2009/4/15 Eric Noulard
<address@hidden>
2009/4/13 Максим Гаммер <address@hidden>:
> Hello,
>
> Yeah, the tick2 function greatly improoved performance.
How much? (in terms of steps per seconds?)
Before: 15-20, after: 50. That includes remaking the program in order to use a separate thread. That's just what we need. :)
Nice to ear that however as suggested by Petr may
be you can use tick(INF,0) in order to use "standard" API.
> Considering other,
> commercial RTIs: it obviously works faster than PORTICO,
I think PORTICO is not to be considered as "commercial" since
it uses an open source license ().
> and in my opinion a little bit faster than PITCH.
Opinion is good but do you have figures, may be in terms
of number of steps per second using your test cases.
We had max of 30 steps per second on PITCH...
Note however that in order to be fair one should use
same API, thus may be tick(INF,0) and not tick2() which is not standard.
If your tests results may be publicly released I think we may all
be interested in reading those.
Ok, we'll publish them as soon as we have a chance. It's gonna be here:
http://cde/publications_en/ and maybe on
openscenegraph.org. in "screenshots" chapter.
Another thing - we use National Instruments Labview programs withing our simulation complex based on CERTICO. If you need, we could share a Labview source code of an example and a source code of a dll that is used in that example. A Linux version of that dll coming soon too... :)
--