<span class="_ _2"></span>s<span class="_ _7"> </span>V8<span class="_ _5"> </span>JS<span class="_ _7"> </span>compiler<span class="_ _2"></span>.<span class="_ _5"> </span>Our<span class="_ _7"> </span>system<span class="_ _5"> </span>generates<span class="_ _7"> </span>particularly<span class="_ _5"> </span>ef<span class="_ _2"></span>ficient<span class="_ _7"> </span>code<span class="_ _5"> </span>for<span class="_ _7"> </span>programs<span class="_ _5"> </span>that<span class="_ _7"> </span>benefit<span class="_ _5"> </span>most<span class="_ _7"> </span>from</div><div class="t m0 x2f h5 y163 ff2 fs3 fc0 sc0 ls0 ws0">type<span class="_ _3"> </span>specialization,<span class="_ _5"> </span>which<span class="_ _5"> </span>includes<span class="_ _3"> </span>SunSpider<span class="_ _5"> </span>Benchmark<span class="_ _3"> </span>programs<span class="_ _5"> </span>that<span class="_ _3"> </span>perform<span class="_ _5"> </span>bit<span class="_ _3"> </span>manipulation.<span class="_ _5"> </span>W<span class="_ _b"></span>e<span class="_ _3"> </span>type-specialize<span class="_ _5"> </span>the<span class="_ _3"> </span>code<span class="_ _5"> </span>in<span class="_ _3"> </span>question</div><div class="t m0 x2f h5 y164 ff2 fs3 fc0 sc0 ls0 ws0">to<span class="_ _3"> </span>use<span class="_ _3"> </span>integer<span class="_ _3"> </span>arithmetic,<span class="_ _3"> </span>which<span class="_ _3"> </span>substantially<span class="_ _6"> </span>impro<span class="_ _2"></span>ves<span class="_ _3"> </span>performance.<span class="_ _3"> </span>For<span class="_ _3"> </span>one<span class="_ _3"> </span>of<span class="_ _3"> </span>the<span class="_ _3"> </span>benchmark<span class="_ _3"> </span>programs<span class="_ _3"> </span>we<span class="_ _6"> </span>e<span class="_ _2"></span>xecute<span class="_ _3"> </span>25<span class="_ _3"> </span>times<span class="_ _3"> </span>faster<span class="_ _3"> </span>than</div><div class="t m0 x2f h5 y165 ff2 fs3 fc0 sc0 ls0 ws0">the<span class="_ _5"> </span>SpiderMonkey<span class="_ _5"> </span>interpreter<span class="_ _2"></span>,<span class="_ _5"> </span>and<span class="_ _5"> </span>almost<span class="_ _5"> </span>5<span class="_ _5"> </span>times<span class="_ _3"> </span>f<span class="_ _2"></span>aster<span class="_ _5"> </span>than<span class="_ _5"> </span>V8<span class="_ _5"> </span>and<span class="_ _3"> </span>SFX.<span class="_ _5"> </span>F<span class="_ _2"></span>or<span class="_ _5"> </span>a<span class="_ _5"> </span>large<span class="_ _5"> </span>number<span class="_ _5"> </span>of<span class="_ _5"> </span>benchmarks<span class="_ _3"> </span>all<span class="_ _5"> </span>three<span class="_ _5"> </span>VMs<span class="_ _5"> </span>produce<span class="_ _5"> </span>similar</div><div class="t m0 x2f h5 y166 ff2 fs3 fc0 sc0 ls0 ws0">results.<span class="_ _5"> </span>W<span class="_ _b"></span>e<span class="_ _3"> </span>perform<span class="_ _5"> </span>w<span class="_ _2"></span>orst<span class="_ _5"> </span>on<span class="_ _5"> </span>benchmark<span class="_ _3"> </span>programs<span class="_ _5"> </span>that<span class="_ _5"> </span>we<span class="_ _5"> </span>do<span class="_ _5"> </span>not<span class="_ _5"> </span>trace<span class="_ _5"> </span>and<span class="_ _5"> </span>instead<span class="_ _3"> </span>f<span class="_ _2"></span>all<span class="_ _5"> </span>back<span class="_ _5"> </span>onto<span class="_ _5"> </span>the<span class="_ _3"> </span>interpreter<span class="_ _b"></span>.