Re: [abc] ecoop

> take some with different characteristics:
>
> product_lines-time2 : mostly itds
> 10.6_authorization-aspectj-v2: interesting cflow + around
> ants (with all aspects!): all sorts
>
> incidentally, looking at the numbers that I got on nikko running
> benches three times by hand, abc -O0 is *always* faster. are we
> 100% sure there isn't something that's taking a long time with
> the script, the ant task, or something else that we shouldn't
> be measuring?
The intended behavior of the script is for both of them to simply run
and time ant with the respective xml file. The -O0 version of the xml
file is based on the normal one, but is run through sed first to change
the -O option (but sed is not part of the measurement).

I'll try to find out what's going on now.

>
> -O
>
> On Mon, 20 Dec 2004, Sascha Kuzins wrote:
>
>
>>It's probably easiest if I run them since I have it all set up on
>>cardinal. Does anybody have suggestions which ones to pick?
>>
>>
>>
>>
>>Damien Sereni wrote:
>>
>>>>Sorry, but I don't have time to work on the paper today...
>>>>
>>>
>>>
>>>No problem, thanks for the help! I'll start working on this section
>>>now to put in your points.
>>>
>>>If anyone can help with running benches on larger examples (or just
>>>letting me know of good examples) that would be very helpful
>>>
>>>Cheers,
>>>D
>>>
>>
>
Received on Mon Dec 20 14:07:34 2004