Continuous Problems With the 9800 Pro

[Michael]

I've always had problems with my setup and my sapphire radeon 9800 pro
128 when it came to 3D, etc. I could never get any catalyst drivers
beyond 3.9 to work for me, and if I even attempted to play games for
more than half an hour to an hour, the game would freeze- either with
corrupted graphics or the graphics and audio in an endless loop.
Frequent restarts during anything that stressed the video card
occurred often as well.

Whenever I would try to install a new catalyst driver, whether it was
the sapphire drivers, ati drivers, or omega drivers (was worth a
shot), I would get a black screen at the point where one would choose
the account to log in with. I would then have to boot into safe mode,
remove the drivers, reboot, and rollback the version. Very annoying,
but no way to know what was causing the problem.

After a little research, I decided to try a variety of things. I tried
setting the videocard to 4x AGP and shut off fast writes from the ATI
control panel, but I got the black screen again. Safe mode, uninstall,
reboot, reinstall, take another shot at it.

I ran memtest86+ for a few hours, and an unusual thing started to
happen. At the end of test 5, somewheres around 96-99%, lots of errors
accumulated, usually starting at the 525 MB region and continuing up.
I switched the dimms to opposite slots, and there was no change! A
little research showed that this was not uncommon with newer chipsets,
and it was suggested by an op in IRC to try checking my CAS settings.
I changed them to the settings reccomended by Geil, but no change in
that either. I decided it wasn't worth pursuing as it was the weekend
and I'd better call my manufacturer (ABS) for more help on monday, but
it still left me suspicious of the dimms or L2 cache.

Here's where I started to get frustrated. I had some leads, but
nowhere to go with them. This is where I decided to try tapping the
help of some newsgroups. This is the list of possible causes that I
have heard of on the net:

-PSU is not powerful enough... probsbly not, I have a 530 watt psu
with the right amps on the 12v line...

-radeons need to be on their own dedicated line on the psu... already
done

-some extra voltage needs to be delievered to the card...possibly, but
i tried increasing 0.1V, no change, still this may be a solution.

-bad card...unlikely, but possible, as I have used this for months
with 2d, and it is rock solid. Still a possiblity, though.

-the GA-8IPE1000 Pro2 motherboard that I have might be confusing the
audigy 2 with the video card as far as IRQ's are concerned...this is a
possiblity, could someone shed some light on this?

-heat issues, not enoug fans, etc.. i don't think this is the prob,
but it might be- i have fan in the front, 120 mm fan in back, and the
card itself has a heatsink and fan.

All in all, I'm frustrated and at my wit's end. This system is
marvelous, but for the incredible graphic instability. Posted below
are my specs... if someone would possibly help me out on this one, I
would be very grateful

-Mike Glukhovsky

SPECS:

Pentium 4 3.00 GHz, 800 MHz FSB
GA-8IPE1000 Pro2 motherboard
1 GB Geil PC3200 400Mhz DDR RAM (2 DIMMS)
200GB Western Digital SATA HDD, 7200 RPM, 8MB cache
Sapphire Radeon 9800 Pro 128 MB
Creative Audigy 2 ZS
Netgear 108 mbps WLAN PCI card (WAG311)
Sony DUW-401A 4x DVD-RW
Plextor 52x DVD/CD-RW

 

[HornDog]

[...]

I ran memtest86+ for a few hours, and an unusual thing started to
happen. At the end of test 5, somewheres around 96-99%, lots of errors
accumulated, usually starting at the 525 MB region and continuing up.
I switched the dimms to opposite slots, and there was no change! A
little research showed that this was not uncommon with newer chipsets,
and it was suggested by an op in IRC to try checking my CAS settings.
I changed them to the settings reccomended by Geil, but no change in
that either. I decided it wasn't worth pursuing as it was the weekend
and I'd better call my manufacturer (ABS) for more help on monday, but
it still left me suspicious of the dimms or L2 cache.

Any error in Memtest is unacceptable. Most MBs with the Intel 865-875 chipset
work better when the ram voltage is bumped up by a tenth of a volt. This has
been reported
by many users. I would think that your Video card is not the cause of your
problem.

[...]

 

[McGrandpa]

Mike,
I have a GA8-IPE1000-G, not pro, with 1 gig Xerox PC3200 in dual
channel, P4 3.0E/800/1m L2, SATA disabled, have 2 WD hd's on IDE1, a
160g and a 60g. ATI 9800 Pro 128. I'm having a few problems with
texture tearing and the like, but nothing like you're having.
Try this, when you get into your BIOS, at the main screen, hit CTL+F1
and check out the extra settings. Mine was originally 'boot PCI' not
AGP, and aperture was 64 megs, not 128 and both did make some
difference. You're going to have a fun time assigning an IRQ to just
the AGP slot, a bunch of those USB ports share that.
This 9800 is new, I installed the Cat 4.5s fresh, tried the 4.6, went
back to 4.5 to see if it made any diff with the texture tearing in OGL
especially (Q3A). It did do it some in D3D, with Halo and Far Cry. It
didn't seem to with UT2004.
HIH!
McG.