<span class="_ _5"> </span>This<span class="_ _5"> </span>includes<span class="_ _3"> </span>the<span class="_ _5"> </span>recursi<span class="_ _2"></span>v<span class="_ _2"></span>e</div><div class="t m0 x2f h5 y167 ff2 fs3 fc0 sc0 ls0 ws0">benchmarks<span class="_ _5"> </span><span class="ff7">access-binary-trees<span class="_ _5"> </span></span>and<span class="_ _3"> </span><span class="ff7">control-flow-recursive<span class="_ _2"></span><span class="ff2">,<span class="_ _5"> </span>for<span class="_ _5"> </span>which<span class="_ _5"> </span>we<span class="_ _3"> </span>currently<span class="_ _5"> </span>don’<span class="_ _2"></span>t<span class="_ _5"> </span>generate<span class="_ _5"> </span>any<span class="_ _5"> </span>nativ<span class="_ _2"></span>e<span class="_ _5"> </span>code.</span></span></div><div class="t m0 x2f h5 y16a ff2 fs3 fc0 sc0 ls0 ws0">In<span class="_ _3"> </span>particular<span class="_ _2"></span>,<span class="_ _5"> </span>the<span class="_ _3"> </span><span class="ff7">bitops<span class="_ _5"> </span></span>benchmarks<span class="_ _3"> </span>are<span class="_ _5"> </span>short<span class="_ _3"> </span>programs<span class="_ _3"> </span>that<span class="_ _5"> </span>per-</div><div class="t m0 x2f h5 y16b ff2 fs3 fc0 sc0 ls0 ws0">form<span class="_ _3"> </span>many<span class="_ _3"> </span>bitwise<span class="_ _3"> </span>operations,<span class="_ _5"> </span>so<span class="_ _3"> </span>TraceMonke<span class="_ _2"></span>y<span class="_ _3"> </span>can<span class="_ _3"> </span>co<span class="_ _2"></span>ver<span class="_ _3"> </span>the<span class="_ _3"> </span>en-</div><div class="t m0 x2f h5 y16c ff2 fs3 fc0 sc0 ls0 ws0">tire<span class="_ _5"> </span>program<span class="_ _5"> </span>with<span class="_ _5"> </span>1<span class="_ _5"> </span>or<span class="_ _5"> </span>2<span class="_ _3"> </span>traces<span class="_ _5"> </span>that<span class="_ _5"> </span>operate<span class="_ _5"> </span>on<span class="_ _5"> </span>inte<span class="_ _2"></span>gers.<span class="_ _5"> </span>T<span class="_ _2"></span>raceMon-</div><div class="t m0 x2f h5 y16d ff2 fs3 fc0 sc0 ls0 ws0">key<span class="_ _5"> </span>runs<span class="_ _5"> </span>all<span class="_ _3"> </span>the<span class="_ _5"> </span>other<span class="_ _3"> </span>progr<span class="_ _2"></span>ams<span class="_ _5"> </span>in<span class="_ _3"> </span>this<span class="_ _5"> </span>set<span class="_ _3"> </span>almost<span class="_ _5"> </span>entirely<span class="_ _5"> </span>as<span class="_ _3"> </span>nati<span class="_ _2"></span>v<span class="_ _2"></span>e</div><div class="t m0 x2f h5 y16e ff2 fs3 fc0 sc0 ls0 ws0">code.