Date: Wed, 1 Apr 92 08:54:39 -0500
From: heflin AT pods01 DOT dnet DOT dupont DOT com (Roger Heflin)
To: "gatekeeper.esgate.dupont.com!sun.soe.clarkson.edu!djgpp"@esds01.dnet.dupont.com (DJGPP Mailing List)
Subject: Re: Video Modes
Status: O


Christoph Kukulies writes:

>The graphics package I have ported by means of djgpp is VOGL.
>It's a quite nice implementation of SGI's graphics 3D library.
>(SGI=Silicon Graphics). The missing pieces are the double buffering
>code which I could not implement because of my lack of knowledge of
>VGA internals. if you could contribute your code/knowledge it would be 
>great.
>BTW, what is 800x600x32768 mode? Which manufacturer supports this?
>Does your double buffering code work for ET4000 in 800x600x256?

I don't implement what is called double buffering, I buffer it into
cpu memory and copy it to the video card as fast as possible (for 800x600
 < 1/2 sec).  For what I expect you are doing this would not be workable, and
probably each video card would have to be specially programmed for this 
mode.  Also the way I am buffering the data I am not limited by the 1M of
video memory, I can buffer up to about 8M for a single image.  

The 800x600x32768 is a ET4000 mode.  If the ET4000 has a Sierra RAMDAC.  The
mode is 800 x 600 x 5 bits of {R,G,B} to give 32768 colors, with no palette.
The lack of a palette causes problems with some applications, such as anything
that whats the palette to change since this becomes very time consuming.

Most of the applications that I work with are image processing type 
applications so the double buffering it not as important as for 3D type
displays.  

If you want to know more about programming the ET4000, I have heard that
TSENG has a tech. manual that goes into detail about how to do things
like double buffering.  It can be done, my video card has some windows modes
that are 1600x1200x16 virtural screen and can be displayed on anywhere from
640x480 up to 1024x768 and as far as I can determine they are only using
the video memory, just moving around starting position of the screen.
I may have to look at how those virtural screen drivers are working, they
probably aren't too compilcated. 


	Roger Heflin