memtest86 random errors, but Err_Bits consistent: is my MOBO/CPU dead?

[PvtB]

Two weeks ago my system had suddenly began experiencing a series of BSOD's in Windows and eventually just refused to boot. After some testing, I became suspicious of the memory so I have decided to purchase an identical stick of CM64SD256-2400 C2 to test the original one.

I have used memtest86 to test both modules. First I had clear CMOS and set FSB to 100MHz to make the system boot with the old module (it stopped posting at 133). The old module seemed to run at first although it was producing thousands of errors in memtest. Unfortunately after a few reboots it just died again. The system wouldn't post at all, even at 100 FSB, sometimes producing "memory error" beeps on power-on. It didn't matter what DIMM slot I put it in.

At that point, I've decided to buy that second module. Unfortunately the freshly purchased module also produces 1000's of errors in memtest86. They appear to be at random and happen during different tests. Here are my observations:


1. All errors have EXACTLY same Err_Bits pattern: 0000xx00 (xx depends on the current bit test pattern).
2. Errors seem to show up at random, usually consecutive locations, at every test.
3. The longer the system runs, the more errors are reported by memtest (at the beginning no errors at all, after about 1/2 hour, errors are reported at almost every address and go into 100's of 1000's).
4. Changing DIMM slots, disabling cache, has no effect.

The observations seem to suggest that either CPU or Motherboard is damaged (am I right?). The gradual increase in errors may be heat related, however, the system appears to be running within range (CPU is only at 40C and all fans are working).

Here are my questions:

1. Is there any way to tell whether the CPU, motherboard, or the new stick is damaged? (I would assume it's probably not the brand-new stick).
2. Is my original module dead? Did it damage the motherboard in the process, since the new module isn't operating properly?

Here are my test system specs (I have never overclocked any component):

Motherboard: AD11 1.2 M00604 (FSB jumper set to 100MHz)
BIOS: ABA42 (cleared to default)
Athlon TB 1333MHz (actually runs at just 1000MHz, due to the 100 FSB setting)
256MB CORSAIR XMS 2400 (CM64SD256-2400 C2) (either the one that I've just purchased, or the original one)
Nvidia Riva TNT AGP (I had tried several other video boards to no
avail, so the video card is not the problem)
Power supply: 350W

System run perfectly for almost a year now, until I've started getting random blue screen errors. Soon after it wouldn't boot at all at 133 FSB, later at 100 FSB. I did my tests at 100FSB and default, flashed CMOS settings.


Thanks for any help,

 

[LiLithTecH]

Memtest86 usually reports pattern errors and block move errors in large quantities.
These problems could be caused by the degraded signals that result from the a low
impedance mismatch on the motherboard.

More than likely the problem is not in the memory.
The problem may to be caused by the motherboard's interaction with the CPU and/or memory.
Other memory testers may not report memory errors, because the memory itself is OK.
Memtest86 just happens to do pattern tests that end up in corrupted data. Since it is a
memory tester, it reports corrupted data as "memory errors" even though the errors are not
caused by bad memory.

Memtest86 actually tests the CPU, Cache and Memory all at the same time, and it can?t tell
exactly where the problem is. If the memory really is bad, Memtest will find errors at
specific locations in the memory chip, and should find them on any motherboard. When Memtest
finds the same ?errors? on every memory chip, but only on one motherboard and at one CPU speed,
it points to the motherboard and CPU as the problem

Try running the tests with the CACHE turned off. Press "C" to turn it OFF.
That would help pinpoint a bad motherboard.

 

[PvtB]

Disabling cache, or changing to a different DIMM slot has no effect. I think both sticks were behaving in a similar way in memtest, unfortunately the original one now refuses to post, so I can't verify it.

For the one working stick, all errors seem to happen at random and always have exactly the same Err_Bits pattern, I am also pretty convinced that there is something wrong with the CPU/MOBO. Is there any way to test which one is it? Unfortunately I don't have a second system to swap components and test them one at a time.

 

[PvtB]

I have repeated the same test with the new memory stick at 133FSB - same results, same error patterns. The original stick won't even post at either 100 or 133.

 

[LiLithTecH]

I would have to suspect the Motherboard is bad.

Bad system cache or failing support chip.

Rarely do CPU's die without OC'n them.