</div><div class="t m0 x34 h5 yd4 ff7 fs3 fc0 sc0 ls0 ws0">regexp-dna<span class="_ _1"> </span><span class="ff2">is<span class="_ _e"> </span>dominated<span class="_ _1"> </span>by<span class="_ _e"> </span>regular<span class="_ _1"> </span>expression<span class="_ _1"> </span>matching,</span></div><div class="t m0 x2f h5 yf4 ff2 fs3 fc0 sc0 ls0 ws0">which<span class="_ _5"> </span>is<span class="_ _5"> </span>implemented<span class="_ _3"> </span>in<span class="_ _5"> </span>all<span class="_ _5"> </span>3<span class="_ _5"> </span>VMs<span class="_ _5"> </span>by<span class="_ _5"> </span>a<span class="_ _3"> </span>special<span class="_ _5"> </span>re<span class="_ _2"></span>gular<span class="_ _5"> </span>expression</div><div class="t m0 x2f h5 yf5 ff2 fs3 fc0 sc0 ls0 ws0">compiler<span class="_ _2"></span>.<span class="_ _3"> </span>Thus,<span class="_ _3"> </span>performance<span class="_ _3"> </span>on<span class="_ _3"> </span>this<span class="_ _3"> </span>benchmark<span class="_ _3"> </span>has<span class="_ _3"> </span>little<span class="_ _6"> </span>relation</div><div class="t m0 x2f h5 yf6 ff2 fs3 fc0 sc0 ls0 ws0">to<span class="_ _5"> </span>the<span class="_ _5"> </span>trace<span class="_ _3"> </span>compilation<span class="_ _5"> </span>approach<span class="_ _5"> </span>discussed<span class="_ _5"> </span>in<span class="_ _5"> </span>this<span class="_ _5"> </span>paper<span class="_ _2"></span>.</div><div class="t m0 x34 h5 yf7 ff2 fs3 fc0 sc0 ls0 ws0">T<span class="_ _2"></span>raceMonke<span class="_ _2"></span>y’<span class="_ _2"></span>s<span class="_ _5"> </span>smaller<span class="_ _3"> </span>speedups<span class="_ _3"> </span>on<span class="_ _3"> </span>the<span class="_ _5"> </span>other<span class="_ _3"> </span>benchmarks<span class="_ _3"> </span>can</div><div class="t m0 x2f h5 y26a ff2 fs3 fc0 sc0 ls0 ws0">be<span class="_ _5"> </span>attributed<span class="_ _5"> </span>to<span class="_ _5"> </span>a<span class="_ _5"> </span>few<span class="_ _5"> </span>specific<span class="_ _5"> </span>causes:</div><div class="t m0 x36 h4 y26b ff3 fs2 fc0 sc0 ls0 ws0">•</div><div class="t m0 x37 h5 y26c ff2 fs3 fc0 sc0 ls0 ws0">The<span class="_ _e"> </span>implementation<span class="_ _e"> </span>does<span class="_ _e"> </span>not<span class="_ _e"> </span>currently<span class="_ _e"> </span>trace<span class="_ _e"> </span>recursion,<span class="_ _e"> </span>so</div><div class="t m0 x37 h5 y26d ff2 fs3 fc0 sc0 ls0 ws0">T<span class="_ _2"></span>raceMonke<span class="_ _2"></span>y<span class="_ _e"> </span>achie<span class="_ _2"></span>ves<span class="_ _1"> </span>a<span class="_ _e"> </span>small<span class="_ _1"> </span>speedup<span class="_ _e"> </span>or<span class="_ _e"> </span>no<span class="_ _e"> </span>speedup<span class="_ _1"> </span>on</div><div class="t m0 x37 h5 y26e ff2 fs3 fc0 sc0 ls0 ws0">benchmarks<span class="_ _1b"> </span>that<span class="_ _1b"> </span>use<span class="_ _e"> </span>recursion<span class="_ _1b"> </span>extensi<span class="_ _2"></span>vely:<span class="_ _1b"> </span><span class="ff7">3d-cube</span>,<span class="_ _e"> </span><span class="ff7">3d-</span></div><div class="t m0 x37 h5 y26f ff7 fs3 fc0 sc0 ls0 ws0">raytrace<span class="ff2">,<span class="_ _3"> </span></span>access-binary-trees<span class="ff2">,<span class="_ _6"> </span></span>string-tagcloud<span class="ff2">,<span class="_ _3"> </span>and</span></div><div class="t m0 x37 h5 y270 ff7 fs3 fc0 sc0 ls0 ws0">controlflow-recursive<span class="ff2">.