Path: utzoo!utgpu!news-server.csri.toronto.edu!mailrus!ncar!boulder!wallwey
From: wallwey@boulder.Colorado.EDU (WALLWEY DEAN WILLIAM)
Newsgroups: comp.windows.ms
Subject: Re: MS-WINDOWS IN 1024x768 RESOLUTION
Message-ID: <18652@boulder.Colorado.EDU>
Date: 21 Mar 90 00:09:21 GMT
References: <327@spot.wbst128.xerox.com>
Sender: news@boulder.Colorado.EDU
Reply-To: wallwey@boulder.Colorado.EDU (WALLWEY DEAN WILLIAM)
Organization: University of Colorado, Boulder
Lines: 67

In article <327@spot.wbst128.xerox.com> tomer@rainbow.UUCP (Shmuel Tomer) writes:
>
>I am using an ATI-VGA Wonder card (in a Gateway 386/20 machine, 4MB
>RAM, 65 MB hdd). I have a serious performance problem when using
>Windows (either 286 or 386) in the 1024x768 mode, using the driver
>provided by ATI, revision 1.03. 
>
>Two operations in particular take unreasonable time. Moving a window
>around on the screen results in a VERY slow repainting of the
>window - this take between 10 to 30 (!) seconds. Scrolling a text
>editor (notepad, write etc.) by a single line, either using arrow keys
>or by clicking on the arrows in the scroll bar, is also VERY slow.
>In contrast, stretching a window is done reasonably fast (though not
>as fast as in the standard VGA mode) to the point that the best way to
>move a window is to stretch two sides or two corners. Also, paging
>through a file (page up/down buttons or clicking on the scrollbar) is
>reasonably fast.
>
>I called ATI and was told that it wasn't me - that's the way it works,
>and they can do nothing about it until Microsoft comes up with Windows
>that has a larger buffer (?). Yet, I can't imagine anybody actually using
>Windows with this kind of performance.
>
>Does anybody experience the same problems, or, better yet, does
>anybody know of a way to improve the situation? Any information will be
>appreciated.
>
>Shmuel
>
>Tomer.wbst128@Xerox.COM

Sounds like somebody at ATI needs to get to work and actually write a 
decent driver!!!!  Since the screen is fast in stretching windows, I
VERY SERIOUSLY doubt that it is the hardware!!  To stretch a window
actually takes more programming and is harder to do than to move a window.
To move a window requires simply copying the video bit map to memory,
actually moving the window logically, and copying the memory back to the
window.  What might (pure speculation-----) be the problem with the ATI
driver is that they use many of Microsoft's prewritten (in the sense
that if the driver doesn't include them, Windows uses its own) GDI
routines.  Many of MicroSofts BitMap GDI routines are supposidly very
slow--Especially if the bitmaps are bigger than 64K even if it supports
them.  Of course the
problem might be just the opposite---that ATI BitMap routines are real
slow and MicroSoft's real fast!  CAN SOMEONE PLEASE COMMENT ABOUT THE
VALIDITY OF THE ABOVE STATEMENTS!!!!!!!!!!!!!!

In either case I think that ATI needs to get to work, esp since 
many posters have said that the Video Seven Board is just fine---
I don't think VRAM should make that much difference--ANY COMMENTS???
How well do other boards work, like the Logix, Paradise, and others????

Finally, could somebody please post some more time trials like---

"To Move a window that fills 2/3rds of the screen about 30 pixels
left on my Board Brand X took 3 Seconds"

-or-

"To stretch my window from using the upper right 1/3rd of the screen
to almost the full screen (note -not the full screen button) on my
Board Brand X took 1.5 Seconds"

This also might help pin down the problem.

	Thanks 
	   Dean Wallwey