GA-880GMA-UD2H: Random freezes with corrupt video

[bytheway_r]

You haven't mentioned your temperatures but I'd try an external GPU just in case.

DM, his PSU is much more powerful than needed. He's using only a dual-core with on-board graphics. He's probably using about 150-200W under maximum load on everything.

[Dark Mantis]

bythway_r I know you are right if you add up the figures but often I have found that especially on startup when the load is greater a significantly larger PSU is sensible.

[j5]

I can't get a temp read off of the onboard GPU. I'm not sure if there's no sensor or if the softs won't recognize it (e.g. HW Monitor)
Using Sandra, board temps are reading about 40C and CPU temps are reading in the 30s.

[Dark Mantis]

Try GPU-Z. That should give you the tempertures of the card.

[j5]

Well, the extended extended memtest came up snake eyes.

Now to determine if it's the stick or the controller.

blast!

ETA: When the first stick went bad, I ran extensive memory tests on the single stick with no issues. Either the stick deteriorated, or....something else.

[absic]

It's always difficult when you have a problem with RAM knowing exactly what is best. If you bought the original RAM modules as a kit and one stick becomes faulty it is always better to RMA all of the sticks not just one of them. Whilst in the vast majority of cases you can get away with mixing and matching RAM in reality it is not the best way.

The fact that you have now run an extended memory test and errors have come up would suggest to me that you have been unlucky with your RAM. Although the memory controller on the CPU doesn't like running RAM above 1333MHz, the truth is it is pretty uncommon for it to fail.

The options you now face though are:
1: Replacing the RAM (again) this is the option I would go for as it is more likely to be faulty RAM than the CPU.
2: Contacting AMD and seek an RMA on the processor. If you chose to RMA the CPU, it will be tested by AMD and returned if it is not faulty but if there is a problem it will be replaced. Either way you will know that your CPU is OK.

As I said right at the beginning, these kind of problems are often the most frustrating and difficult to cure because there is a 50/50 chance that you will replace one thing only to find that it is something else causing the issue. 

[Dark Mantis]

Well, the extended extended memtest came up snake eyes.

ETA: When the first stick went bad, I ran extensive memory tests on the single stick with no issues. Either the stick deteriorated, or....something else.

It sounds like from your post that yoiu just returned the one faulty module for RMA when it went. If this is so then you are now trying to run with an unmatched pair of modules which is likely to give problems.

[j5]

Yeah, I asked G.Skill if I should send both back, they didn't indicate it was needed. So I didn't.

I'm going to be systematic here with the following:

Run the new stick on memtest in channel 0
Run the new stick on memtest in channel 1
run the old stick on memtest
Relax the timings back to 9-9-9-24 and run both sticks on memtest again.

Depending on the results, I'll send both sticks back this time, regardless.

I hope the CPU is good since that is the most invasive replacement.

[Dark Mantis]

Yes as they are a matched set(or were anyway) they should always stay together. If you return any you return the whole set.

Try loosening the timings to 9-9-9-27

[j5]

My battery of tests are complete.

I have a bad channel 0

After successful completion with memtest of both single sticks  in channel1, I started fresh with channel 0 this morning.
When  I returned from work, there was no video.
Hard power down, moved the stick to the back bank of channel 0, restarted, the screen went corrupt as windows started to load.
After that, the box wouldn't even post.
Moved the stick back to channel 1, all good.
Back to channel 0, no post.
Now I have both sticks in channel 1 with no issues so far.


So...CPU or MB?

[Dark Mantis]

Most likely a failed  controller on the CPU I reckon. RMA it.

[j5]

I think I either have multiple issues or the second memory controller is going bad also.

Running all memory in Channel 1, the unit locked up (on the media center 7 screen saver) after about 2 days of continuous operation.

I was also running a disk synch on my RAID 1 partition at the time, not sure if that makes a difference.