</span></div><div class="t m0 x36 h4 y271 ff3 fs2 fc0 sc0 ls0 ws0">•</div><div class="t m0 x37 h5 yfe ff2 fs3 fc0 sc0 ls0 ws0">The<span class="_ _8"> </span>implementation<span class="_ _6"> </span>does<span class="_ _8"> </span>not<span class="_ _8"> </span>currently<span class="_ _8"> </span>trace<span class="_ _8"> </span><span class="ff7">eval<span class="_ _8"> </span></span>and<span class="_ _8"> </span>some</div><div class="t m0 x37 h5 yff ff2 fs3 fc0 sc0 ls0 ws0">other<span class="_ _1"> </span>functions<span class="_ _1"> </span>implemented<span class="_ _d"> </span>in<span class="_ _1"> </span>C.<span class="_ _1"> </span>Because<span class="_ _d"> </span><span class="ff7">date-format-</span></div><div class="t m0 x37 h5 y100 ff7 fs3 fc0 sc0 ls0 ws0">tofte<span class="_ _8"> </span><span class="ff2">and<span class="_ _8"> </span></span>date-format-xparb<span class="_ _8"> </span><span class="ff2">use<span class="_ _8"> </span>such<span class="_ _8"> </span>functions<span class="_ _8"> </span>in<span class="_ _8"> </span>their</span></div><div class="t m0 x37 h5 y101 ff2 fs3 fc0 sc0 ls0 ws0">main<span class="_ _5"> </span>loops,<span class="_ _5"> </span>we<span class="_ _3"> </span>do<span class="_ _5"> </span>not<span class="_ _5"> </span>trace<span class="_ _5"> </span>them.</div><div class="t m0 x36 h4 y272 ff3 fs2 fc0 sc0 ls0 ws0">•</div><div class="t m0 x37 h5 y98 ff2 fs3 fc0 sc0 ls0 ws0">The<span class="_ _8"> </span>implementation<span class="_ _8"> </span>does<span class="_ _d"> </span>not<span class="_ _8"> </span>currently<span class="_ _8"> </span>trace<span class="_ _8"> </span>through<span class="_ _d"> </span>re<span class="_ _2"></span>gular</div><div class="t m0 x37 h5 y99 ff2 fs3 fc0 sc0 ls0 ws0">expression<span class="_ _6"> </span><span class="ff7">replace<span class="_ _d"> </span></span>operations.<span class="_ _6"> </span>The<span class="_ _d"> </span>replace<span class="_ _6"> </span>function<span class="_ _d"> </span>can<span class="_ _6"> </span>be</div><div class="t m0 x37 h5 y9a ff2 fs3 fc0 sc0 ls0 ws0">passed<span class="_ _3"> </span>a<span class="_ _5"> </span>function<span class="_ _5"> </span>object<span class="_ _3"> </span>used<span class="_ _5"> </span>to<span class="_ _3"> </span>compute<span class="_ _5"> </span>the<span class="_ _3"> </span>replacement<span class="_ _5"> </span>text.</div><div class="t m0 x37 h5 y9b ff2 fs3 fc0 sc0 ls0 ws0">Our<span class="_ _6"> </span>implementation<span class="_ _8"> </span>currently<span class="_ _6"> </span>does<span class="_ _8"> </span>not<span class="_ _8"> </span>trace<span class="_ _6"> </span>functions<span class="_ _8"> </span>called</div><div class="t m0 x37 h5 y9c ff2 fs3 fc0 sc0 ls0 ws0">as<span class="_ _5"> </span>replace<span class="_ _5"> </span>functions.<span class="_ _5"> </span>The<span class="_ _5"> </span>run<span class="_ _5"> </span>time<span class="_ _5"> </span>of<span class="_ _5"> </span><span class="ff7">string-unpack-code<span class="_ _5"> </span></span>is</div><div class="t m0 x37 h5 y9d ff2 fs3 fc0 sc0 ls0 ws0">dominated<span class="_ _5"> </span>by<span class="_ _5"> </span>such<span class="_ _3"> </span>a<span class="_ _5"> </span><span class="ff7">replace<span class="_ _5"> </span></span>call.