X58A UD3R DMI Data pool

[Gary123]

Yes I understand but it is even easier to just press F12 to boot from some other device that once rather than having a slow boot every time.

Thanks! I'd forgotten all about F12.

[Dark Mantis]

That's often the problem isn't it. You tend not to use something for ages and then can't remember how to do it when you want to.

[Trinitrotoluene]

Your BSODs form what I could see on the other forum could well still be memory linked. Are your memory modules listed as compatible with this motherboard?

You could be very right. I ran out of options on whatever I could try software wise, updating driver, deleting and cleaning software that could possibly mess up. I ran memtest again yesterday, errors. Rebooted, ran it again: errors. Took all modules out, swapped them in between slots, overnight run: no errors. Weird indeed. Memory doesn't just die and then decide to get back to work.

Anyway, not the right topic to start a new subject here. Though any comments on this are welcome. But I guess the software side of troubleshooting here is irrelevant and not necessary as long as the memory keeps acting up. Read more on this in my thread on the OCZ forum if you like.

http://www.ocztechnologyforum.com/forum/showthread.php?78922-Possible-faulty-module&p=569224&viewfull=1#post569224

[Dark Mantis]

Thanks for the update and of course if anything progresses please let us know.

[Trinitrotoluene]

Progress!  

I ran prime95 blend overnight for the past two nights, no problems or crashes. So CPU and memory modules are OK.

Inspected the memory slots on the motherboard, everything looked fine. Did find a very small piece of material in slot #1. That may indeed have caused all this trouble, though I'm not sure yet. Doesn't seem to be in line with everything. Though it might explain that reseating the modules resolved the errors in memtest a while ago.
One would say that if the #1 module is obstructed in its connection with the motherboard, you'd get 0-2000MB errors in memtest. Which I did get, but also above 2000MB. Or perhaps triple channel setup might explain that. Anyway, I'm sure all the slots are clean and in perfect condition now.

Will do some more stress testing, also from cold boot after leaving the PC off for a good night. The most important thing is getting the system to run stable again. But I do want to know for sure what caused all this trouble (if resolved now...). Kind of pisses me off not knowing it (yet) haha  

Finally OCZ told me there's still the option left to fiddle around with the QPI/vtt voltage a bit, if problems persist. Some boards/modules like higher values, some like lower.


Fingers crossed! I always thought blue was my favorite color, but I'm starting to dislike it lately...

[Dark Mantis]

Well I am happy that things seem to be progressing for you and the stability is better. It could well have been the foreign body in the memory slot that was causing your problems, that's all it takes.

I always thought blue was my favorite color, but I'm starting to dislike it lately...

I think they must have heard you because it looks like the next generation of boards is going to be black

http://forum.giga-byte.co.uk/index.php/topic,2849.0.html

[Trinitrotoluene]

Indeed! I really hope that was the problem. Good to have been brought to my attention, and other google users out there, the importance of absolutely clean memory slots. So worth double checking.

Black looks so nice as well. I don't dislike the blue color of the UD3R board. The color of the BSOD was what I meant, sorry!
I'll get back to this thread after a while of having a stable running system. In case I forget and don't post back here, that would be a good sign

Thanks again for the support.

[Dark Mantis]

I really don't know what you find to dislike about the blue of the BSOD

[Trinitrotoluene]

Oh boy, if only that was a touchscreen and I could press ANY key to continue...

What?! It IS?!

[absic]

Hey DM, does your wife know that you have this kind of image on your PC?

[Dark Mantis]

Don't know what you mean absic....that is the wife

[absic]

In your dreams! 

[Dark Mantis]

Yeah that's what she said as long as I'm only dreaming it   

[Trinitrotoluene]

Problem not yet solved. I still get blue screens, system service exception and PFN list corrupt are the latest. I bet they are memory related. Everytime on cold boot I get errors in memtest. I tried QPI/VTT voltage 1.360, 1.340. 1.320 and 1.300. Always errors in memtest during coldboot. When it is warmed up, it doesn't seem to give errors in memtest but it isn't stable either. Now please let me copy paste the following from my thread on OCZ.

So this morning I did another memtest coldboot run, QPI/VTT voltage on 1.300V.

http://img823.imageshack.us/img823/5422/20101104085244.jpg
http://img143.imageshack.us/img143/1632/20101104104509.jpg
(both screenshots from same run)
The first time I started memtest, the screen stayed blank; normal memtest screen, without any content. The 2nd time it worked, got errors quite fast as you can see. And they seem to be the same as here.

update: I just decided to restart the system, after using it for 2 hours or so. Had a BSOD when it was about to restart. System_service_exception 0x0000003b, not the first time I got that one (also with qpi/vtt voltage on 1.360v).
Ran memtest again, came up with errors. Keep in mind this time it wasn't a cold boot. Picture below.

http://img8.imageshack.us/img8/3991/20101104124038.jpg

And another BSOD, PFN list corrupt. First my firefox crashed, then that error. Ran memtest again, QPI/VTT voltage on 1.360 this time in case it was too low, but no joy:

http://img149.imageshack.us/img149/5263/20101104130703.jpg
http://img573.imageshack.us/img573/5605/20101104130831.jpg

more updates: I also ran each module individually in every white slot (1,3,5) and gave it one pass in memtest, no errors in any slot or on any module. Though the first time I put module B in slot 3 (2nd white slot) memtest would be totally blank, same thing as I mentioned before. Then put back all the modules and also gave them one pass in memtest, no errors.

There seems to be no logic, except for coldboot memtest errors that do seem to persist and also BSOD's are still here. If you ask me, those two facts should be linked together.

Remaining QPI/VTT voltages I can try are: 1.280v 1.260v 1.240v 1.220v 1.200v and of course anything lower and higher, but I don't think we want that. I started testing at 1.360v and went down.

But what does all this tell you?

[Dark Mantis]

As you say there doesn't seem to be any logic to it, but it might be worth trying to loosen off the timings a little to 8-8-8-24 and run them at the recommended voltage with QPI/Vtt voltage set just 0.5V below.