[LJC]

I'm not normally one to have to ask for any assistance with these kinds of things,
but this one really has me going. The scenario I'm having is that I built a new
Pentium D 820 based system with a GeForce 7800GT PCIe card and 1GB of ram recently
and am having some issues with stability. More precisely, when I put the system
under very heavy load (the most extreme being running into Ironforge during peak
times while playing World of Warcraft and sometimes having a Skype conference also
going. This often involves lots of drive thrashing as it quickly tries to load massive
amounts of textures from the HD to the graphics card) my screen will occasionally
black out for 0.5 to 1 second, then come back. Most of the time it will do that three or four times
over about 5 minutes, then eventually it will flicker black, then immediately proceed
to a bluescreen. (IRQL_NOT_LESS_OR_EQUAL_TO most of the time)

Now, to make it even more interesting, CPU and GPU load doesn't seem to have too much of an effect on it, as I have run Prime95 (two copies simultaneously to max both cores) for about 12 solid hours with no errors at all. Also, using WoW as an example situation, I can play it and have Skype going for hours on end as long as I don't do something that involves massive loading sequences (such as it trying to load textures for hundreds of characters all at once when approaching the bank in a major city)

My system specifications are as follows:

Pentium D 820
Intel Desktop Board D945PSN
2x 512 MB DDR2-533
1x 200GB Seagate HD
GeForce 7800GT 256MB Graphics
Antec 500w Neo HE PSU
Sound Blaster Audigy

Now, here is what I have done so far...

I've tested the memory thoroughly, replaced the two 512MB sticks, and tested them
again in both single and dual channel modes with various timings just in case. (memtest86+)

Run a couple HD diags like Hitachi's DFT and Ontrack's little tool to check that.

Tested the power supply for issues with a couple different ATX PSU testers and a multimeter.

Checked for overheating issues on the CPU, northbridge, and graphics card. (all are falling well within acceptable temperature ranges)

Tried turning off many of the miscellaneous peripherals that are built onto the motherboard
(sound, parallel port, etc.) and have removed ALL non-necessary pci cards from the system while testing.


My next step will be to exchange the graphics card, as often the blue screen will
mention either nv4_disp.dll or win32k.sys

I'm going to zip and attach a set of about 5 minidumps, as they're itty bitty and may share some insight I'm missing when I looked them over with the Windows debugging tools.

Any ideas or suggestions would be greatly appreciated, as I'm past the point of frustrated

Thanks!

Attached Files

•  Minidump.zip   100.44KB   71 downloads

[Advertisements]

Hi and welcome to Geeks to Go. Random events like you described are ususally caused by heat, memory, driver issues or the power supply. I see you have ruled out heat and memory already. I would concentrate on driver issues and the power supply. Make sure you have the current driver for your video card. If you update the driver, use Driver Cleaner to clean the previous installation first. (it is free, google for it or search this forum as I have posted the link many times). As for the power supply, checking with a multimeter is good, but probably won't lead to a definitave conclustion. The real question here is, does the power supply have adequate wattage and current capability for your system, especially during graphically intense gaming activity. This type of activity puts a draw on the power supply and also generates heat. Look at the lable on the side of the power supply. It will give you the manufacturer, model number, wattage, and amp ratings on various rails (+12v, +5v, etc). Make note of these. Then go to http://www.hardforum...isplay.php?f=93 and check out the sticky posts at the top of the forum. This is the best collection of information on power supplies that I have found. You will be able to check your power supply against the information in the sticky posts and determine how it measures up. There are power supply calculators where you can detail your systems components and come up with watts needed to power your system as configured. Is it adequate? See that sticky post where a multitude of common power supplies are identified as less than desireable because the over state their ratings by they way they apply the term "under ideal circumstances". In these cases you need to deduct 30 percent from the stated rating to come up with a realistic rating. This was how I figured out my 420 Watt Raidmax power supply that came with my case was only capable of 294 watts (I needed 370 based on the calculator) and not adequate for my system. I had random reboot problems too. I replaced my power supply with a good quality Forton 520 watt PSU and reboot problem went away.

[austin_o]

Hi and welcome to Geeks to Go. Random events like you described are ususally caused by heat, memory, driver issues or the power supply. I see you have ruled out heat and memory already. I would concentrate on driver issues and the power supply. Make sure you have the current driver for your video card. If you update the driver, use Driver Cleaner to clean the previous installation first. (it is free, google for it or search this forum as I have posted the link many times). As for the power supply, checking with a multimeter is good, but probably won't lead to a definitave conclustion. The real question here is, does the power supply have adequate wattage and current capability for your system, especially during graphically intense gaming activity. This type of activity puts a draw on the power supply and also generates heat. Look at the lable on the side of the power supply. It will give you the manufacturer, model number, wattage, and amp ratings on various rails (+12v, +5v, etc). Make note of these. Then go to http://www.hardforum...isplay.php?f=93 and check out the sticky posts at the top of the forum. This is the best collection of information on power supplies that I have found. You will be able to check your power supply against the information in the sticky posts and determine how it measures up. There are power supply calculators where you can detail your systems components and come up with watts needed to power your system as configured. Is it adequate? See that sticky post where a multitude of common power supplies are identified as less than desireable because the over state their ratings by they way they apply the term "under ideal circumstances". In these cases you need to deduct 30 percent from the stated rating to come up with a realistic rating. This was how I figured out my 420 Watt Raidmax power supply that came with my case was only capable of 294 watts (I needed 370 based on the calculator) and not adequate for my system. I had random reboot problems too. I replaced my power supply with a good quality Forton 520 watt PSU and reboot problem went away.

[LJC]

Thanks for the reply.

After reading a few...less than splendid threads about these NeoHE power supplies, I'm going to exchange mine and go with a different one, just in case. I'm also going to try a few different driver versions, though I find it odd that mine would have issues with the same drivers my roommate is running on his virtually identical system. (he does have a different power supply and soundcard, but everything else is exactly the same).

Anyway, I'll give these a shot, and will post an update.

[LJC]

Revisiting this, I think you were right on by suspecting the power supply.

I actually wound up not replacing this one though. As I was peeking inside the case to prepare to take out the PSU, I noticed that I had hooked my two hard drives up on the same +12v rail as my 7800GT! It then hit me that it being that way fits EXACTLY what my problem was, as I would get the blue screens almost always when my graphics card was being pushed hard and the drives were also being accessed.

I moved my hard drives over to the second +12v rail, and have yet to have any lockups since. Obviously, it hasn't been quite long enough for me to declare victory yet, but I'm pretty sure this has solved the issue.

Thanks again for the ideas!