CAT's: Needless Mode Switching. Weird bug

[Tony DiMarzio]

Ok. Let's assume your WinXP desktop is running at 1024x768x32bpp@100hz.
Let's also assume you've used a program like reforce to strip all unwanted
refresh rates. Now you go to run a small fullscreen direct3d or opengl app
that is configured to use the same exact settings as your desktop
(1024x768x32bpp@100hz). Would you expect that the monitor should not have to
change modes in this scenario and that if it did change modes it was a waste
of a mode switch? I would!

Call me neurotic but this bug is really bothering me. I've been trying to
isolate it for some time now... but I just can't put my finger on it. All of
my googling has also turned up nothing. I'm sure I'm not the only one
experiencing it... but maybe I'm the only one bothered by it.

With any of my previous video cards Voodoo3, Geforce2,3,ti etc... there
would be NO monitor mode switch in the aforementioned scenario. Likewise, my
ATI 9800 Pro 128mb will not do mode switches in that scenario (just as I'd
hope and expect) until the first time I change the Catalyst drivers from
what was installed immediately after installing WindowsXP. So basically.. it
goes like this:

1. Install WindowsXP
2. Install Cat 4.5. and run reforce.
3. Everything is honky-doree!! no unnecessary monitor mode switches.
4. Catalyst 4.9 Doom3 beta comes out.
5. I get that feeling in the pit of my stomach because I know I'm going to
install them and in the process will probably start seeing the unnecessary
mode switching bug.
6. I install the 4.9's (drivercleaner, safemode, etc etc <-- very clean cat
install) and DAMNIT the bug shows it's face.
7. What I didn't mention before is that now the bug is there forever!!!! I
mean it! The bug won't die now until I reinstall WinXP from scratch!
Reverting to Cat 4.5 will not fix this. I tried removing every trace of ATI
files or reg keys that I could to no avail.

So whatever file or registry setting is ****ing this up in the first place
is something that does NOT get removed by either ATI's uninstall utility or
drivercleaner.

Has anyone experienced this annoying bug before? I wouldn't care that much
but my monitor doesn't really "like" switching modes all the time.

-------------System Spec----------
Windows XP (2600-Professional) w/ SP1
AMD Socket 939 Athlon 64 FX-53, 1MB L2 Cache
ASUS "A8V Deluxe Wi-Fi" K8T800 Pro
VIA Hyperion 4in1 451v
1024mb (2 x (512mb OCZ Platinum EL (2-3-2-5) DDR400/PC3200) Dual Channel)
ATI Radeon 9800 Pro (Catalyst 4.5) (C-378mhz / M-337.5mhz) OC to
(C-411.75mhz / M-360.00mhz)
SbAudigy 5.1 (PCI-Slot 3)

120gb Western Digital WD1200JB primary/master
60gb IBM 60gxp primary/slave
16x HI-VAL DVDrom (BDV316C) secondary/master
48x/24x/48x plextor (PX-W4824A) secondary/slave

LianLi PC-6070 *Silent* Aluminum Mid-Tower Case | 2x80mm intake fans, 1x80mm
exhaust fan
PowerMax 470watt Aluminum PSU | +3.3V@26A +5V@47A +12V@28A

 

[Inglo]

Ok. Let's assume your WinXP desktop is running at 1024x768x32bpp@100hz.
Let's also assume you've used a program like reforce to strip all unwanted
refresh rates. Now you go to run a small fullscreen direct3d or opengl app
that is configured to use the same exact settings as your desktop
(1024x768x32bpp@100hz). Would you expect that the monitor should not have to
change modes in this scenario and that if it did change modes it was a waste
of a mode switch? I would!

As soon as I switched from nVidia to ATI a year ago I stopped using
Refresh Force, it caused too many problems. Using the Catalyst display
options I have the refresh rate override set to 85 Hz. I use 1280, 1152
and 1024 and they all refresh at 85, no mode switching or anything. My
monitor can do 100 Hz at the res's below 1280, but 85, 100 difference
isn't noticeable.
I think that reforce program is useless now, though it was indispensable
on my GeForce. My guess is it's at fault, could be wrong.

 

[Tony DiMarzio]

It's not reforce. I experience the same bug/issue using or not using reforce
to strip the unwanted refresh rates. Also, "refresh rate override" in the
ATI Control Panel does not cure this bug.

To clarify... this issue never prevented me from playing games or doing
anything at the requested refresh rate. The problem is that the monitor
blanks for about a second as if doing a mode switch whenever it enters or
exits a 3d application. Its apparently switching modes because not only does
the monitor blank but also the monitor light momentarily turns orange from
green then back to green - this is the behavior you'd see for example when
switch resolutions from 800x600 to 1024x768

thanks anyway though

 

[GTX_SlotCar]

Maybe it's an XP thing with ATI. My 9800Pro doesn't to that, but I'm running
W2k.

Gary

 

[Tony DiMarzio]

Possibly... I don't have a W2k Partition setup right now to verify. It is,
however, an obscure bug... those are the ones that always hit me it seems.
Every issue I've ever posted in this group I've never found a solution too
not puting this group down - just saying the bugs I find are obscure
and probably don't bother others as much as they bother me.

 

[Tony DiMarzio]

I should also note that by reverting to Cat 3.7 the problem goes away... I
can't and won't run cat 3.7 though for other reasons so staying at 3.7 is
not an option.

 

[Tony DiMarzio]

WOW!!! I actually fixed this bug!!! I pretty much just accepted it after
weeks of trying to isolate it and fix it. In the end the culprit, I think,
turned out to be radlinker. I say "I think" because I distinctly remember
the bug showing its face after upgrading from cat 4.4 to 4.7. The bug didn't
arise after installing radlinker. Nevertheless, radlinker has allowed me to
fix this.

Through a combination of selecting and un-selecting the option "mode change"
for a custom profile via radlinker, the behavior has returned to normal.

feels like a dream come true

This poses the question though... which registry setting does the radlinker
option "mode change" correlate to????? I can't find it in the registry and
obviously don't have the ability to look through the radlinker source code.