</div><div class="t m0 x8 h4 y273 ff3 fs2 fc0 sc0 ls0 ws0">•</div><div class="t m0 xb2 h5 y16a ff2 fs3 fc0 sc0 ls0 ws0">T<span class="_ _b"></span>wo<span class="_ _8"> </span>programs<span class="_ _d"> </span>trace<span class="_ _8"> </span>well,<span class="_ _d"> </span>b<span class="_ _2"></span>ut<span class="_ _8"> </span>have<span class="_ _6"> </span>a<span class="_ _d"> </span>long<span class="_ _8"> </span>compilation<span class="_ _d"> </span>time.</div><div class="t m0 xb2 h5 y16b ff7 fs3 fc0 sc0 ls0 ws0">access-nbody<span class="_ _7"> </span><span class="ff2">forms<span class="_ _5"> </span>a<span class="_ _7"> </span>large<span class="_ _7"> </span>number<span class="_ _7"> </span>of<span class="_ _5"> </span>traces<span class="_ _7"> </span>(81).<span class="_ _7"> </span></span>crypto-md5</div><div class="t m0 xb2 h5 y16c ff2 fs3 fc0 sc0 ls0 ws0">forms<span class="_ _3"> </span>one<span class="_ _3"> </span>v<span class="_ _2"></span>ery<span class="_ _3"> </span>long<span class="_ _3"> </span>trace.<span class="_ _3"> </span>W<span class="_ _b"></span>e<span class="_ _3"> </span>e<span class="_ _2"></span>xpect<span class="_ _3"> </span>to<span class="_ _3"> </span>impro<span class="_ _2"></span>ve<span class="_ _3"> </span>performance</div><div class="t m0 xb2 h5 y16d ff2 fs3 fc0 sc0 ls0 ws0">on<span class="_ _3"> </span>this<span class="_ _5"> </span>programs<span class="_ _3"> </span>by<span class="_ _5"> </span>improving<span class="_ _5"> </span>the<span class="_ _3"> </span>compilation<span class="_ _5"> </span>speed<span class="_ _3"> </span>of<span class="_ _5"> </span>nano-</div><div class="t m0 xb2 h5 y16e ff2 fs3 fc0 sc0 ls0 ws0">jit.</div><div class="t m0 x8 h4 y23 ff3 fs2 fc0 sc0 ls0 ws0">•</div><div class="t m0 xb2 h5 y274 ff2 fs3 fc0 sc0 ls0 ws0">Some<span class="_ _d"> </span>programs<span class="_ _d"> </span>trace<span class="_ _1"> </span>v<span class="_ _2"></span>ery<span class="_ _d"> </span>well,<span class="_ _d"> </span>and<span class="_ _1"> </span>speed<span class="_ _d"> </span>up<span class="_ _d"> </span>compared<span class="_ _d"> </span>to</div><div class="t m0 xb2 h5 y170 ff2 fs3 fc0 sc0 ls0 ws0">the<span class="_ _6"> </span>interpreter<span class="_ _2"></span>,<span class="_ _6"> </span>but<span class="_ _6"> </span>are<span class="_ _6"> </span>not<span class="_ _6"> </span>as<span class="_ _6"> </span>fast<span class="_ _6"> </span>as<span class="_ _6"> </span>SFX<span class="_ _6"> </span>and/or<span class="_ _8"> </span>V8,<span class="_ _6"> </span>namely</div><div class="t m0 xb2 h5 y275 ff7 fs3 fc0 sc0 ls0 ws0">bitops-bits-in-byte<span class="ff2">,<span class="_ _c"> </span></span>bitops-nsieve-bits<span class="ff2">,<span class="_ _29"> </span></span>access-</div><div class="t m0 xb2 h5 y276 ff7 fs3 fc0 sc0 ls0 ws0">fannkuch<span class="ff2">,<span class="_ _3"> </span></span>access-nsieve<span class="ff2">,<span class="_ _6"> </span>and<span class="_ _3"> </span></span>crypto-aes<span class="ff2">.