CPU Channel 1 memory error

[AltecBX]

So I found the event logs in the IPMI. Here it is:

 

[mjws00]

That narrows it. Can't recall if you shifted slots, but I believe so?
So you get another option. Pull that dimm, and see if everything runs perfect. If so.... voila. Else... back to elimination starting with one memory variable at a time. 1 dimm, 2 dimms etc.

 

[cyberjock]

What I'd do is switch B1 DIMM with another DIMM. If it moves it's the RAM, if it doesn't move it's something else (probably the slot).

 

[pschatz100]

I can't speak to DDR4 and this motherboard specifically, but I have seen many cases where systems get "fussy" when one uses four or more memory sticks. It is not always a matter of something being bad (although it certainly could be.) All electronics are built to certain tolerances and with memory in particular, if some of the sticks are operating at the edge of their rated performance range then you can get random errors that are difficult to diagnose when using them in groups.

Perhaps you should try running the memory at a slower speed. Easy to do, and if the problem goes away then you know you've got timing issues.

 

[Ericloewe]

I can't speak to DDR4 and this motherboard specifically, but I have seen many cases where systems get "fussy" when one uses four or more memory sticks. It is not always a matter of something being bad (although it certainly could be.) All electronics are built to certain tolerances and with memory in particular, if some of the sticks are operating at the edge of their rated performance range then you can get random errors that are difficult to diagnose when using them in groups.

Perhaps you should try running the memory at a slower speed. Easy to do, and if the problem goes away then you know you've got timing issues.

Nehalem had trouble keeping memory stable at stock voltages with two DIMMs per channel, but it's been a while since I've seen something of the sort (except for the whole Kingston story...). In any case, since we're looking at one DIMM per channel, that's unlikely to be the case.

 

[Knowltey]

Yep, bad slots can be surprisingly common, get them all the time at work. Our troubleshooting process for RAM follows to have them do 1 stick of RAM at a time, see which one is causing errors/now showing to the system. Then once they've pinpointed that have them stick the RAM in a slot that was already verified as good with a different stick and see if the problem goes away. If it does, then take a stick that was already earlier verified to be a good stick and stick it in the suspect spot and see if the problem comes back.