BSOD, Super-Pi errors, and more...

[DavidREmett]

System:
MSI K8N Neo2 Platinum (nForce3 250)
Athlon64 3000+ (Winchester), at stock (1.8GHz clock)
2x Crucial Ballistix PC3200, dual channel (128 bits), at stock (200MHz DDR401?? according to memtest86, 2-2-2-8) ("BL6464Z402.16TG" 512MB 184-pin DIMM 64Mx64 DDR PC3200 CL2 UNBUFFERED)
Akasa PowerPax 480W PSU
2 Hitachi 7K250 SATA drives (one 160GB, the other 250GB) using the SATA-3/SATA-4 mobo ports
A Liteon DVD drive, a Liteon CD-writer, and an NEC DVD-writer
Creative SoundBlaster Audigy 2 ZS
"Standard" ATI 9800 Pro 128MB

I have had a few BSODs over the past few weeks. I wrote the last two down:

"STOP: 0x00000024 (0x001902FE, 0xF79FF8F0, 0xF79FF5EC, 0x001F0003)"
According to MS, this indicates an error with ntfs.sys. I'm going to run a low-level disk check on both drives soon (I will post the results).

"PFN_LIST_CORRUPT -- STOP: 0x0000004E (0x00000099, 0x00000000, 0x00000000, 0x00000000)"
According to MS, this indicates broken memory.

SuperPi gives random errors (NOT CONVERGENT IN SQR, NOT EXACT IN ROUND) very quickly even on the early tests (64K/128K). These errors are random -- I don't get the same error every time and I don't get them at the same point.

Prime95 takes longer to give errors, but will give them before passing a single test. The blend test only uses 5-10% CPU, yet still causes my PC to run as tho it were using 99%, is this normal?!? The errors are again random.

Memtest86 reports the same errors every time it is run:

Fail address Good Bad
--------------------- -------- ---------
0000ea25f34 - 234.3MB fffffffb dffffffb
0002de368cc - 734.3MB 48901c54 48905c54
0000ea25f34 - 234.3MB ffffffff dfffffff

I switched the 2 modules around physically. The error addresses changed to (I didn't note down the good/bad stuff):

Fail address
---------------------
00012199734 - 289.5MB
000254486F8 - 596.5MB
00012199734 - 289.5MB

I'm assuming this means that the memory is faulty.

The voltages:

Vcore (should be 1.4V i think): 1.39V according to SpeedFan, 1.37/1.36 in BIOS, 1.38V in MSI core center.
+12V: 12.71V (??) according to SpeedFan, 12.09/12.16V in BIOS, 11.85V/11.91V in MSI core center.
3.3V: 3.18V according to SpeedFan, 3.34/3.32 in BIOS, 3.18/3.20 in MSI core center.
VCC (5V): 4.99V according to SpeedFan, 5.07V in BIOS, 5.03V in MSI core center.
(SpeedFan) Vin2: 0.00V (??)
5Vsb: 4.88V according to SpeedFan, 4.94/4.96 in BIOS, not reported in MSI core center.
(SpeedFan) Vbat: 0.00V (??)
(BIOS) Battery: 3.09--3.14V

The temperatures seem fine:
System (case): 23/24C
CPU: 41C (rises to about 49/50C under load, EG prime95 CPU test)
Chipset (My mobo has only one chipset, no NB/SB): 36C
HDs: 35C and 33/32C (One is right in front of the intake fans)

I haven't encountered any problems in general use of the PC other than the occassional BSODs.

The memtest86 results seem to indicate that the memory is faulty, however, memtest86 results show only 2 dodgy memory addresses, with only 1 dodgy bit in each, surely this wouldn't be enough to cause super-pi and prime95 to crash so quickly?

Any suggestions/help/whatever would be greatly appreciated. Thank you!

(Sorry about the length of this post, just wanted to post everything I knew...)

[ski]

Memtest should show no errors.
Clean off the RAM modules' contacts with hi-purity alcohol.
If no luck, then run Memtest with one module installed at a time.
If they both modules show errors, then suggest replacing both.
If one shows errors, then replace just that one.

[DavidREmett]

One of the modules was faulty, removing it solved all the problems (I've requested an RMA from crucial). Thanks! I am rather surprised that all those problems could be caused by only 2 faulty bits in memory!

[SuperSparks]

Yep, that's all it needs, it's one of the reasons why memory problems are so common. At leat you won't have any problems getting it replaced - Crucial are excellent in that respect.

[ski]

You're welcome.
Glad to hear that you were able to fix the problem, and thanks for posting back with the solution.

[DavidREmett]

I sent the faulty module back to crucial and got a new module on friday. Installed it and the problems are back again!
Modules in slots 1+2/3+4 (the only way to get dual channel with two modules on this board): Super-Pi/Prime95 errors and I've had a BSOD and a random reset (although no errors in 3 passes of memtest86 :S)...
Modules in slots 1+3 (memory forced to operate in single channel): No problems (as far as i can tell).
Either module in slot 1/3: No problems (as far as i can tell).

It seems to me as though there is some problem with dual channel, or some problem with having the two modules physically next to each other, EG overheating -- however, prime95 and super-pi fail within seconds, not long enough for them to get too hot?

[SuperSparks]

In that case the problem is almost certainly with the motherboard. Unfortunately there is no easy way to test it though

[DavidREmett]

I updated to the latest BIOS and cleared the CMOS, resetting everything to defaults. Rather dubiously, I'm no longer getting any errors (have been running Prime95 for about 7 hours now with no errors).