3d apps crashing

[Felix]

Hey folks, great to have the forums back up.

I've finally got my comp connected to the net, and most of its problems (like random restarting) have been solved. But I'm still having trouble with 3d progs, like 3dmark, quake3. 3dmark sometimes gets through (just under 10k score), but more likely either hangs or exits back to windows. Interestingly, Im pretty sure the only times it does this is during the 4 games, never during the other tests.

Quake3 hangs randomly. It will actually run fine for most of the time, but occasionally it decides to hang. Im sure if i tried out unreal tourament I'd get something similar.

Heating shouldnt be a problem, volcano 7+ with carefully applied as3. The case temp feels ok.

Drivers and updates Ive done are:

win2k service pack 3
nvidia 30.82 detonator drivers (the ones just released)
latest via4in1s
bios update

Ive been contemplating the scary idea of a vid card firmware update. Is this likely to do anything?

Or what else could I do?

I really want to try out some light overclocking, but definately not until I have a stable system.

Specs are:

leadman 400w psu
asus a7v333 mobo
athlon xp 1800+
256mb ddr333 kingmax ram
leadtek geforce4 ti4400
seagate barracuda 7200rpm 40gig hd
lucent win 56k modem

And heres a couple of messages in event log. This one doesnt seem to happen when it hangs, not sure what causes it:

Event Type: Warning
Event Source: EventSystem
Event Category: Event Service
Event ID: 4106
Date: 10/08/2002
Time: 3:59:50 p.m.
User: N/A
Computer: CARMEN
Description:
The COM+ Event System detected a corrupt IEventSubscription object. The COM+ Event System has removed object ID {A5978620-5B3F-F1D1-8ED2-00FA0035B753}. The subscriber will no longer be notified when the event occurs.

And occasionally when it hangs an event log like this is written:

Event Type: Information
Event Source: DrWatson
Event Category: None
Event ID: 4097
Date: 10/08/2002
Time: 5:14:01 p.m.
User: N/A
Computer: CARMEN
Description:
The application, , generated an application error The error occurred on 08/10/2002 @ 17:14:00.984 The exception generated was c0000005 at address 00740075 (<nosymbols>)

Note how it oh so helpfully leaves a blank space after application :p

[H2gO]

Quite simple my fellow

bad CARMEN

resolve = change compter name



Yeah yeah I know - sick j/k

back to your issue

I'd be inclined to nuke the whole thiing & rebuild it for all the trouble its giving you

- sorry

:(

[Felix]

rofl, its a nice name ok :)

Unfortunately, Ive already done that, and windows has been reinstalled several times.

[dustyslapper]

Maybe you could try running Prime95 for a couple of hours or so - just to double check that its a graphics problem as opposed to something else.

[Felix]

Thanks, I hadnt heard of that program. Im just dling it now, and will see how it goes soon.



Just tried it, and I did indeed get an error:

Test 1, 400 Lucas-Lehmer iterations of M19922945 using 1024K FFT length.
FATAL ERROR: Rounding was 0.5, expected less than 0.4
Hardware failure detected, consult stress.txt file.
Torture Test ran 0 minutes - 1 errors, 0 warnings.
Execution halted.

The txt file doesnt say anything specific about what the hardware failure might be.

[dustyslapper]

Excellent - at least its not your graphics card.

What are your temperatures like? Safe 'n sound below 50? Try whipping the side of the case off if its a little warm.

If its cool maybe try adding a little voltage, just to give the cpu every chance at stability.

Otherwise....?

[Felix]

Well that depends on what monitoring prog u ask :)

27, or 40, or 42, or 55. Im positive the 55 is way high. Ill be getting a thermal probe I think. Heh, asus probe sometimes tells me my hsf is going at -25000rpm. Ive previously upped the core voltage, which helped somewhat when i had random restarts, but not this.

Ill take the case off and up the voltage tonight just to make sure tho.

p.s. Just remembered to mention this. Occasionally it doesnt post, secondary master drive fail (thats the richoh cdrw drive). I originally thought it was due to an inadequate psu, but its still around after getting a 400w leadman.

[I.R]

the apps quitting sounds like not enuff Vcore...... are you OCing?? and what is you Vcore at now???
Don't know about the POST failure....... warm or cold boot???

[Felix]

No overclocking just now. I dropped back to 1.75v, will try going back up to 1.775 and 1.800.

I think its always cold boot.

Edit: Unfortunately, going up to 1.800v with side panel off hasnt made a difference.

[Wibber]

what board/vid ect?

[Felix]

Originally posted by Felix

Specs are:

leadman 400w psu
asus a7v333 mobo
athlon xp 1800+
256mb ddr333 kingmax ram
leadtek geforce4 ti4400
seagate barracuda 7200rpm 40gig hd
lucent win 56k modem


:p

[Binky Stunt Cat]

go search for NVMax..... and i do belive i've hidden it on my webpage...http://binky.orcon.net.nz in the files section.
that should fix the problem, just by enabling the AMD+VIA hotfix under System

*points out yet again to mods that a faq of troubleshooting needs to be made, with this mentioned (3rd or 4th time i've mentioned this fix to the same problem)*

[Felix]

Thanks. I dled it, but I dont see any AMD+VIA hotfix anywhere in system?

[Binky Stunt Cat]

Try reading this thread
http://forums.overclockers.co.nz/showthread.php?s=&threadid=1350

[Felix]

I ticked the win2k + athlon stability box, but it doesnt appear to make any difference. :(

Update: I just ran 3dmark and got a Kmode_exception_not_handled stop error during multi-texturing.

[Binky Stunt Cat]

well, a full reinstall wouldnt hurt.....

[Deviant]

99.999% of the time it's 1 of 3 things, and you've only tried 2 of them from what I've read.

1. CPU - not enough vcore or clockspeed too high - DONE
2. Drivers - Seems you've doneenough here too.
3. Memory - What about this?

I have seen problems like yours on my machine before, and on others. Try changing memory timings in your BIOS, even try defaulting all BIOS settings to the slowest, then work your way up. The other thing is try different memory, your stick (or both or whatever) may have a problem even at default speeds. Try some one elses in your machine, or if you have 2 sticks, remove one, and try it, then try the other stick on it's own.

[swiftynz]

i've had prime95 errors caused by setting the memory timings too high. what spec ram are you running and at what fsb?

try dropping it back to CL2.5 if it isn't already set.

[Antimatter]

Try testing with the CDR/W unplugged(power and IDE both unplugged).
I agree with the above comments that it sounds like a RAM problem though. The Kmode_exception_not_handled stop error during multi-texturing error would support that, as video textures are stored in main memory initially. Try changing the AGP aperture size in BIOS to see if that has any effect.

[Felix]

Im running at the default (optimal) memory settings. I tried increasing them. 3 of the 4 settings go ok when set to a lower timing, but one of them mucks it (didnt get thru a long post, and stop error booting into windows). So Ive kept it all back at optimal for now.

However, I ran a memory diagnostic program which found a number of errors. I might see about returning the ram. Unfortunately, I dont have any other sticks to try out.

FSB is default, 133. Ill try changing the agp and cdrw as well. Thanks all, will see how that goes.



mmm just tried the aperture size and cdrw, no luck, just exited out of 3dmark.