MP 1,1-5,1 5,1 ram mismatch

[basslik]

I have three different pairs of memory in my Mac Pro. speed and cas Latency are the same, even the model numbers are almost the same except a letter here or there different. Total of 128.

I read on OWC that the memory sticks should all match?

I really haven't had any issues so far except the usual hiccups pro Tools is known for.

I do have 48 megs from OWC. Should I use it instead of the mismatch, even though it's half the memory, yet running triple channel.

THANK YOU

 

[h9826790]

In general, as long as the cMP can boot, you are good to go.

Use all identical DIMM is a precaution, not requirement.

 

[basslik]

Thanks for the info Martin. Greatly appreciate it.

 

[tsialex]

In general, as long as the cMP can boot, you are good to go.

Use all identical DIMM is a precaution, not requirement.

Different DIMMs have different SPDs, more data cached inside MemoryConfigs and less space avaliable inside the NVRAM main VSS store.

Several people here that have 8x16GB RDIMMs with different models have just ~20KB free right after a multiple NVRAM reset, while people with 8x16GB Samsung RDIMMs have ~35KB, also after a a multiple NVRAM reset. ~20KB is not enough for sealed containers software updates staging and GC will have to run while the updates are also running - this is a brick paradise.

Nowadays, I'd call it a basic requirement that the DIMMs must be identical and from a model that don't have any form of Intel XMP, like several Kingston and Hynix have. I've helped @Antc172 last week with this exact problem, 6x16GB Hynix HMT42GR7AFR4C-RD installed and after a complete reset his mid-2010 MacPro main VSS store had just 18646 bytes available. Completely reproducible, this one seems to be the worst case that I've found recently.

 

[basslik]

Different DIMMs have different SPDs, more data cached inside MemoryConfigs and less space avaliable inside the NVRAM main VSS store.

Several people here that have 8x16GB RDIMMs with different models have just ~20KB free right after a multiple NVRAM reset, while people with 8x16GB Samsung RDIMMs have ~35KB, also after a a multiple NVRAM reset. ~20KB is not enough for sealed containers software updates staging and GC will have to run while the updates are also running - this is a brick paradise.

Nowadays, I'd call it a basic requirement that the DIMMs must be identical and from a model that don't have any form of Intel XMP, like several Kingston and Hynix have. I've helped @Antc172 last week with this exact problem, 6x16GB Hynix HMT42GR7AFR4C-RD installed and after a complete reset his mid-2010 MacPro main VSS store had just 18646 bytes available. Completely reproducible, this one seems to be the worst case that I've found recently.

Alex the reconstruction that was accomplished you mentioned that you adjusted the Bootrom.
Ok, that explains the mismatch. I'll adjust the BootROM image to the case.

I do have 48 gig all matching OWC ram I can swap out if that is a better route ?

 

[tsialex]

Alex the reconstruction that was accomplished you mentioned that you adjusted the Bootrom.

Ok, that explains the mismatch. I'll adjust the BootROM image to the case.

This has nothing to do with the BootROM clean-up/upgrade/reconstruction on itself, but the way the memory configuration for each DIMM is stored inside the NVRAM and the high usage of the NVRAM when staging software updates since Big Sur, like I wrote earlier.

I do have 48 gig all matching OWC ram I can swap out if that is a better route ?

You should investigate which RAM configuration/set of DIMMS works best for your Mac Pro. Just do a deep NVRAM reset after a week and then check if one or another set of DIMMs have more available space.

 

[basslik]

That's a great idea. I have a particular Pro Tools session that force quits every once in awhile with a lot of plugins. I'll see if it makes a difference.