Carsten Neumann wrote: > Hello Pablo, > > Pablo Carneiro Elias wrote: >> Hi Dirk, thanks for the replies... >> >> well, as the problem occurs only in window, some frame time count might >> not be applying a minimum value check leading to some zero division when >> frame rate is too high .. Is there a possibility of you to put a minimum >> non-zero value check there? > > yes, I just committed such a fix (r1838). Not sure what to do about the > different timer resolution on windows though - I'll take a look at what > other systems use there and see if we can draw inspiration from that ;)
QueryPerformanceCounter gives micro or nanosecond resolution. It's what is typically used for hi-res timing on Windows. However, it comes with warnings that it might be inaccurate if the cpu-frequency changes. One could use a lower-resolution timer to detect drift beacause of that and re-request the resolution. It hasn't bitten me, yet, because I mostly run on desktops or with my laptop powered from the net. Cheers, /Marcus ------------------------------------------------------------------------------ _______________________________________________ Opensg-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/opensg-users
