Zen 2 Platform for HEDT - Improvements over Last-Gen

Section by Andrei Frumusanu

The platform architecture of the new Threadripper 3000 series is fundamentally different, and a massive departure from what we’ve seen in the past on the first and second generation Threadripper products. Previously, AMD still made use of its monolithic die design used in Zen and Zen+ Ryzen, Threadripper and EPYC products. The approach was an economically smart one for AMD in regards to having to design only a single silicon die that would be used across the three product lines, however it had some fundamental technical disadvantages when it came to power efficiency as well as having to make some performance compromises.

The biggest disadvantage exhibited by the Threadripper 2000 series was the platform’s weakness in regards to its memory architecture, an issue that was particularly prevalent in the 32-core Threadripper 2990WX. As explained in our review of the TR2 products last year, the main problem with that SKU was that in order to achieve a 32-core product, AMD had to make use of 4 “Zeppelin” dies. Unlike the server-oriented SP3 socket however, Threadripper products come on the TR4 platform. While the two sockets are physically identical, they’re electrically incompatible with each other. In practice, the biggest difference between the two platforms is the fact that Threadripper products supports 4-channel memory setups, while the EPYC variants support the full 8-channel memory configuration possible.

The main conundrum for a product such as the 2990X which had to make use of 4 dies, each integrating 2 memory controller channels, is the decision on how you split up the memory controller setup between the dies and choose which 4 active controllers you’ll end up using. AMD’s approach here is that instead of using only one memory controller per die, the company chose to have two dies each with both memory controllers active, while the other two dies wouldn’t have any memory controllers enabled at all. The issue here is that the CPUs located on these dies would only have to access memory by hopping through the infinity fabric to the adjacent dies which did have memory controllers, and incur quite a large memory latency and bandwidth penalty. This penalty was large enough, that in situations where applications weren’t properly NUMA-aware and scaled across all core, the 2990WX ended up sometimes lagging behind the 16-core 2950X in performance.

Chiplet Architecture To The Rescue

Of course, AMD was aware of this drawback, but wasn’t planning to stay with this compromise forever. The new Ryzen 3000 series earlier this summer introduced the chiplet architecture for the first time ever, with some quite astounding success. The main differences here is that AMD is decoupling the actual CPU cores and cluster from the rest of the traditional SoC. The CPU chiplet contains nothing more than the CPU cores themselves, the CPU clusters L3 caches, and the I/O interface which communicates with the rest of the “traditional” system, which is now located on a separate silicon die.

AMD Ryzen 3000 Consumer IOD - Credit Fritzchens Fritz

For the Ryzen 3000 products, this I/O die is seemingly quite familiar in terms of design to what we saw in the first- and second-generation Zen architecture products. We find your various I/O IP blocks which take care of various connectivity such as USB, Ethernet, SATA, alongside the critical components such as the PCIe controllers and of course the memory controllers. In general, what’s found on the Ryzen 3000 IOD isn’t all too different in functionality than what we previously saw on the monolithic Zen dies from past years – of course, except for the CPUs themselves.

AMD EPYC2 / Threadripper 3000 sIOD - Credit Fritzchens Fritz

As we move on to the new Threadripper 3000 products (and new EPYC 2 processors), we however see the AMD’s main chiplet design advantage. Although the new Threadripper and EPYC products use the very same 7nm CPU chiplet dies (CCDs), they are using a different IO die, what seems to be called by AMD as the sIOD (server IO die?).

What’s interesting about the sIOD is that it’s not much of a “monolithic” design, but actually more similar to four consumer IO dies put together on one chip. In the above die shots (credit to Fritzchens Fritz), we actually see that AMD is employing an identical physical design of large parts of the chip’s IP blocks, with the main "central" block cluster going as far as being essentially identical. Of course, the layout of the various surrounding blocks is quite different. AMD here is essentially reusing design resources across its product ranges.

