Video Card Rebooting! Can't control, Eject, Eject!

Nephtys · Post by **Nephtys** » 2006-05-05 11:00pm

*cues crashing and parachute opening noises*

Whenever I run a game that has a lot of various color blending effects (Eve Online, Freelancer when in hyperspace, Weird Worlds: Return to Infinate Space or Plasma Pong), my computer screen suddenly goes black, and makes a loud, choppy HURRRRRRRR noise for about 3 seconds before rebooting.

I've got a 128mb Radeon X600XT PCI-E. I'm running the newest ATI Drivers. It still happens and bothers me greatly, making some games unplayable due to crashes.

This is most notable on Plasma Pong, where about 10 seconds worth of gameplay guarentees a crash. The 'VPU recover' settings are on for my card, but it keeps happening. Any help here?

Arrow · Post by **Arrow** » 2006-05-05 11:55pm

Has it always done this, or only started recently? Did it start after the new driver install, or after you added new hardware? What are the rest of your system specs?

Nephtys · Post by **Nephtys** » 2006-05-06 01:54am

Arrow wrote:Has it always done this, or only started recently? Did it start after the new driver install, or after you added new hardware? What are the rest of your system specs?

It's always done this, but it's become less common recently. No new hardware, and the updated drivers didn't correct it. I've got a P3 3.2ghz, 512SDRAM machine with no other components of note. I have no idea motherboard make, or power source or whatnot. It's not heat that's causing it, I know that for a fact though.

Pu-239 · Post by **Pu-239** » 2006-05-06 03:16am

I've got a P3 3.2ghz, 512SDRAM machine with no other components of note.

Wow, a 3.2GHz Pentium 3?

Netko · Post by **Netko** » 2006-05-06 03:26am

It is the kind of anomaly that can suggest hardware malfunction. Had a ti4200 back in the day that always crashed after about 15min if the DX7 style T&L engine was running while being perfectly stable in DX8 shader games. Turns out it got damaged somehow so that the regular heat of running caused the damaged component to crash the card.

But I would try other things before writing the card off (or hopefuly having it replaced under warranty).

Uraniun235 · Post by **Uraniun235** » 2006-05-06 04:09am

Does your computer gracefully handle memory-intensive - but not graphically intensive - programs? How about CPU-heavy programs?

Nephtys · Post by **Nephtys** » 2006-05-06 06:44am

Pu-239 wrote:
I've got a P3 3.2ghz, 512SDRAM machine with no other components of note.
Wow, a 3.2GHz Pentium 3?

Typo.

The compy works fine with cPU intensive stuff, but struggles with card-heavy stuff. It dies to all hell in snow maps in Call of Duty 2 I've recently realized, for one.

Arrow · Post by **Arrow** » 2006-05-06 07:36am

Sounds like a power problem, then. Either your PSU isn't up to the job, or your card has a bad power regulator (or the card itself has some other defect). If you can, open your case and find the label on your power supply, and post the information. It should tell you the make and model, as well as the current on the 3.3v, 5.0v and 12.0v rails. An additional thing you can do is post the voltage levels read by your motherboard (you should be able to find them in the power menu of your BIOS, or through a third party utility).

Uraniun235 · Post by **Uraniun235** » 2006-05-06 01:56pm

The important thing is really the manufacturer and the maximum amperage on the 12v line - extremely unlikely that it'd be a shortage of 3.3v or 5v power.

Jaepheth · Post by **Jaepheth** » 2006-05-06 10:17pm

this was happening to me, it'd blue screen on ram intensive stuff and blame the video driver.

But it turns out it was my system RAM. I just lowered the voltage on my system RAM, and haven't gotten that crash again since.

Nephtys · Post by **Nephtys** » 2006-05-06 10:18pm

Jaepheth wrote:this was happening to me, it'd blue screen on ram intensive stuff and blame the video driver.

But it turns out it was my system RAM. I just lowered the voltage on my system RAM, and haven't gotten that crash again since.

I'm not BSODing. The screen blacks out then goes HURRRRRRRRRRRK awkwardly for a few seconds. Quite different it'd see. I really don't think it's the power supply.

Uraniun235 · Post by **Uraniun235** » 2006-05-06 10:20pm

Yeah, you might try running Memtest86+ overnight and seeing if it comes back with any errors.

Jaepheth · Post by **Jaepheth** » 2006-05-06 10:23pm

Nephtys wrote:
Jaepheth wrote:this was happening to me, it'd blue screen on ram intensive stuff and blame the video driver.

But it turns out it was my system RAM. I just lowered the voltage on my system RAM, and haven't gotten that crash again since.
I'm not BSODing. The screen blacks out then goes HURRRRRRRRRRRK awkwardly for a few seconds. Quite different it'd see. I really don't think it's the power supply.

Do you have auto-reset on?

(control panel->system->advanced->startup and recovery)

Still, you could try underclocking stuff to see if it helps.

Nephtys · Post by **Nephtys** » 2006-05-06 10:28pm

Jaepheth wrote:
Nephtys wrote:
Jaepheth wrote:this was happening to me, it'd blue screen on ram intensive stuff and blame the video driver.

