[Shootout-list] bench mark test run time, max bad score
Bengt Kleberg
bengt.kleberg@ericsson.com
Wed, 29 Sep 2004 11:49:35 +0200
Brandon J. Van Every wrote:
...deleted
> Because I don't trust you, me, or anyone else to stay on the ball with
> this project for years and years and years. When's the last time the
> original Shootout was updated? I'd like to be able to relate the
> results of yesterday to the results of today.
again, why?
the resluts are relative performance of a certain versions of languages,
on a certain hardware/os platform.
why should that be possible to relate to other versions of (other?)
languages on other hardware/os platforms?
if i am willing to invest the kind of work neccessary to create a system
where this kind of relations are meaningful, it would seem to me that
much less work would be requiered to just rerun the tests again. on that
days hardware/os and that days versions of languages.
...deleted
> Who said g++ is no longer supported? Yet it has changed over the years.
why do you care about the realtive performance of g++ today versus g++
of yesterday? and if you are interested, why not add an old g++ version
to the shootout? we are working under the assumption that g++ is supported.
>>or have i missed something?
>
>
> I believe so.
this sounds very likely. i tend to miss things. but i have faith in
testing to cath errors. i test my ideas by asking others.
so, again, why do we want historical data that can be compared/related
to present data?
bengt