WTF did I do to my computer?!

[disko33]

Ok, this could be a fairly long post, so be warned!

 

A bit over a week ago, I get a new Opteron 165 from Newegg to put in a computer I am building for my parents. It turns out to be a CCBBE0610 stepping. I figure since I am not planning on OC'ing their computer very much, so I will take out my CCBWE0550 stepping from one of my computers, and give that to them, and take the 0610 for myself. Very simple process, pop out the 0550 and put in the 0610, right?

 

I leave all the BIOS settings the same, since a 0610 *should* be able to handle the 2610mhz that I was running the 0550 at (@ stock voltage). Well, turns out its not stable. So I drop all timings and everything back to stock. Still not stable - at 1800mhz! I figure I must not have seated the heatsink properly or something - there is no way it can't be stable at 1800mhz. Take off heatsink, take out CPU, clean off and reapply thermal paste, etc etc. Still can't get it stable. Clear CMOS, and load optimized settings. Still not stable. Temps are all fine, low 30's, but for some reason I keep getting random reboots and freezes. So I think maybe I just got a bad CPU. Switch the 0610 with the 0615 from my other computer. The 0610 now works fine in my other computer at 2700mhz, now the 0615 (which was running great at 2700mhz @ stock voltage in other computer) is now not stable even at stock 1800mhz speed. I now figure maybe something weird happened to memory, even though I didn't touch it in the move. Swap out the memory between my two computers. Still getting random freezing and reboots. Both pairs of memory sticks run fine and pass Memtest in my 'South' computer, both fail in my 'North' computer. When I put them in the yellow slots, the computer will not even get past the BIOS screen, it just freezes. In the orange slots, it will boot, but usually will freeze up Windows.

 

Does anyone know what I could have possibly done to cause this behavior? It really seems like it is something wrong with the board (DFI DR-Expert). The odd thing is that it is on a motherboard tray in my case, so I literally didn't even touch the board itself, the tray just slides out, I unscrew the heatsink from the CPU bracket, and screw it back in. I have cleared CMOS, and loaded optimal settings, done everything I can think of - but with no success. Any advice would be greatly appeciated!

[Outlander]

I personally like to load optimized defaults in BIOS when installing a new chip and bringing up the computer at least once in this state before overclocking the new chip. Not to mention clearing CMOS before installing the new chip. Of course this probably is not going to help you now. Try to load optimized defaults and see what happens. YMMV. Good luck.

[tasr]

OK any time you change CPU, Video cards or RAM do yourself a favor.

 

Clear the CMOS > Load Defaults > Shut Down > Change Components > Boot Clear CMOS > Load Defaults and boot.

 

Some will disagree. Takes 5 mintues not including changing the components.

[disko33]

Yeah, under normal circumstances I would have done just that, like changing from a single core to dual core, or even just a different CPU. However, this time I figured since they were both Opty 165's, I would be ok just changing it out as is.

 

I have since cleared CMOS and loaded optimized defaults, and am still having serious issues. Assuming the CPU and memory are both ok - as I have tested them in my other computer - it seems that the motherboard is jacked. Is it really possible I messed something up just by changing the CPU without clearing CMOS first? Just seems kind of odd to me.

[tasr]

Pull the chip and inspect it for bent pins. Also did you try a long CMOS clear?

[disko33]

Like I said, the same thing is happening with 2 different CPU's, so its definitely not bent pins. No I haven't tried a long CMOS clear yet, but I will try it overnight tonight - although I'm not holding out too much hope that is the problem.

[noneed4me2]

Its gotta be the board but as to what it is i am not sure, I would bet the psu is failing maybe. Try swapping those components, and its funny how many times people tell you to "reset the cmos" when you already posted that is what you did. Hope this is a more constructive suggestion. Swap PSUs without swapping anything else and see if instablity switches.

[ReelFiles]

Your Windows install might have gotten corrupted from running it unstable and crashing.

[PhoetuS]

Your Windows install might have gotten corrupted from running it unstable and crashing.

 

I second this notion...

 

While you are eliminating factors, it might be worth running memtest to make sure the ram is ok.

[disko33]

That's the thing, I have been running Memtest and getting all kinds of errors, and it is happening with both sets of sticks. When I switch the sticks to my other computer and run Memtest - no errors - so I don't think the actual memory is bad. The crazy part about it is that I am even getting errors running DDR500 RAM at DDR400 speeds at 3-4-4-8.

 

As far as the power supply, I can definitely try switching it out I guess, but again I have a really hard time believing that all of the sudden - at the exact same time I switched the CPU - it started failing out of the blue. Its less than a year old, and Seasonics are supposedly pretty good power supplies. But I will try that just to rule it out. If that isn't it, I'm going to have to RMA the board I guess because if the CPU, RAM and power supplies are all known to be good in other systems - it has to be the board, right?

[PhoetuS]

if the CPU, RAM and power supplies are all known to be good in other systems - it has to be the board, right?

 

Yep, if everything else tests ok, then it has to be the board.

 

 

At this point, having ruled out everything else, you have very limited choices. If you want to try one last thing before an RMA to DFI, you could try contacting Tmod about getting a new pre-programmed BIOS chip. It could be that the BIOS got totally fuxxored to the point that a CMOS clear won't fix it. IF that is the case it should fix it.

 

But, if that isn't the case, & it doesn't fix the problem, then you will still have to RMA it...

[disko33]

Alright, here is where I am at now with the problem. I basically took everything out of the case, kept it to just mobo, ram, cpu, video card, main hard drive and the power supply from other computer. Had cleared BIOS overnight, so loaded the optimized defaults, and enabled Memtest. Turned on computer, let it run 2 passes of Memtest, and passed with no errors. Good sign, much better than before when it failed Memtest with tons of errors. Excited, I then restart, go into BIOS, disable Memtest, restart again, and go into Windows. I have the drive partitioned as 2 drives, one for XP, one for Vista. I select XP from boot menu, kicks me out almost instantly with some message about the memory, hardware change, turning off shadowing in the BIOS, etc - I don't remember the exact message, probably should have wrote it down, but of course I didn't. Anyhow, restarted and tried Vista instead, thinking maybe I had jacked up my Windows XP install. Actually got to the Vista desktop, before it froze on me.

 

So, after all this, does it sound like my hard drive is now corrupted and will have to reinstall Windows, or something likely wrong with the board? I'm pulling my hair out over this crap! As far as the BIOS, I don't know how it could have got messed up just going from one Opty 165 to a different one - but I guess I could try that.

 

If I do end up having to RMA the board, how quick is DFI's RMA touraround usually? Also, I installed an Evercool chipset fan a while back, would I have to take that off and put the original back on, or would they just repair the board and send it back as is?

 

By the way, I appreciate everyones help and suggestions - this board rules!