 

[johns]

Mem test software is crap, and has been for the last
4 - 5 years. None of it works. The stupid thing has
probably generated a buffer overrun because you had
it in loop mode. I strongly suspect that your problems
are because you left remnants of the old drivers in place
and installed new drivers over them. And the more
you mess with that situation, the worse it will get. If
you can do it, I think you should reinstall from scratch
and see if that fixes it. Generally, if there is a hardware
problem, it will louse during the install, and is easy to
spot.

johns

 

[Michael]

Any error in Memtest is unacceptable. Most MBs with the Intel 865-875 chipset
work better when the ram voltage is bumped up by a tenth of a volt. This has
been reported
by many users. I would think that your Video card is not the cause of your
problem.

Thanks a lot... I'll try that and see how it goes. I'll report back
with the results.

 

[Michael]

Any error in Memtest is unacceptable. Most MBs with the Intel 865-875 chipset
work better when the ram voltage is bumped up by a tenth of a volt. This has
been reported
by many users. I would think that your Video card is not the cause of your
problem.

Thanks a lot... I'll try that and see how it goes. I'll report back
with the results.

 

[Michael]

Mike,
I have a GA8-IPE1000-G, not pro, with 1 gig Xerox PC3200 in dual
channel, P4 3.0E/800/1m L2, SATA disabled, have 2 WD hd's on IDE1, a
160g and a 60g. ATI 9800 Pro 128. I'm having a few problems with
texture tearing and the like, but nothing like you're having.
Try this, when you get into your BIOS, at the main screen, hit CTL+F1
and check out the extra settings. Mine was originally 'boot PCI' not
AGP, and aperture was 64 megs, not 128 and both did make some
difference. You're going to have a fun time assigning an IRQ to just
the AGP slot, a bunch of those USB ports share that.
This 9800 is new, I installed the Cat 4.5s fresh, tried the 4.6, went
back to 4.5 to see if it made any diff with the texture tearing in OGL
especially (Q3A). It did do it some in D3D, with Halo and Far Cry. It
didn't seem to with UT2004.
HIH!
McG.

That's VERY interesting. First off, I noticed that it was "boot PCI"
and very quickly changed it to "boot AGP". This seemed to only make
the problems worse! VPU recover would continually kick in, Windows
would boot to a blac screen for 10 seconds and only then show up, and
disabling vpu recover made it stall in a black screen as it usually
does during a new catalyst install. And my aperture already was 128...
but at least that eliminates the two of those.

As far as assiging an IRQ to the agp video card in conjunction with
the usb ports (and remembering that I do have a PCI audigy 2) how
would you reccomen I go about doing that?

I'm happy that you've had success with yours, and I'm sure the texture
tearing is a catalyst driver problem that will soon be resolved. But
then again, who am I to be giving advice on how to make video cards
work?

At least I know that the 9800 Pro DOES work with the gigiabyte mobo...
and now I strongly suspect either the ram or the audigy. If anyone
could shed some light on this aspect of my problem, it would be
incredibly helpful.

 

[Michael]

Mike,
I have a GA8-IPE1000-G, not pro, with 1 gig Xerox PC3200 in dual
channel, P4 3.0E/800/1m L2, SATA disabled, have 2 WD hd's on IDE1, a
160g and a 60g. ATI 9800 Pro 128. I'm having a few problems with
texture tearing and the like, but nothing like you're having.
Try this, when you get into your BIOS, at the main screen, hit CTL+F1
and check out the extra settings. Mine was originally 'boot PCI' not
AGP, and aperture was 64 megs, not 128 and both did make some
difference. You're going to have a fun time assigning an IRQ to just
the AGP slot, a bunch of those USB ports share that.
This 9800 is new, I installed the Cat 4.5s fresh, tried the 4.6, went
back to 4.5 to see if it made any diff with the texture tearing in OGL
especially (Q3A). It did do it some in D3D, with Halo and Far Cry. It
didn't seem to with UT2004.
HIH!
McG.

That's VERY interesting. First off, I noticed that it was "boot PCI"
and very quickly changed it to "boot AGP". This seemed to only make
the problems worse! VPU recover would continually kick in, Windows
would boot to a blac screen for 10 seconds and only then show up, and
disabling vpu recover made it stall in a black screen as it usually
does during a new catalyst install. And my aperture already was 128...
but at least that eliminates the two of those.

As far as assiging an IRQ to the agp video card in conjunction with
the usb ports (and remembering that I do have a PCI audigy 2) how
would you reccomen I go about doing that?

I'm happy that you've had success with yours, and I'm sure the texture
tearing is a catalyst driver problem that will soon be resolved. But
then again, who am I to be giving advice on how to make video cards
work?

