I accidentally left out the telco benchmark, which is bad since cdecimal makes it just scream on Python 3.3 (and I verified with Python 3.2 that this is an actual speedup and not some silly screw-up like I initially had with spectral_norm):<div>

<br></div><div><div>### telco ###</div><div>Min: 0.897108 -&gt; 0.016880: 53.15x faster</div><div>Avg: 0.899742 -&gt; 0.017443: 51.58x faster</div><div>Significant (t=692.55)</div><div>Stddev: 0.00283 -&gt; 0.00032: 8.8470x smaller</div>

<br><div class="gmail_quote">On Sun, Sep 30, 2012 at 7:12 PM, Brett Cannon <span dir="ltr">&lt;<a href="mailto:brett@python.org" target="_blank">brett@python.org</a>&gt;</span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">

I am presenting the talk &quot;Python 3.3: Trust Me, It&#39;s Better Than 2.7&quot; as PyCon Argentina and Brasil (and US if they accept the talk). As part of that talk I need to be able to benchmark Python 3.3 against 2.7 (both from tip) using the unladen benchmarks (which now include benchmarks from PyPy that can be relatively easily ported to Python 3).<div>


<br></div><div>To make sure the unladen benchmarks run fine against Python 3.3, I did a fast run of the benchmarks. I figured people might be interested in the quick-and-dirty results on my 2 GHz Intel Core i7 MacBook Pro w/ 8 GB RAM and no attempt to control for performance beyond not actively browsing the web. As I said, quick-and-dirty and not authoritative; all done just to make sure all the benchmarks could run to completion (including the django, html5lib, and genshi benchmarks which are only on my laptop ATM until those projects cut a release with official Python 3 support).</div>


<div><br></div><div>One thing to keep in mind is that many benchmarks use a raw str for things, so the benchmarks often compare Python 2.7 str vs. Python 3.3 str (i.e. str vs. unicode). While this might seem unfair, this is what real-world comparisons in performance will be from users so it&#39;s an (somewhat unfair) comparison that we just have to live with. I might take the time to try to make some tests run under both raw strings and unicode so both comparisons are available.</div>


<div><br></div><div>If you care about helping out with the benchmarks (e.g. helping spot where the iteration counts should be higher, etc.) then head over to the speed@ mailing list.</div><div><br></div><div><br></div><div>


<br></div><div>&gt; python3 perf.py -T --basedir ../benchmarks -f -b py3k ../cpython/builds/2.7-wide/bin/python ../cpython/builds/3.3/bin/python3.3 </div><div><br></div><div>... output about the command line for the benchmarks ...</div>


<div><br></div><div><div>### 2to3 ###</div><div>0.785234 -&gt; 0.722169: 1.09x faster</div><div><br></div><div>### call_method ###</div><div>Min: 0.491433 -&gt; 0.414841: 1.18x faster</div><div>Avg: 0.493640 -&gt; 0.416564: 1.19x faster</div>


<div>Significant (t=127.21)</div><div>Stddev: 0.00170 -&gt; 0.00162: 1.0513x smaller</div><div><br></div><div>### call_method_slots ###</div><div>Min: 0.492749 -&gt; 0.416280: 1.18x faster</div><div>Avg: 0.497888 -&gt; 0.419275: 1.19x faster</div>


<div>Significant (t=61.72)</div><div>Stddev: 0.00433 -&gt; 0.00237: 1.8304x smaller</div><div><br></div><div>### call_method_unknown ###</div><div>Min: 0.575536 -&gt; 0.427234: 1.35x faster</div><div>Avg: 0.577286 -&gt; 0.433428: 1.33x faster</div>


<div>Significant (t=66.09)</div><div>Stddev: 0.00117 -&gt; 0.00835: 7.1621x larger</div><div><br></div><div>### call_simple ###</div><div>Min: 0.413011 -&gt; 0.338923: 1.22x faster</div><div>Avg: 0.415862 -&gt; 0.340699: 1.22x faster</div>


<div>Significant (t=111.94)</div><div>Stddev: 0.00223 -&gt; 0.00134: 1.6616x smaller</div><div><br></div><div>### chaos ###</div><div>Min: 0.375286 -&gt; 0.435456: 1.16x slower</div><div>Avg: 0.382798 -&gt; 0.459515: 1.20x slower</div>


<div>Significant (t=-5.01)</div><div>Stddev: 0.01116 -&gt; 0.03234: 2.8980x larger</div><div><br></div><div>### fastpickle ###</div><div>Min: 0.853560 -&gt; 0.770580: 1.11x faster</div><div>Avg: 0.879498 -&gt; 0.776249: 1.13x faster</div>


