Announcement

Collapse
No announcement yet.

AMD Announces Ryzen 7000 Series "Zen 4" Desktop CPUs - Linux Benchmarks To Come

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

  • #71
    Originally posted by numacross View Post
    ~9% of mobile ones have it enabled.
    Since the statement was that a W680-chipset motherboard was needed, I don't know why anyone would assume the mobile parts had support.

    Anyway, thanks for the info. That's disappointing about the lower-end desktop die.

    Comment


    • #72
      Originally posted by Dukenukemx View Post
      Probably because Ryzen 7000 series are not mobile parts.
      I was talking about the Zen 4 core, but obviously in its laptop incarnation. It's just speculation, though. It should become clear, in due time.

      Don't forget that Apple's cores were designed, from the ground-up, with efficiency as the number 1 priority. Also, Apple can afford to trade die size for energy-efficiency more so that AMD.

      Comment


      • #73
        Originally posted by coder View Post
        If you look above the table, the Zen 4 entry cites several sources, including a TechPowerUp interview with Robert Hallock, in which he specifies VNNI and BF16 will be supported. BF16 seems the only controversial part, since everything else is just replicating what Intel has supported since Ice Lake. If we're lucky, the current list is still incomplete!
        I missed that, thanks. Looking at those sources does not inspire confidence: the first one is a rumor, the second is hearsay, the third is a leak with the instruction list being "As covered by other outlets" but doesn't which ones, and the last one being a trusted TPU interview only mentions VNNI and BFLOAT16. Which leaves us with no official confirmation beyond the last source.

        Originally posted by coder View Post
        Also, it would be a little surprising for AMD to keep this stuff under wraps. Intel usually publishes ISA support of their products a couple years in advance, so that more software is ready by the time it launches.
        I am disappointed by this as well. There isn't much stuff about "znver4" either, so it might take a while for proper compiler support.

        Thanks.

        Comment


        • #74
          Originally posted by coder View Post
          Since the statement was that a W680-chipset motherboard was needed, I don't know why anyone would assume the mobile parts had support.
          Those are being supported by the WM690 chipset.

          Comment


          • #75
            Originally posted by coder View Post
            That can't be right, because the P-cores are essentially the same as Alder Lake's. The only thing that should help Raptor Lake's IPC are bigger L2 caches and tweaks to the ring bus. Those are probably worth a few % at best, and nowhere close to 15%.

            Perhaps what you're thinking of is single-thread performance, which is approximately the product of IPC ratio and clock speed ratio. If that's what you mean, then you have to compare it with AMD's claim of 29% faster single-threaded performance than Zen 3. Using your figure of 2% better IPC than Alder Lake would give Zen 4 a 16.4% single-threaded advantage over Alder Lake.


            Since you can't buy either, right now, might as well wait and see. What if Intel's stated numbers require a custom water cooling loop + a $1k external chiller to achieve, in real life?
            They're not the same P-cores, Alder Lake used Golden Cove while Raptor Lake is using Raptor Cove cores. While they both use the same Gracemont E cores. The L2 caches changes from WIllow Cove to Golden Cove introduced a 20% IPC increase. Alder Lake was ~30% faster than Zen 3 in single-threaded performance due to the 11% better IPC/clock speed difference. According to their benchmarks Zen 4 is 6.6% faster in single-threaded performance.

            Edit: I want to add I already have a custom loop, only thing I need is a $4 back plate for my water block.
            Last edited by WannaBeOCer; 30 August 2022, 03:22 PM.

            Comment


            • #76
              Originally posted by WannaBeOCer View Post
              They're not the same P-cores, Alder Lake used Golden Cove while Raptor Lake is using Raptor Cove cores.
              From what I've read, the Raptor Lake P-cores and E-cores are both essentially the same as Alder Lake.

              Anyway, I'm not going to argue over rumors, leaks, and speculation. Believe whatever you like -- we'll find out the truth in a couple months.

              Comment


              • #77
                Originally posted by coder View Post
                From what I've read, the Raptor Lake P-cores and E-cores are both essentially the same as Alder Lake.

                Anyway, I'm not going to argue over rumors, leaks, and speculation. Believe whatever you like -- we'll find out the truth in a couple months.
                Pretty sure most of the Raptor Lake performance increases are coming from increasing clock speeds and adding extra e-cores. Not IPC gains.

                Last rumor I saw was 8-15% faster single-threaded, which would probably put it in the 0-10% faster than Zen 4 range. And 30-40% faster multi-threaded, which would maybe put it 0-10% slower vs Zen 4? It's hard to say, we'll just have to wait for benchmarks and it will probably depend on what you want to do with the chips at each tier. It sounds like AMD may have a X3D part coming pretty soon that would dominate gaming.

                Comment


                • #78
                  Originally posted by smitty3268 View Post
                  Pretty sure most of the Raptor Lake performance increases are coming from increasing clock speeds and adding extra e-cores. Not IPC gains.
                  Yes. Also 2x L2 cache and faster DDR5 speeds. Each worth a couple %, at most.

                  Not to knock Raptor Lake. It'll be nice to see a "tuned up" version of Alder Lake, with some of the kinks ironed out. However, it's disappointing to hear how much they're raising power limits (even with a 350 W enthusiast option!), apparently relying on extreme overclocking, in a desperate attempt to keep a lead over Zen 4.

                  It seems particularly tone deaf, in these days of high energy prices and ecological disasters becoming more commonplace.

                  Comment


                  • #79
                    Originally posted by atomsymbol

                    Well, but does the transition from x86 to ARM/RISC-V have more momentum and more potential than the transition of IPC (instructions per clock) from approximately 1.5 instr/cycle to 10 instructions/cycle? All general-purpose high-performance CPU cores (ARM, RISC, x86) are undergoing the transition to higher IPC.

                    The fact is that neither the legacy features of x86 cores (such as: segmentation), nor (from programmer's perspective) the variable-lengthiness of the x86 ISA, nor the smaller number of 64-bit general-purpose registers ---- none of that has proved to be blocking x86 from regularly achieving higher IPC with each new x86 CPU core generation.
                    ARM/RISC-V are doing it at lower power/higher efficiency, that's why.

                    Comment


                    • #80
                      Oh actually, are these latest platforms still as slow to boot? I think my i7 spend maybe 5 seconds at most from power to GRUB whereas my Ryzen takes a lot lot longer. It's not a big deal since I hardly reboot, but it'd be nice if they fixed that and it's kind of annoying how the Linux ecosystem only gets faster but the hardware maker the other way.

                      Comment

                      Working...
                      X