Index: [thread] [date] [subject] [author]
  From: Martin Lexa <mlexa@atlas.cz>
  To  : ggi-develop@eskimo.com
  Date: Tue, 03 Aug 1999 21:22:43 +0200

Re: ANNOUNCEMENT: New ViRGE driver

Hi Niklas!

Niklas Höglund wrote:

> I tested on a fresh update from CVS. I've updated again now, though,
> and it seems to work well except for one thing. I've seen this
> behaviour earlier on an old Tseng card in DOS. This bug only occurs on
> modes "higher" (bandwidth-wise) than 640x480x8, and the effect is more
> severe the higher the mode. It seems that drawing operations somehow
> delay the card in "starting a new line". When the screen is not drawn
> on nothing happens, but when it is drawn on some lines get shifted to
> the right for one vertical refresh. The next vertical refresh other
> lines get shifted, and when the drawing is finnished the display
> stabilizes.
> 
> The effect is easily reproducible with
> GGI_DEFMODE=1024x768x16 ./demo
> 
> I've looked some more at it and it doesn't start at the left edge of
> the screen, but about an inch to the right of the left edge. Exactly
> how it looks seem to depend on what is drawn.
  Yes, I know what do you mean. There are2 solutions to solve this:
  1) for 24/32 bpp: in ramdac code go to line 177 and insert this:
    virge_streams[22].data = 0xc000;
    then go to line 181 and insert this:
    virge_streams[22].data = 0x10000;
   
    It will probably work only ViRGE (dx, gx). This is not fully solved,
    because I wait for Jos permission to do this.

  2) Reduce vertical refresh rate of your monitor (60 Hz will do the
work).


> 
> --
>                                                 Niklas

Bye, Martin
		[E-mail: mlexa@atlas.cz]

Index: [thread] [date] [subject] [author]