Announcement

Collapse
No announcement yet.

AMD Launches EPYC Embedded 3000 & Ryzen Embedded V1000 Series

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • AMD Launches EPYC Embedded 3000 & Ryzen Embedded V1000 Series

    Phoronix: AMD Launches EPYC Embedded 3000 & Ryzen Embedded V1000 Series

    AMD is taking their Zen microarchitecture to the embedded space now with the announcement of the AMD Launches EPYC Embedded 3000 and Ryzen Embedded V1000 series...

    Phoronix, Linux Hardware Reviews, Linux hardware benchmarks, Linux server benchmarks, Linux benchmarking, Desktop Linux, Linux performance, Open Source graphics, Linux How To, Ubuntu benchmarks, Ubuntu hardware, Phoronix Test Suite

  • #2
    A listed feature of both the EPYC and the Ryzen, from the linked article:
    On-board secure processor for crypto co-processing, SME to defend against unauthorized physical memory access, and SEV for encrypting VM memory to help protect against various administrator attacks without disrupting application codes.
    That was the first thing I checked for. It's the first thing I check for on every new CPU I look into.

    I want to know:
    • Is the source code and tool chain available for this secure processor?
      • If yes, can we compile the code ourselves and have our processor run our personally compiled version?
      • Can we modify this code and run the modified version?
      • (so pretty much the 4 essentials freedoms)
    • What features does this processor have? In particular, can it interface with networking hardware without the host OS knowing?
    • Presumably user-land or at least kernel-level code can interact with this co-processor: how well has it been audited to ensure it can't be accessed and controlled by a malicious third party?
    Until these questions have been asked positively, I will look sceptically upon this, otherwise incredible, hardware that both AMD and Intel are producing.

    I'd love for AMD (or Intel) to take a risk and produce just 1 privacy, freedom and security focused, modern CPU. I'd buy it and encourage every soul I met to do the same to help make it a commercial success and signal to the market that we really want this stuff.

    Comment


    • #3
      The V1000 is yet another set of products based on the Raven Ridge die. This one fills the pretty large gap between the 12W-25W Ryzen Mobile parts and the 65W Ryzen APUs.

      I don't know if AMD has plans for a NUC style device, but a part with this power envelope would work great for one. Being embedded these processors probably cost more than a consumer part would be, so I'm not expecting AMD NUCs to start rolling out with this.

      I'm kind of surprised no motherboard makers have put out one of their own already based on a Ryzen Mobile part. A Ryzen Mobile 2700U based sub-ITX desktop would be a much better tiny gaming machine than any Intel NUC.

      Comment


      • #4
        Meanwhile, Ryzen CPUs are still plagued by this: https://bugzilla.kernel.org/show_bug.cgi?id=196683

        AMD didn't really ever properly comment on this issue.

        Comment


        • #5
          Originally posted by shmerl View Post
          Meanwhile, Ryzen CPUs are still plagued by this: https://bugzilla.kernel.org/show_bug.cgi?id=196683

          AMD didn't really ever properly comment on this issue.
          Something tells me we can dig up some pretty terrible Intel bugs they haven't properly addressed. This appears to be an attempt to derail a conversation about a new product with what-about-ism. I hope this line of discussion does not continue here. It's off topic.

          Comment


          • #6
            Originally posted by existensil View Post
            A Ryzen Mobile 2700U based sub-ITX desktop would be a much better tiny gaming machine than any Intel NUC.
            There are a lot of places where such setup would be _very_ useful. We recently benchmarked few boards and found the GPU performance to be inadequate.

            Comment


            • #7
              Originally posted by existensil View Post

              Something tells me we can dig up some pretty terrible Intel bugs they haven't properly addressed. This appears to be an attempt to derail a conversation about a new product with what-about-ism. I hope this line of discussion does not continue here. It's off topic.
              The question is whether it affects any other AMD chip. I see no point in pretending it's a non existing problem.

              Comment


              • #8
                Originally posted by shmerl View Post

                The question is whether it affects any other AMD chip. I see no point in pretending it's a non existing problem.
                Be ignorant of the OP retort. Let's go down the list of equivalency to Intel issues. I'll wager it's a mountain of BS that awaits for you to swallow.

                Comment


                • #9
                  Originally posted by shmerl View Post

                  The question is whether it affects any other AMD chip. I see no point in pretending it's a non existing problem.
                  Had Ryzen since launch, Never had that problem TBH.

                  Built 2 others, both with ryzen CPU's , Again no issues.

                  So the bug is rather random it would seem.

                  Comment


                  • #10
                    Originally posted by Marc Driftmeyer View Post

                    Be ignorant of the OP retort.
                    Did you see extensive testing on this? OP report doesn't say anything of the sort. So contribute something useful to this, unless you have no data either.
                    Last edited by shmerl; 21 February 2018, 03:28 PM.

                    Comment

                    Working...
                    X