On 1 May 2013 21:48, Maciej Fijalkowski <fijall@gmail.com> wrote:
On Wed, May 1, 2013 at 11:16 AM, Armin Rigo <arigo@tunes.org> wrote:
> Hi Michael,
>
> On Wed, May 1, 2013 at 3:54 AM, Michael Hudson-Doyle <micahel@gmail.com> wrote:
>>   File "/home/mwhudson/src/benchmarks/unladen_swallow/perf.py", line 170, in
>> TScore
>>     return (avg(sample1) - avg(sample2)) / math.sqrt(error * 2)
>> ZeroDivisionError: float division by zero
>
> The code could be more robust indeed :-(  It's because hexiom2 runs
> only once in --fast mode, so it ends up with error == exactly 0.0.
> Fixed in f7abffc04667.
>
>
> A bientôt,
>
> Armin.
> _______________________________________________
> pypy-dev mailing list
> pypy-dev@python.org
> http://mail.python.org/mailman/listinfo/pypy-dev

Btw, this is a benchmark run (without --fast) summary:

http://paste.pound-python.org/show/32751/

Thanks for that.
 
It seems CPython on ARM is kinda bad (despite the fact that our
assembler is bad too)

Yeah.  Maybe that's something we (Linaro) can do something about...

Cheers,
mwh