<span class="_ _3"> </span>The<span class="_ _3"> </span>reason<span class="_ _6"> </span>is</span></div><div class="t m0 xb2 h5 y277 ff2 fs3 fc0 sc0 ls0 ws0">not<span class="_ _d"> </span>clear,<span class="_ _d"> </span>b<span class="_ _2"></span>ut<span class="_ _d"> </span>all<span class="_ _1"> </span>of<span class="_ _d"> </span>these<span class="_ _d"> </span>programs<span class="_ _d"> </span>have<span class="_ _d"> </span>nested<span class="_ _d"> </span>loops<span class="_ _d"> </span>with</div><div class="t m0 xb2 h5 y278 ff2 fs3 fc0 sc0 ls0 ws0">small<span class="_ _5"> </span>bodies,<span class="_ _3"> </span>so<span class="_ _5"> </span>we<span class="_ _5"> </span>suspect<span class="_ _5"> </span>that<span class="_ _3"> </span>the<span class="_ _5"> </span>implementation<span class="_ _5"> </span>has<span class="_ _3"> </span>a<span class="_ _5"> </span>rela-</div><div class="t m0 xb2 h5 y279 ff2 fs3 fc0 sc0 ls0 ws0">tiv<span class="_ _2"></span>ely<span class="_ _5"> </span>high<span class="_ _5"> </span>cost<span class="_ _5"> </span>for<span class="_ _5"> </span>calling<span class="_ _5"> </span>nested<span class="_ _5"> </span>traces.<span class="_ _5"> </span><span class="ff7">string-fasta<span class="_ _5"> </span></span>traces</div><div class="t m0 xb2 h5 y27a ff2 fs3 fc0 sc0 ls0 ws0">well,<span class="_ _5"> </span>b<span class="_ _2"></span>ut<span class="_ _5"> </span>its<span class="_ _5"> </span>run<span class="_ _7"> </span>time<span class="_ _5"> </span>is<span class="_ _5"> </span>dominated<span class="_ _5"> </span>by<span class="_ _5"> </span>string<span class="_ _5"> </span>processing<span class="_ _7"> </span>builtins,</div><div class="t m0 xb2 h5 y27b ff2 fs3 fc0 sc0 ls0 ws0">which<span class="_ _5"> </span>are<span class="_ _3"> </span>unaf<span class="_ _2"></span>fected<span class="_ _5"> </span>by<span class="_ _5"> </span>tracing<span class="_ _5"> </span>and<span class="_ _3"> </span>seem<span class="_ _5"> </span>to<span class="_ _5"> </span>be<span class="_ _3"> </span>less<span class="_ _5"> </span>ef<span class="_ _2"></span>ficient<span class="_ _5"> </span>in</div><div class="t m0 xb2 h5 y27c ff2 fs3 fc0 sc0 ls0 ws0">SpiderMonkey<span class="_ _5"> </span>than<span class="_ _5"> </span>in<span class="_ _5"> </span>the<span class="_ _5"> </span>two<span class="_ _5"> </span>other<span class="_ _5"> </span>VMs.</div><div class="t m0 x33 h5 yd2 ff1 fs3 fc0 sc0 ls0 ws0">Detailed<span class="_ _5"> </span>perf<span class="_ _2"></span>ormance<span class="_ _5"> </span>metrics.<span class="_ _5"> </span><span class="ff2">In<span class="_ _5"> </span>Figure<span class="_ _5"> </span>11<span class="_ _7"> </span>we<span class="_ _5"> </span>show<span class="_ _5"> </span>the<span class="_ _5"> </span>frac-</span></div><div class="t m0 x32 h5 y205 ff2 fs3 fc0 sc0 ls0 ws0">tion<span class="_ _5"> </span>of<span class="_ _3"> </span>instructions<span class="_ _5"> </span>interpreted<span class="_ _5"> </span>and<span class="_ _5"> </span>the<span class="_ _3"> </span>fraction<span class="_ _5"> </span>of<span class="_ _5"> </span>instructions<span class="_ _5"> </span>exe-</div><div class="t m0 x32 h5 y94 ff2 fs3 fc0 sc0 ls0 ws0">cuted<span class="_ _5"> </span>as<span class="_ _5"> </span>nati<span class="_ _2"></span>v<span class="_ _2"></span>e<span class="_ _5"> </span>code.<span class="_ _5"> </span>This<span class="_ _7"> </span>figure<span class="_ _5"> </span>shows<span class="_ _5"> </span>that<span class="_ _7"> </span>for<span class="_ _5"> </span>many<span class="_ _5"> </span>programs,<span class="_ _5"> </span>we</div><div class="t m0 x32 h5 y95 ff2 fs3 fc0 sc0 ls0 ws0">are<span class="_ _5"> </span>able<span class="_ _5"> </span>to<span class="_ _3"> </span>e<span class="_ _2"></span>x<span class="_ _2"></span>ecute<span class="_ _5"> </span>almost<span class="_ _5"> </span>all<span class="_ _5"> </span>the<span class="_ _3"> </span>code<span class="_ _5"> </span>nati<span class="_ _2"></span>v<span class="_ _2"></span>ely<span class="_ _2"></span>.