Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site meccts.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!mhuxr!mhuxn!ihnp4!stolaf!umn-cs!meccts!jlt
From: jlt@meccts.UUCP (James L. Thompson)
Newsgroups: net.micro.pc
Subject: Trying to access vertical retrace.
Message-ID: <273@meccts.UUCP>
Date: Mon, 18-Nov-85 12:12:11 EST
Article-I.D.: meccts.273
Posted: Mon Nov 18 12:12:11 1985
Date-Received: Thu, 21-Nov-85 05:22:54 EST
Reply-To: jlt@meccts.UUCP (James L. Thompson)
Organization: MECC Technical Services, St.Paul, MN
Lines: 17


  Does anyone know how to determine if the Color Graphics Adapter is
performing a vertical retrace?  I have written a routine to display
text by directly addressing the screen memory.  When I use this
routine on the monochrome card I don't get any snow.  When I use
this on the color card, however, it looks like hell.
  The Tech. Manual states '...the CPU should access the display buffer
during the vertical retrace time(if not, then the display will be
affected with random patterns...)'.  It also refers to vertical retrace
as Interrupt Level 2 (I assume that means the IRQ2 interrupt line).
After looking at the schematics I could find no connection to any IRQ
line on the bus.  This leads me to believe that the only retrace that
can be access by software is the horizontal retrace (IBM calls it Display
Enable or Horizontal Drive) via 6845 port X'3DA', bit 0.  The
horizontal retrace does not provide sufficient time to display a whole
line and therefore must be called for each character displayed.  This
technique is just too slow.