Constant BSOD W/ Same Bug

[Brando569]

i have an AMD 64 X2 3800 (which i think might be my problem for everything since the guy sent it to me wrapped in tissue paper ) and my computer wont run any other version of windows besides xp 64. i have xp 64 corp. installed and whenever i have DC++ open for extended amounts of time and i'm downloading alot, after a few hours ill come back and see the BSOD there to great me.

this is usually the error i get:

(the normal stuff here)

Bugcode_NDIS_Driver

(more of the normal stuff)

Stop: 0x0000007C (0x0000000000000005,0xFFFFFADFE6571228,0xFFFFFADFE6559500,0xFFFFFADFE6570700)

i have all of the lastest patches installed, is there any other way to fix this?? I also got the "IRQ Less Than Or Equal To Zero" error. this is gettin really annoying, im think about sellin all this 64 crap and goin back to 32 bit

Edited September 13, 2006 by Brando569

[nfm]

Before going into hardware, it could be a simple driver problem. You need to make tests on your own. Other thing may be unstable ram, cpu or on-die memory controller used in AMD processors. Make sure that your ram cooparates with processors on-die memory controller, you can test that by running two instances of super pi 32 million calculations on each core. If that passes move on to test #5 in memtest86, you should be stable if you can do 100 loops of tets #5. Other thing it may be unsatble cpu, you can alternatively up cpu's voltage but that's not recommended, cpu should run at deafult volts without any problems. Again, run two instances of prime95 or burn7k to test cpu stability, if it passes at least 8 hours then you're in good shape.

It also may be simple BIOS update adressing issues with x86_64 operating systems. I would check your power supply, since you switched to dual core cpu, measuring with multimeter would be my best bet. I highly doubt this is a hardware problem but you never know, I'm 75% sure this has to do with software.

Have you tried Azureus or BitComet?

[ripken204]

http://msdn.microsoft.com/library/default....9a86e98.xml.asp

looks like its ur enthernet port/driver. update its driver and if that doesnt work, check if its sharing an irq.

[Brando569]

thanks for the info guys, it looks like its DC++ thats causing the problem. just for the hell of it i didnt run it last night and when i woke up (11 hours later) it was fine. (I use DC++ because my University has a HUGE hub with about 700 users and 22TB of stuff )

As for the powersupply and RAM i bought them to match the processor. i upgraded everything at once.

Ripken, i was lookin at that lastnight but it didnt really give me any info i could use. i have an Asus A8N-SLI (NForce Chipset). I'll upgrade it and see if that fixes anything.

I'm led to belive that it is hardware since ive had problems in both windows and linux. I'll run the tests you mentioned nfm and ill post the results.

Edited September 15, 2006 by Brando569

[Brando569]

results:

Super Pi would only let me run one instance at at time for some reason. it would give me some error message that idk what it meant (Not Convergent in SQR05). it was able to calculate Pi to the 32 millionth place in 37 minutes. theres no special procedure to follow to run it on each core is there? just open it up twice and let it run, correct??

Ran Sisoft Sandra 2007 today using the burn in to test the cpu and memory and then i went to class for like 2 hours and came back and it had crashed. this is what the BSOD said:

Driver IRQL Not Less Or Equal

Stop:0x0000000D1

*** NDIS.sys - AddressFFFFFADFE3E6A0F6 base at FFFFFADFE3E25000, Date Stamp 42436112

i also updated my Nforce drivers 2 nights ago and that didnt seem to help any :-/

[cluberti]

It's still a driver issue, likely the ethernet driver. When you get a STOP 0xD1, what happened is that a driver attempted to access a page or a set of pages of kernel memory that was paged to disk (likely in kernel paged pool). It did so at an IRQ level of DPC Dispatch (or higher), which will immediately cause a bugcheck, as this is a no-no. Any time something running at a dispatch interrupt level or higher attempts to access kernel memory that is non-resident (paged out), this invokes the scheduler. This is strictly NOT ALLOWED in kernel, and the box will bugcheck.

This is NOT a Windows problem, but a driver problem - Windows is crashing to make sure that the kernel doesn't become corrupted due to the invalid access caused by the driver. Contact the vendor to get a driver that doesn't have this problem .

Edited September 15, 2006 by cluberti

[nfm]

results:

Super Pi would only let me run one instance at at time for some reason. it would give me some error message that idk what it meant (Not Convergent in SQR05). it was able to calculate Pi to the 32 millionth place in 37 minutes. theres no special procedure to follow to run it on each core is there? just open it up twice and let it run, correct??

Ok so we know it's hardware. To run superpi or or prime correctly you need to make two folders, and put an .exe into each, then fire them up, go into task manager and for one exe set core 0 and for the other core 1. Not Convergent in SQR05 in superpi means usually unstable processor, "not exact in round" in superpi is a problem with ram. RMA your processor if you can.

37min calculation is also very long. My 3800 X2+ at stock speed can do that in 30min. superpi is the most popular benchmark among the overclokers today, new overclocked core duo 2 can do 32 million pi calculation in 10-12min.

[RJARRRPCGP]

This may be a sign of the motherboard giving the wrong default Vcore. You may be required to set the Vcore manually.

[Brando569]

i just re-burned xp64 thinkin that maybe the cd was scrathed and it was screwing up stuff. now its even worse, when i do something dealing with graphics (scrolling a window, chaning the size of a window, etc..) its really choppy and slow, and randomly my sound doesnt work even though i have the drivers installed and the connections are fine.

this is really starting to aggrevate me, i thought building a 64 bit based computer would be awesome and its been nothing but problems for me.

As for RMAing the processor, would it be possible for me to do so since i bought it OEM off of ebay??

[nfm]

no, only retail can be rma'ed

[Brando569]

thats what i feared..... i was hoping that maybe there was a way to do it through AMD but i guess not. the choppy video problem was easy to fix, i didnt have the drivers installed. Both runs of super pi completed successfully taking 37 mins 50 secs on each core.