<div>Significant (t=8.24)</div><div>Stddev: 0.02771 -&gt; 0.00407: 6.7995x smaller</div><div><br></div><div>### float ###</div><div>Min: 0.476596 -&gt; 0.391101: 1.22x faster</div><div>Avg: 0.486164 -&gt; 0.411553: 1.18x faster</div>


<div>Significant (t=9.07)</div><div>Stddev: 0.01049 -&gt; 0.01511: 1.4411x larger</div><div><br></div><div>### formatted_logging ###</div><div>Min: 0.346703 -&gt; 0.451643: 1.30x slower</div><div>Avg: 0.351218 -&gt; 0.454626: 1.29x slower</div>


<div>Significant (t=-51.50)</div><div>Stddev: 0.00376 -&gt; 0.00246: 1.5265x smaller</div><div><br></div><div>### genshi ###</div><div>Min: 0.275107 -&gt; 0.294309: 1.07x slower</div><div>Avg: 0.287433 -&gt; 0.299026: 1.04x slower</div>


<div>Significant (t=-3.82)</div><div>Stddev: 0.01077 -&gt; 0.00467: 2.3044x smaller</div><div><br></div><div>### go ###</div><div>Min: 0.719160 -&gt; 0.781042: 1.09x slower</div><div>Avg: 0.729322 -&gt; 0.798135: 1.09x slower</div>


<div>Significant (t=-8.54)</div><div>Stddev: 0.01300 -&gt; 0.01248: 1.0415x smaller</div><div><br></div><div>### hexiom2 ###</div><div>203.842661 -&gt; 187.107363: 1.09x faster</div><div><br></div><div>### iterative_count ###</div>


<div>Min: 0.145088 -&gt; 0.153285: 1.06x slower</div><div>Avg: 0.146369 -&gt; 0.154425: 1.06x slower</div><div>Significant (t=-9.21)</div><div>Stddev: 0.00134 -&gt; 0.00142: 1.0569x larger</div><div><br></div><div>### json_dump_v2 ###</div>


<div>Min: 3.512367 -&gt; 4.040813: 1.15x slower</div><div>Avg: 3.521879 -&gt; 4.057966: 1.15x slower</div><div>Significant (t=-64.29)</div><div>Stddev: 0.01071 -&gt; 0.01526: 1.4247x larger</div><div><br></div><div>### json_load ###</div>


<div>Min: 1.024560 -&gt; 0.642353: 1.60x faster</div><div>Avg: 1.025255 -&gt; 0.644000: 1.59x faster</div><div>Significant (t=426.59)</div><div>Stddev: 0.00049 -&gt; 0.00194: 3.9240x larger</div><div><br></div><div>### mako_v2 ###</div>


<div>Min: 0.137584 -&gt; 0.287701: 2.09x slower</div><div>Avg: 0.140620 -&gt; 0.293204: 2.09x slower</div><div>Significant (t=-296.14)</div><div>Stddev: 0.00243 -&gt; 0.00272: 1.1195x larger</div><div><br></div><div>### meteor_contest ###</div>


<div>Min: 0.284739 -&gt; 0.254285: 1.12x faster</div><div>Avg: 0.286174 -&gt; 0.255323: 1.12x faster</div><div>Significant (t=38.02)</div><div>Stddev: 0.00124 -&gt; 0.00133: 1.0725x larger</div><div><br></div><div>### nbody ###</div>


<div>Min: 0.491416 -&gt; 0.336127: 1.46x faster</div><div>Avg: 0.493339 -&gt; 0.337467: 1.46x faster</div><div>Significant (t=185.50)</div><div>Stddev: 0.00164 -&gt; 0.00092: 1.7927x smaller</div><div><br></div><div>### normal_startup ###</div>


<div>Min: 0.639285 -&gt; 0.898157: 1.40x slower</div><div>Avg: 0.645513 -&gt; 0.901586: 1.40x slower</div><div>Significant (t=-90.10)</div><div>Stddev: 0.00575 -&gt; 0.00270: 2.1309x smaller</div><div><br></div><div>### nqueens ###</div>


<div>Min: 0.399351 -&gt; 0.429575: 1.08x slower</div><div>Avg: 0.403643 -&gt; 0.430284: 1.07x slower</div><div>Significant (t=-9.83)</div><div>Stddev: 0.00603 -&gt; 0.00053: 11.3092x smaller</div><div><br></div><div>### pathlib ###</div>


