From: Alexander Lechner (alexander.lechner++at++vertigo-systems.de)
Date: 08/15/2003 05:37:42
Hi all!
We're running OpenGL Performer 3.02 (gcc2 and 3) on a Dual Xeon Suse 8.2
Linux System.
Besides some problems with the latest Nvidia 44.96 driver, which results
in a 3-5 seconds delay when starting up with with APP_CULLDRAW, we encounter
problems with pthreads.
Could someone point out the current issue with Performer and pthreads.
We have a pthread running separately (created by APP) to APP and DRAW.
The thread and DRAW sync each other with pthread_mutexes but after some
time the DRAW sigsegvs while accessing the mutex.
The pthread also allocates memory in the shared arena. Is this safe or not?
We tried to use a pfFlux but this resulted in performer warning us that
there is no current flux buffer for the frame and so returning the latest.
This is obvious as the provider pthread is slower than the DRAW and hence
cannot deliver data for the current frame.
In general: What is the best (preferred) way to run a thread/proc?
pthread or fork?
Thanks in advance!
Alex
-- Alexander.Lechner ++at++ vertigo-systems.de Engelbertstraße 30 | phone: +49-221-2405472 D-50674 Köln | fax: +49-221-34892616
This archive was generated by hypermail 2b29 : Fri Aug 15 2003 - 05:42:28 PDT