3.0 for Linux is 8% slower than 2.6.1 ?

[Bernard Devlin]

I've no idea why it would download as *.sh.  It was uploaded as *.rev.  The
stack doesn't just run under Linux, it runs under Vista (as I think I said
in my original post).  For some reason, I couldn't open it under OS X PPC
(although I didn't think endian issues affected Rev stacks).

It has a loop that repeats an exponential number of times, adding lines of
random numbers to a list.  It just records how long the script takes to run
each time.  There are a series of 'send in times' so that the results can be
averaged, and it can be run whilst nothing else is running.

I would think the script in the 'start' button is pretty self-explanatory.
If anyone else has any other benchmarks, I'd be grateful to see them.  Mine
seems pretty simple to me, and tests nothing more than a few functions,
looping and adding to lists.

It will only make any sense if you run the script on the same hardware,
using a version of Rev >= 3 and a version of Rev < 3.

Bernard

On Sun, Oct 5, 2008 at 6:58 PM, Colin Holgate <coiin at rcn.com> wrote:

>
> On Oct 5, 2008, at 11:38 AM, Wilhelm Sanke wrote:
>
>  Downloaded your test stack and could open it under WindowsXP, but could
>> not get it to run.
>>
>> Could you provide a version of your stack that is able to run tests for
>> platforms other than Linux, too?
>>
>
>
> The file downloaded named as perf-test-leg.rev.sh. You would want to
> rename it to perf-test-leg.rev. When you run it, it takes 80 seconds before
> anything happens, and over 6 minutes to do the whole test.
>
> On my Mac the numbers it showed were 21. I guess that lower is better?
>
>
>
> _______________________________________________
> use-revolution mailing list
> use-revolution at lists.runrev.com
> Please visit this url to subscribe, unsubscribe and manage your
> subscription preferences:
> http://lists.runrev.com/mailman/listinfo/use-revolution
>