Over at XS several of the guys have reported having UTT based ram die on them, usually just one of the two. One of the guys posted that after letting his stick sit for a day...it started working again.
Have you tried that by any chance. I will try to find the post and if I do I'll provide you with a link.
Edit - here you go. The possible explanation may or may not correct, but the problem has happened to others and some have started working again so it might be worth giving it a shot. It would be a shame to send those back and get an exchange that doesn't perform as good.
1st post:
http://www.xtremesystems.org/forums/showthread.php?p=819546#post819546
2nd post:
http://www.xtremesystems.org/forums/showthread.php?p=820583#post820583
Possible explanations????
http://www.xtremesystems.org/forums/showthread.php?p=820885#post820885
Linked above but copied below:
****************************
04-17-2005, 10:47 PM #2726 / link
uwackme
Xtreme Member
Join Date: Jun 2003
Posts: 459 hmmm
--------------------------------------------------------------------------------
Yes, One of my sticks has been exhibiting this phenomenon. Runs awesome for weeks, then one day .... BEEEEP. Its one stick. Nothing would "clear" the error. But taking it out, leaving it aside for a few hours while I used a different stick and VIOLA... "bad" stick is back running for weeks flawlessly.
Like some latch-up problem (a problem that can effect CMOS circuits) gets the ram in a locked up state, a long sit with no power and the chips bleed to no electrons and presto its back up and running.
This would be pointing to motherboard signal integrity and power supply (the VDimm circuit) issues. The high voltage, and high current and high speed we are pushing these things too is taxing the rest of the system beyond what it is used to. This isnt a permanent damage issue, just a symptom of poor signal integrity, PCBoard and component quality. The RAM itself is fine, the latchup is occuring probably to UNDERSHOOT on some of the signal lines....ie: instead of a "0" being 0 volts, it was driven lower due to circuit bounce to some -X volts low enough to trigger the latchup, which then freeze's the circuit inside the DDR chip(s) until they are truly turned OFF and all charge is drained away.
This is all speculation, but the symptoms (bad ram miraculously coming back to life) sure sounds like the situtation.
I wish I had a scope fast enough to "SEE" the signals while operating, but that's some seriously expensive equipment...500Mhz dataline data rates would need a 2GHz digital storage scope to properly sample.
Above all dont panic if the BEEEEEEP pays you a visit.