1

Hardware questions! Watching some tech reviews about the newest and bluest Dell Poweredge servers and I noticed in this link that NVDIMM's for the new Dell Poweredge 740XD servers require R-DIMM's to operate properly. Why can't NVRAM-N DIMM's simply operate independently on the board? (What's preventing Dell from supporting say 1 NVRAM-N DIMM per CPU Socket and nothing else? How are R-DIMM's that different beyond the backup capability?) Thanks!

1

Current operating system implementation of NVDIMM-Ns shows this memory as a block device (ie storage) where you typically mount a filesystem and use it as a regular (but really fast) disk.

See Dell EMC NVDIMM-N Persistent Memory, chapters 8 & 9.

So you still need RAM for the OS.

  • Oh ok that makes a lot more sense, will there be implementations for NVRAM-N as a memory instead of storage solution? – Tmanok Feb 16 '18 at 17:26
  • 1
    Ultimately, you will need to organise data in this persistent storage. Files (and therefore file system) is a convenient way to do it. So I'm not sure that another implementation/semantic is needed at all. And you can still nmapa file located in nvdimm should your application require persistent memory. And remember this principle: en.wikipedia.org/wiki/Everything_is_a_file (not sure this applies to Windows, but it definitely does for Unix/Linux) – sfk Feb 17 '18 at 20:10

Your Answer

By clicking "Post Your Answer", you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.