But it turns out it was my system RAM. I just lowered the voltage on my system RAM, and haven't gotten that crash again since.
I'm not BSODing. The screen blacks out then goes HURRRRRRRRRRRK awkwardly for a few seconds. Quite different it'd see. I really don't think it's the power supply.
Do you have auto-reset on?

(control panel->system->advanced->startup and recovery)

Still, you could try underclocking stuff to see if it helps.

Auto-reset is on right now. Should it be? I'll try playing plasma pong as the test while underclocked to see how it goes.

Edit:
Okay. I turned off autoreset for the hell of it, and well. When I played Plasma Pong, it crashed as expected after about 30 seconds (level 2!). Except it didn't restart. It maintained the HRRRRRRRK sound, as I hoped the VPU would come and save it. Nope. Instead, after the screen went black, it became a solid and flat dark blue afterwards, and I shut the machine manually then. Meep Veep.

Edit, The Second: Underclocking it to 2/3rds speed doesn't work. Also, the chances of heat being an issue are zero. Compy 386 (my little machine) is well ventilated, and has a wind tunnel inside, so it shouldn't be that.

Uraniun235 · Post by **Uraniun235** » 2006-05-07 01:30am

Right-click on My Computer, click on Manage, and you should get this window:

Left click on System, within the Event Viewer tree. Sort by Source and scroll down to where the value under the Source column is "System Error".

Double-click on the most recent such entry (if one exists) and copy/paste the error code for us. (example below)

Arrow · Post by **Arrow** » 2006-05-07 08:17am

Couple more questions:

The HURRR sound - is it coming from your speakers or from the video card itself?

Always, what procedure did you follow to install your drivers. Did you the old drivers from the Add/Remove programs control panel before installing the new ones? Did you use a program such as Driver Cleaner to make sure the all the old driver files where gone? Or did you just install over the old drivers?

Nephtys · Post by **Nephtys** » 2006-05-07 05:46pm

Error code 00000019, parameter1 00000020, parameter2 ff642e98, parameter3 ff642fd8, parameter4 0a280005.

The HURRRRRRRRRRR is from inside my box, so I'd presume it's the vidcard or something. I cleaned up older drivers by letting ATI's thingie auto-update.

Uraniun235 · Post by **Uraniun235** » 2006-05-07 08:30pm

Do you have any friends with spare video cards that they could loan to you long enough to swap with your own temporarily?

4Tran · Post by **4Tran** » 2006-05-07 09:42pm

This sounds like the video card is overheating. The sound is likely the video card fan tryig to rev up, but is having problems doing so. Try taking the video card out, and see if its fan can still spin properly. Uranium235's advice of swapping out the video card is also a good idea.

Nephtys · Post by **Nephtys** » 2006-05-07 09:50pm

4Tran wrote:This sounds like the video card is overheating. The sound is likely the video card fan tryig to rev up, but is having problems doing so. Try taking the video card out, and see if its fan can still spin properly. Uranium235's advice of swapping out the video card is also a good idea.

No spare card here, I'm afraid. Nor friends with available spares.

But it's very odd. I can play some games (Half Life 2, Dawn of War, etc) on high settings and it works fine. Then when a certain type of particle or color blending effect happens, the compy dies hideously.

I really doubt my graphics card is overheating 10 seconds after my computer boots up when I try to run Plasma Pong.

Ace Pace · Post by **Ace Pace** » 2006-05-08 03:25am

I'm not sure, my 6800GT overheats very fucking fast playing Galciv2.

Nephtys · Post by **Nephtys** » 2006-05-08 04:18am

Ace Pace wrote:I'm not sure, my 6800GT overheats very fucking fast playing Galciv2.

Weird as hell. I have ZERO problems with 3 hour GalCiv2 sessions, and the X600 is a lot worse than the 6800GT. I can't be overheating within 10 seconds of measly plasma pong, can I? I can't imagine heating my board in an oven for 10 seconds and that happening.

Ace Pace · Post by **Ace Pace** » 2006-05-08 04:23am

Nephtys wrote:
Ace Pace wrote:I'm not sure, my 6800GT overheats very fucking fast playing Galciv2.
Weird as hell. I have ZERO problems with 3 hour GalCiv2 sessions, and the X600 is a lot worse than the 6800GT. I can't be overheating within 10 seconds of measly plasma pong, can I? I can't imagine heating my board in an oven for 10 seconds and that happening.

Well, I'm not exactly sure why my 6800GT is overheating, I mean, it dosn't happen in any other games(I get hard D3D crashs that force restart and complete lockups), and I can't trace it to anything except GPU tempature spiking VERY rapidly.

yes, I'm patched and GPU throttling enabled.

Jaepheth · Post by **Jaepheth** » 2006-05-08 05:03am

I'd suggest opening your case and observing it while it runs. pinpoint where the sound is coming from.

Arrow · Post by **Arrow** » 2006-05-08 08:08am

IIRC, ATITool has a temperature logger. You could download it, and you should be able to set it to write to the log file once a second (since Plasma Pong is crashing in 10), and that should determine if heat is your problem.

However, I'm also doubting if temperature is the problem. When I've seen a GPU core lock due to heat, the last couple of seconds of sound usually, loops, and when its the VRAM, the graphics are corrupted. But since its puking on shader effects, I'm starting to think something in hardware (most likely the GPU) is defective.