Very Computer

Hi NG,

two questions:

1) how can determine the size of the window decoration of
a glut window? my window is slightly displaced when i re-init it....

2) i want to put a fps-throttle in my program so i added the following
code to the end of my display func which is registered via
glutDisplayFunc():

...
        ltime.tv_sec=ttime.tv_sec;
        ltime.tv_usec=ttime.tv_usec;
        gettimeofday(&ttime,NULL);
        if (ttime.tv_sec-ltime.tv_sec)
sec=(1000000-ltime.tv_usec)+ttime.tv_usec;
        else                            sec=ttime.tv_usec-ltime.tv_usec;

        if(sec<40000)
        {       sec=((40000-sec)*1000);
                wtime.tv_sec=0; wtime.tv_nsec=sec;
                nanosleep(&wtime,NULL);
        }
        glutSwapBuffers();

Quote:}

while

        struct timeval ltime,ttime;
        struct timespec wtime;
        long sec;

and a
        gettimeofday(&ttime,NULL);

is done in the normal init of the program so the time struct is not
empty

but it differs about 7 milliseconds (31fps instead of 25 fps)
i tried usleep() and select(0,NULL,NULL,NULL,&timeout) too.. same
effect
i also tried to do it via a glut timer callback but its too slow then
(20fps)

gettimeofday seems to give me correct values (~15000microsecs/frame)

i have no idea what i could do ...

thanks in advance

For greater accuracy I would recommend using timeGetTime() or the high
performance counters (QueryPerformanceCounter(), QueryPerformanceFreq()).
Also - I haven't checked whether this is relevant - but unsure your maths is
entirely float or integer as necessary or you may experience rounding
errors.

Hope that helps
Ian

Quote:> Hi NG,

> two questions:

> 1) how can determine the size of the window decoration of
> a glut window? my window is slightly displaced when i re-init it....

> 2) i want to put a fps-throttle in my program so i added the following
> code to the end of my display func which is registered via
> glutDisplayFunc():

> ...
>         ltime.tv_sec=ttime.tv_sec;
>         ltime.tv_usec=ttime.tv_usec;
>         gettimeofday(&ttime,NULL);
>         if (ttime.tv_sec-ltime.tv_sec)
> sec=(1000000-ltime.tv_usec)+ttime.tv_usec;
>         else                            sec=ttime.tv_usec-ltime.tv_usec;

>         if(sec<40000)
>         {       sec=((40000-sec)*1000);
>                 wtime.tv_sec=0; wtime.tv_nsec=sec;
>                 nanosleep(&wtime,NULL);
>         }
>         glutSwapBuffers();
> }

> while

>         struct timeval ltime,ttime;
>         struct timespec wtime;
>         long sec;

> and a
>         gettimeofday(&ttime,NULL);

> is done in the normal init of the program so the time struct is not
> empty

> but it differs about 7 milliseconds (31fps instead of 25 fps)
> i tried usleep() and select(0,NULL,NULL,NULL,&timeout) too.. same
> effect
> i also tried to do it via a glut timer callback but its too slow then
> (20fps)

> gettimeofday seems to give me correct values (~15000microsecs/frame)

> i have no idea what i could do ...

> thanks in advance

---
This email has been checked for viruses
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.265 / Virus Database: 137 - Release Date: 18/07/2001

sorry, i forgot to mention that i'm running linux so these routines are
not available i think
i found a way to make it work by realtime-adjusting ( (un-)correcting the
wait-time) ...
greetings
leander

(i'm still interested if someone knows a better way or can tell me something
about the window decoration)

Quote:> For greater accuracy I would recommend using timeGetTime() or the high
> performance counters (QueryPerformanceCounter(), QueryPerformanceFreq()).
> Also - I haven't checked whether this is relevant - but unsure your maths is
> entirely float or integer as necessary or you may experience rounding
> errors.