Thanks for the reply!

This fix took care of the OpenSG issues..... now I'm running in to 
nvidia driver problems...... urgh. That's for another message board.

dan

Marcus Lindblom wrote:
> Daniel Sperka wrote:
>   
>> [snip]
>>
>> 1) Am I required to use OpenSG Threads? If so, am I limited to just two?
>>   
>>     
> Nope & nope.
>
> OpenSG uses some thread local data that need to be set in order to be 
> able to dereference FCPtr's. Failure to do so will result in crashes in 
> getAspect().
>
> You need to call osg::ExternalThread::get(0)->initialize(0) from 
> external threads that you want to make OpenSG calls in.
>
> The 0 argument to initalize set's the aspect for that thread. If you 
> manage locking yourself, you could use a single aspect for all of your 
> threads. If not, you could use several aspects (default is two, one 
> renders while the other mutates) and
>   
>> 2) If I limit the usage to FieldContainers to a single thread - i.e. all 
>> FC's are created and used in that one thread, will that work? Even if I 
>> use pthread_create? With a little more care I can easily segregate 
>> OpenSG FC's to my "command thread".....
>>   
>>     
> Yup. Be sure to call osginit() from that thread.
>   
>> 3) The new website has a nice general discussion of threading in OpenSG. 
>> It alludes to changes in OpenSG 2.0 - what is the status of those 
>> changes? Were there improvements in OpenSG 1.8?
>> 4) Can OpenSG be built as a single-threaded library? If so, will it buy 
>> me anything (like reduced overhead, improved speed, etc)? My app will 
>> remain multithreaded anyways..... just wondering.
>>   
>>     
> I don't think it's possible as such at the moment. It would by you some 
> reduced overhead in space/time (at least w.r.t 1.8) but there's some 
> ongoing work to fix it:
>
>  1. In 2.0 FCPtrs might be abolished totally which reduces the time 
> overhead on pointer access. (You'd have to map the pointer between 
> threads though)
>  2. Dynamic setting of number of aspects (before osginit). Setting it to 
> 1 will halve memory usage for most FieldContainers, compared to the 
> current default of 2 aspects.
>
> Someone from the core team ought to know more about the progress here...
>
> Cheers,
> /Marcus
>
>
> -------------------------------------------------------------------------
> This SF.net email is sponsored by DB2 Express
> Download DB2 Express C - the FREE version of DB2 express and take
> control of your XML. No limits. Just data. Click to get it now.
> http://sourceforge.net/powerbar/db2/
> _______________________________________________
> Opensg-users mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/opensg-users
>
>   

-- 
Daniel J. Sperka, Ph. D. 
UC Davis Center for Neuroscience


-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/
_______________________________________________
Opensg-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/opensg-users

Reply via email to