At least I know that the 9800 Pro DOES work with the gigiabyte mobo...
and now I strongly suspect either the ram or the audigy. If anyone
could shed some light on this aspect of my problem, it would be
incredibly helpful.

-Mike Glukhovsky

 

[Michael]

Mem test software is crap, and has been for the last
4 - 5 years. None of it works. The stupid thing has
probably generated a buffer overrun because you had
it in loop mode. I strongly suspect that your problems
are because you left remnants of the old drivers in place
and installed new drivers over them. And the more
you mess with that situation, the worse it will get. If
you can do it, I think you should reinstall from scratch
and see if that fixes it. Generally, if there is a hardware
problem, it will louse during the install, and is easy to
spot.

johns

I know that memory testing software is an issue of controversy for
many, but nonetheless errors in the results would serve to concern me
(as it would for most people) rergardless of the possilibility of
flawed tests or results.

In any case, I'm not about to make any decisions about the quality of
the ram this moment. You may very well be right about the "remnants of
the drivers" being left behind... so what I think I'll do is create a
second small parition on my drive, make it the active partiton and
hide my old windows partition, and try a fresh install to see whether
or not it actually makes a difference. Thanks so much for all your
help, and I'll post again as far as whether it works or not.

-Mike

 

[borolad]

I know that memory testing software is an issue of controversy for
many, but nonetheless errors in the results would serve to concern me
(as it would for most people) rergardless of the possilibility of
flawed tests or results.

In any case, I'm not about to make any decisions about the quality of
the ram this moment. You may very well be right about the "remnants of
the drivers" being left behind... so what I think I'll do is create a
second small parition on my drive, make it the active partiton and
hide my old windows partition, and try a fresh install to see whether
or not it actually makes a difference. Thanks so much for all your
help, and I'll post again as far as whether it works or not.

-Mike

What do you get if you :

-uninstall the Audigy
-pull the card from the slot
-reset the BIOS
-reboot ?

BoroLad
N.B. apart from no sound that is !

 

[McGrandpa]

That's VERY interesting. First off, I noticed that it was "boot PCI"
and very quickly changed it to "boot AGP". This seemed to only make
the problems worse! VPU recover would continually kick in, Windows
would boot to a blac screen for 10 seconds and only then show up, and
disabling vpu recover made it stall in a black screen as it usually
does during a new catalyst install. And my aperture already was 128...
but at least that eliminates the two of those.

As far as assiging an IRQ to the agp video card in conjunction with
the usb ports (and remembering that I do have a PCI audigy 2) how
would you reccomen I go about doing that?

I'm happy that you've had success with yours, and I'm sure the texture
tearing is a catalyst driver problem that will soon be resolved. But
then again, who am I to be giving advice on how to make video cards
work?

At least I know that the 9800 Pro DOES work with the gigiabyte mobo...
and now I strongly suspect either the ram or the audigy. If anyone
could shed some light on this aspect of my problem, it would be
incredibly helpful.

Mike, I came by this IPE1000G in an interesting way. The mobo I had
used for a year died, it was an GA8iHXP 2.1, and it definitely lost the
voltage regulator to the AGP bus. I called my local shop for help, the
owner is right friendly, he said bring the box over, we'd work something
out. OK, I take it, he has a GA8-848P-L, I drop my 2.66 northwood in,
he drops the 2 x 512m DDR400 in (this is not a dual channel mobo), I put
the 9800 Pro in and all is well. Until I get a 3.0E Prescott which
upped the fsb from 533 to 800 and the DDR400 automatically kicks up from
166 actual to 200. At first, it was lockups in Far Cry, then other
things, then the rig would suddenly reset for no apparent reason.
Turned out that it was the 848P chipset that didn't work so well with
the DDR400 at full speed. He gets me another mobo with the 865P
chipset. This one. It not only supported PC3200+ at full speed, it
does it in Dual Channel. 3rd generation dual channel. There are no
longer any problems with the 9800 Pro locking up, VPU resetting the
card, and no arbitrary hard resets now. I was also told that some of
these newer chipsets are implemented a little differently from brand to
brand, and sometimes a mobo won't like a particular flavor of ram. I
don't know why either, ram is simple enough.
Like me, you're boiling it down to the smallest denominator. In your
case, you may well find the answer in simply moving the Audigy to a
lower slot. Or simply removing it for testing.
It will be hard to remove ram for testing, but as the book says, there
are different things to try.
You have 2 DIMMS, right? 4 slots. You're likely using 1 and 3, well
try 2 and 4. If it won't boot or still gives problems, try both in 1
and 2, which stops Dual Channel, but it's for testing right?
HIH, and keep us posted!
McG.
(if none of this works, then try manually adding latency to the ram
settings in the extra BIOS settings, but you shouldn't have to do that)