We have blade server M610 (Dual CPU) with 32GB memory. Each CPU side has 6 slots to a total of 12 slots. There are 8 slots filled with 4GB modules to reach 32GB. I would like to ask, shall I install 4 modules of 8GB in the remaining 4 slots to make the total memory to 64Gb? In short, on each CPU side, there will be 4x4GB modules and 2x8Gb modules. I want to know whether this configuration is supported or not. Or should be all the slots are to be identical sizes?
Thanks for your support.
According to the information I have, the configuration that you are looking to install is not supported. The information states that the only way to install 64GB of memory in your blade server is to install 8 x 8GB DIMMs, 4 DIMMs for each processor. These would be installed in slots A2, A3, A5, A6, B2, B3, B5, B6 and needs to be set to Advanced ECC mode to get use of all 64GB in the server.
John CDell | Social Outreach Services
Get Support on Twitter @DellCaresPro Download the Dell Quick Resource Locator app today to access PowerEdge support content on your mobile device! (iOS, Android, Windows)
Thanks for the reply, John.
In this case, i have to dump all my existing 4GB memory (8 x4 gb modules)!!!
I will lose the investment protection and Dell will lose the claim of "Pay as you grow"
Remembering i have 4 physical servers and i have to throw 32 modules of 4 Gb to make it 64Gb..!
If you have 4 blade servers and all 4 blade servers have the same memory (8x4GB RDIMMs), what may be an option is to fully populate 2 servers with 4GB RDIMMs giving you 48GB in 2 servers, this would leave you with 8 x 4GB RDIMMs so you could keep one server as it is currently and then fully populate one server to 64GB using 8 x 8GB RDIMMs.
I appreciate that this may be disappointing but I'm glad you asked this before you purchased any upgrade. If you want a list of possible memory configurations, I'd advise you to look here: support.dell.com/.../HOMen.pdf (pages 203 and 204).
Hi John, thanks again.
I respect your advise, but i have an issue. I am using XenServer hypervisor and all these 4 servers are in a pool. I heard that, all servers in the pool should be in identical configuration by cpu, memory and all hardware configuraiton. Is this is right, i may not be able to apply your advise..
The "Pay as you grow" motto is always with several reservations. If you had purchased 12 1GB UDIMMs, you couldn't have expected to be able to 'grow' to 4GB RDIMMs.
You can grow up to 12 4GB RDIMMs, but if you want more memory, you have to adhere to Intel's memory population requirements.
- both CPUs need the same memory layout
- each CPU has 3 memory channels and each channel needs to have the same memory layout or nothing (unused channel)
So 8 4GB dimms is supported if it'll be set up using (per CPU) 2 channels with 2 DIMMs of 4GB each. The 3rd memory channel is unused.
This is why you cannot have 8 x 4GB and 4 x 8GB, as each CPU has 6 slots, and each pair needs the same set of memory.
So you could do (per memory channel) 1 x 8GB + 1 x 4GB (remove 2 x 4GB from each system). This would yield 2 CPUs x 3 memory channels x 12GB = 72GB.
Member since 2003
Very promising suggestion!
Correct me if i am wrong. I will populate the memory as follows:
In this case i will lose only 2 4GB dimms in each server. Please update. Thank you again