</div><div class="t m0 x33 h5 y96 ff2 fs3 fc0 sc0 ls0 ws0">Figure<span class="_ _3"> </span>12<span class="_ _5"> </span>breaks<span class="_ _3"> </span>do<span class="_ _2"></span>wn<span class="_ _3"> </span>the<span class="_ _5"> </span>total<span class="_ _3"> </span>e<span class="_ _2"></span>xecution<span class="_ _5"> </span>time<span class="_ _3"> </span>into<span class="_ _3"> </span>four<span class="_ _5"> </span>activ-</div><div class="t m0 x32 h5 y97 ff2 fs3 fc0 sc0 ls0 ws0">ities:<span class="_ _3"> </span>interpreting<span class="_ _6"> </span>bytecodes<span class="_ _3"> </span>while<span class="_ _6"> </span>not<span class="_ _3"> </span>recording,<span class="_ _6"> </span>recording<span class="_ _6"> </span>traces</div><div class="t m0 x32 h5 y98 ff2 fs3 fc0 sc0 ls0 ws0">(including<span class="_ _8"> </span>time<span class="_ _8"> </span>taken<span class="_ _8"> </span>to<span class="_ _8"> </span>interpret<span class="_ _8"> </span>the<span class="_ _d"> </span>recorded<span class="_ _6"> </span>trace),<span class="_ _d"> </span>compiling</div><div class="t m0 x32 h5 y99 ff2 fs3 fc0 sc0 ls0 ws0">traces<span class="_ _5"> </span>to<span class="_ _5"> </span>native<span class="_ _5"> </span>code,<span class="_ _5"> </span>and<span class="_ _5"> </span>executing<span class="_ _5"> </span>nati<span class="_ _2"></span>ve<span class="_ _5"> </span>code<span class="_ _5"> </span>traces.</div><div class="t m0 x33 h5 y9a ff2 fs3 fc0 sc0 ls0 ws0">These<span class="_ _8"> </span>detailed<span class="_ _d"> </span>metrics<span class="_ _8"> </span>allow<span class="_ _8"> </span>us<span class="_ _8"> </span>to<span class="_ _d"> </span>estimate<span class="_ _8"> </span>parameters<span class="_ _8"> </span>for<span class="_ _d"> </span>a</div><div class="t m0 x32 h5 y9b ff2 fs3 fc0 sc0 ls0 ws0">simple<span class="_ _6"> </span>model<span class="_ _6"> </span>of<span class="_ _6"> </span>tracing<span class="_ _6"> </span>performance.<span class="_ _6"> </span>These<span class="_ _8"> </span>estimates<span class="_ _6"> </span>should<span class="_ _6"> </span>be</div><div class="t m0 x32 h5 y9c ff2 fs3 fc0 sc0 ls0 ws0">considered<span class="_ _8"> </span>very<span class="_ _8"> </span>rough,<span class="_ _8"> </span>as<span class="_ _8"> </span>the<span class="_ _8"> </span>values<span class="_ _8"> </span>observed<span class="_ _8"> </span>on<span class="_ _8"> </span>the<span class="_ _8"> </span>individual</div><div class="t m0 x32 h5 y9d ff2 fs3 fc0 sc0 ls0 ws0">benchmarks<span class="_ _d"> </span>ha<span class="_ _2"></span>ve<span class="_ _8"> </span>large<span class="_ _d"> </span>standard<span class="_ _8"> </span>deviations<span class="_ _d"> </span>(on<span class="_ _8"> </span>the<span class="_ _d"> </span>order<span class="_ _d"> </span>of<span class="_ _8"> </span>the</div></div><div class="pi" data-data='{"ctm":[1.673203,0.000000,0.000000,1.673203,0.000000,0.000000]}'></div></div></div>