<div>Min: 0.137462 -&gt; 0.170506: 1.24x slower</div><div>Avg: 0.145370 -&gt; 0.172849: 1.19x slower</div><div>Significant (t=-11.09)</div><div>Stddev: 0.01232 -&gt; 0.00128: 9.6403x smaller</div><div><br></div><div>### pidigits ###</div>


<div>Min: 0.400265 -&gt; 0.379307: 1.06x faster</div><div>Avg: 0.401755 -&gt; 0.381171: 1.05x faster</div><div>Significant (t=14.65)</div><div>Stddev: 0.00259 -&gt; 0.00178: 1.4496x smaller</div><div><br></div><div>### raytrace ###</div>


<div>Min: 1.770596 -&gt; 1.958350: 1.11x slower</div><div>Avg: 1.773719 -&gt; 1.968401: 1.11x slower</div><div>Significant (t=-44.19)</div><div>Stddev: 0.00439 -&gt; 0.00882: 2.0099x larger</div><div><br></div><div>### regex_effbot ###</div>


<div>Min: 0.076566 -&gt; 0.098124: 1.28x slower</div><div>Avg: 0.077491 -&gt; 0.098696: 1.27x slower</div><div>Significant (t=-54.47)</div><div>Stddev: 0.00052 -&gt; 0.00069: 1.3227x larger</div><div><br></div><div>### regex_v8 ###</div>


<div>Min: 0.091530 -&gt; 0.109116: 1.19x slower</div><div>Avg: 0.092308 -&gt; 0.113627: 1.23x slower</div><div>Significant (t=-5.72)</div><div>Stddev: 0.00088 -&gt; 0.00829: 9.4271x larger</div><div><br></div><div>### richards ###</div>


<div>Min: 0.257974 -&gt; 0.232134: 1.11x faster</div><div>Avg: 0.259248 -&gt; 0.234325: 1.11x faster</div><div>Significant (t=23.80)</div><div>Stddev: 0.00144 -&gt; 0.00185: 1.2823x larger</div><div><br></div><div>### simple_logging ###</div>


<div>Min: 0.326569 -&gt; 0.416797: 1.28x slower</div><div>Avg: 0.331694 -&gt; 0.418844: 1.26x slower</div><div>Significant (t=-36.32)</div><div>Stddev: 0.00523 -&gt; 0.00122: 4.3004x smaller</div><div><br></div><div><div>


### spectral_norm ###</div><div>Min: 0.483011 -&gt; 0.741558: 1.54x slower</div><div>Avg: 0.487128 -&gt; 0.749741: 1.54x slower</div><div>Significant (t=-57.40)</div><div>Stddev: 0.00512 -&gt; 0.00886: 1.7299x larger</div>


</div><div><br></div><div>### startup_nosite ###</div><div>Min: 0.220444 -&gt; 0.374521: 1.70x slower</div><div>Avg: 0.222773 -&gt; 0.376785: 1.69x slower</div><div>Significant (t=-176.17)</div><div>Stddev: 0.00166 -&gt; 0.00221: 1.3331x larger</div>


<div><br></div><div>### threaded_count ###</div><div>Min: 0.171352 -&gt; 0.151892: 1.13x faster</div><div>Avg: 0.183180 -&gt; 0.153634: 1.19x faster</div><div>Significant (t=8.12)</div><div>Stddev: 0.00801 -&gt; 0.00140: 5.7241x smaller</div>


<div><br></div><div>### unpack_sequence ###</div><div>Min: 0.000075 -&gt; 0.000061: 1.23x faster</div><div>Avg: 0.000101 -&gt; 0.000065: 1.54x faster</div><div>Significant (t=206.90)</div><div>Stddev: 0.00001 -&gt; 0.00000: 3.2374x smaller</div>


<div><br></div><div>The following not significant results are hidden, use -v to show them:</div><div>chameleon, fannkuch, fastunpickle, regex_compile, silent_logging</div></div><div><br></div><div><div>### django ###</div>


<div>Min: 0.868956 -&gt; 0.894571: 1.03x slower</div><div>Avg: 0.873620 -&gt; 0.905274: 1.04x slower</div><div>Significant (t=-6.97)</div><div>Stddev: 0.00313 -&gt; 0.00966: 3.0912x larger</div><div><br></div><div>### genshi ###</div>


<div>Min: 0.269615 -&gt; 0.286348: 1.06x slower</div><div>Avg: 0.272206 -&gt; 0.290708: 1.07x slower</div><div>Significant (t=-12.29)</div><div>Stddev: 0.00253 -&gt; 0.00526: 2.0793x larger</div><div><br></div><div>### html5lib ###</div>


<div>12.279808 -&gt; 11.862586: 1.04x faster</div></div>
</blockquote></div><br></div>