On Aug 24, 2009, at 2:09 AM, Dr. David Kirkby wrote:

>
> Harald Schilly wrote:
>> On Aug 10, 9:58 am, Harald Schilly <harald.schi...@gmail.com> wrote:
>>> I was thinking about a benchmark website...
>>
>> ... which is now more concrete:
>> http://www.sagemath.org/tour-benchmarks.html
>> Please send me more examples if you have something else and review  
>> the
>> page.
>
> I think you should state what system was used for the comparison. Just
> because factorial is quick in Sage on system 'A', does not mean it  
> is so
> on system 'B'. Some processors are particularly good at floating  
> point,
> some are pretty dumb at it. (I believe the early Sun CoolThreads T1
> processor has only one FPU to share between all the cores).
>
>> I've CCed me to http://sagetrac.org/sage_trac/ticket/6808 so that I
>> can add more selected benchmarks and probably a link to the script
>> from the website once it is ready.
>> We could also think about expanding http://wiki.sagemath.org/ 
>> Benchmarks
>
> It would be great if someone could write that benchmark. If we can
> collect some results, I could ask Karl if he would link to the page.
>
> Just make sure he, and his web site are mentioned as the source  
> when one
> does
>
> help(benchmark)
>
> or whatever it's called.
>
> Since there are likely to be multiple benchmarks, it would be worth
> thinking about sensible names for them.

I remember a bunch of people did benchmarking stuff at Sage Days 10-- 
I don't know what became of it, but it might be worth digging up  
again (if just for ideas). The key idea was to make it as easy to  
write benchmarks as it was to write doctests, as opposed to  
elaborate, complicated things.

- Robert



--~--~---------~--~----~------------~-------~--~----~
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to 
sage-devel-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URLs: http://www.sagemath.org
-~----------~----~----~----~------~----~------~--~---

Reply via email to