memtest86+, Same error, different DDR2 sticks

[highsnr]

I'm debugging unstable XP PRO SP3 on MSI P35 Neo2 (v1.9 bios) w/ Q6600 (G0 stepping) quad core that was somewhat overclocked, but is no longer. System would periodically go nuts, but never got any errors with lots of Memtest86 or Prime95 testing. System has been running since April '08.

Now BOTH (see types below) memories have ONCE generated the SAME type of error in Memtest86+ V2.10. Happens very rarely and only when starting up Memtest86+ and in first passes.

In MT tests #2 & #3, following a reboot and during first pass, generated ~32 errors at the same memory address range of 133.3M.

Since I don't believe the different sticks have the same errors, it occurs to me it's a MB problem or startup problem w/ MB, and Memtest86+.

Both mem types have passed many 10's of hours of memtest86+ passes WITHOUT error other than as described above.

First mems were 2X 1GB sticks, Crucial ballistix DDR2-800 AA804.8FE5 at 2.2V and 333 MHz. Stop using them because Crucial said no good with this MB.

Now using 2X 1GB non-ballistix replacements they sent (and overnight for free!) that are 1.8V and at 330MHz.

Operating latencies were verified to be within spec for speeds and voltages with Memtest86/CPUz.

I'm looking for stability rather than speed. Gave up on any overclocking.

I'm wondering if the new RAM is fine and the memtest86+ errors are not really ram errors but rather artifacts of memtest86 startup. E.g. voltage instability, chipset warmup, etc...

Power supply is 500W Enermax liberty.

Sorry about the long post.
Any ideas?
Thanks much.

[hunter33]

Definitely try upgrading the mobo BIOS first, that might be all it takes.

[Mike Clements]

See if the MSI BIOS has a feature called LEGACY USB SUPPORT, or someting similar. If yes, disable and try Memtest again. On most ASUS boards, the LEGACY USB SUPPORT setting will cause a false error indication when running Memtest so it's worth a shot. You might also try disabling BIOS SHADOWING just while running Memtest if you find that feature and the first step did not resolve it.

Also, what type of instablity are you having?

[highsnr]

OK, I'll try with Legacy USB support turned off.
Problem is this is the only error I get and 22 successive boots only produced the error the first time!
FWIW, I tried wiggling the USB mouse (keyboard is PS/2) to "exacerbate" the problem, without effect.
If this proves to be a non-real error, then this is, sadly, another example of prime95 + memtest86+ together failing, after much testing, to reveal ram that ultimately results in instability until it's swapped out.

Thanks very much for the replies.

[highsnr]

So several weeks later, my system is now rock solid.
Only thing changed was the memory from Ballistix to Crucial's "standard" memory.
So this is another example of a memory problem that wasn't diagnosed with extensive testing using Memtest86+, Prime95, superprime, etc.
I'm still looking for a diagnostic recipe that would find these sorts of problems.
Windows being unreliable is too unreliable. ;)
Thanks for the help.