ECC Errors - Which RDIMM?

[keith]

Unfortunately, I checked with Supermicro and both boards only support a single DIMM config by populating slot DIMMC1. IE, I can't put a single stick in slot DIMMD1 and leave the rest of the slots empty.

What about 2 sticks? Or just odd or just even slots?

Even if it were possible to do single DIMM configs with each available slot, I still don't understand how this would work since the MemTest logs are always showing the correct channel mappings EXCEPT when displaying an ECC error. I'm guessing you missed my attempts to clarify this earlier in this thread. I'd appreciate it if you could answer this as I want to make sure I'm not totally confused here.

I believe you are referring to the SMBIOS details such as "BankLocator: P0_Node0_Channel3_Dimm0". This is unrelated to the chipset internal memory controller channel mappings, which report the "incorrect" channel (this is not as apparent in the logs). As mentioned previously, there is a possibility that the mapping of the physical DIMM slots (ie. the SMBIOS info you are referring to) to the chipset memory controller may not be one-to-one. By installing DIMMs in different physical slots, we can see which internal memory controller channel it is mapped to by looking at the logs (again, not the SMBIOS details such as "BankLocator: P0_Node0_Channel3_Dimm0").

[lunadesign]

What about 2 sticks? Or just odd or just even slots?

Unfortunately, no. They've got specific positions for the sticks in the 1, 2 and 4 stick cases.

Instead, I have been testing with 8 DIMMs and moving the problematic DIMM through all 8 slots. I should be done with that in an hour or two and will send you the logs.

I believe you are referring to the SMBIOS details such as "BankLocator: P0_Node0_Channel3_Dimm0". This is unrelated to the chipset internal memory controller channel mappings, which report the "incorrect" channel (this is not as apparent in the logs). As mentioned previously, there is a possibility that the mapping of the physical DIMM slots (ie. the SMBIOS info you are referring to) to the chipset memory controller may not be one-to-one. By installing DIMMs in different physical slots, we can see which internal memory controller channel it is mapped to by looking at the logs (again, not the SMBIOS details such as "BankLocator: P0_Node0_Channel3_Dimm0").

Thanks for explaining! You are correct, I've been using the BankLocator lines to know which DIMMs are in which slots since they indicate the slot names (i.e., "DIMMC1"). Since they also mention "channel" I thought that these were the same "channels" that the ECC errors were identifying. I would have never guessed there would be two sets of unrelated memory "channels" in the same system. If there's any way you can clarify that in the logs, that would be super helpful to prevent people from falling into the trap I did.

Question: For future reference, can you please explain where the "internal memory controller channel" information is in the logs?

[lunadesign]

I've finished my testing with 8 DIMMs and rotating a known problematic DIMM through all 8 slots.

Here's what I found:

MB Slot	Channel according to BankLocator	Channel according to ECC error
A1	0	0
B1	1	1
C1	2	3
D1	3	2
E1	4	6
F1	5	7
G1	6	5
H1	7	4

Keith -- I've sent all of the data from these test to the PassMark support e-mail address. Please confirm you received it!