While the chip isn’t completely mirrored – there are still distinct unique IP blocks on each quarter of the die, it is in fact correct to say that it’s divided into quarters. These “quadrants” are in fact physically and logically separate from each other. Where this is important to consider, is in regards to the memory layout. In fact, logically, the layout is actually quite similar to what we’ve seen on the previous generation Threadripper and EPYC chips in terms of memory controller and CPU cluster distinction. Each quadrant still has its own two local memory controller channels, and the CPU CCXs connected to this quadrant have the best latency and bandwidth to memory. The CPUs accessing memory controllers of a different quadrant still have to do this via a hop over the infinity fabric, the biggest difference for this generation however is that instead of this hop being across different dies on the MCM package, it all remains on the same silicon die.

For Rome, AMD had explained that the latency differences between accessing memory on the local quadrant versus accessing remote memory controllers is ~+6-8ns and ~+8-10ns for adjacent quadrants (because of the rectangular die, the quadrants adjacent on the long side have larger latency than adjacent quadrants on the short side), and ~+20-25ns for the diagonally opposing quadrants. While for EPYC, AMD provides options to change the NUMA configuration of the system to optimize for either latency (quadrants are their own NUMA domain) or bandwidth (one big UMA domain), the Threadripper systems simply appear as one UMA domain, with the memory controllers of the quadrants being interleaved in the virtual memory space.

The interesting question here of course is, how is this UMA domain setup for the Threadripper 3950X and 3970X? The SKUs come with 4 chiplets each, with the 3950X employing 3 cores per CCX, totalling 24 cores, and the 3970X employing 4 cores per CCX, totaling 32 cores. However, what we don’t know is how these chiplets are divided and populated across the sIOD’s quadrants. In theory, one could have one chiplet and one memory controller per quadrant – or one could have just two fully populated quadrants with the other two quadrants disabled. Given we have numbers on a fully populated EPYC 7742 to compare against, and that the diagonally opposing quadrant latency penalty is quite big, we should be able to estimate the implementation based on the latency results.


Looking at the latency results, there’s a few comparisons to make. In regards to the L1, L2 and L3 performance, I refer to our original Zen2 analysis in our Ryzen 3000 review article. The numbers here don’t change, which is natural as we’re talking about the very same CPU chiplet across the different product lines.

Going out of the CCD, the DRAM latency is the most interesting difference that we need to have a closer look at. Comparing the new Threadripper 3970X to the 2950X we see a latency degradation of 16.2ns, with the structural DRAM latency rising from 62.2ns to 78.6ns. For this comparison we’re using the very same DRAM sticks with identical timings between the Ryzen and two Threadripper platforms, so any differences here are solely due to the architectural differences of the platforms.

This degradation is actually to be expected. The third generation Threadripper degrades in two aspects compared to its predecessor: First of all, the chiplet architecture does incur a latency penalty as the separation of the CPU cores onto a different silicon die comes with a latency penalty. Secondly, in the first and second generation Threadripper products, each CPU had access to its own die memory controller by default, and it wasn’t possible to use an UMA setup. The third-gen Threadripper comes with an UMA setup by default, and the fact that the IOD is interleaving memory accesses across the quadrant memory controllers again adds another latency penalty.

Looking at the differences between the EPYC 7742 running in NPS4 mode and the new 3970X, we however see that the new TR3000 platform has a definitive latency advantage of almost 25ns – albeit we’re no longer running apples-to-apples here in regards to the DRAM.

Finally, the most interesting comparison is using the very same DRAM and timings between a Ryzen 3000 processor and the new 3970X. Using an 3700X we had at hand, the latency penalty for the new TR chip is “only” 9.2ns, rising from 69.4ns to 78.6ns. Maybe I might sound a bit optimistic here, based on the Rome numbers from earlier this summer I had expected some quite worse results for the new Threadripper 3000 series, so I see this result to be actually quite good. While we don’t have definitive confirmation, it does look like the new 24 and 32-core Threadripper 3000 SKUs are using only two adjacent quadrants of the sIOD.

Of course, the structural latency degradations here don’t necessarily translate to performance degradations. As we saw on the Ryzen 3000 products, AMD’s new doubled L3 cache as well as improved prefetchers have managed to more than compensate for the worse structural latency, actually increasing the memory performance of the new Zen2 chips.

Power Consumption: 6-13W Per Core Test Bed and Setup
Comments Locked


View All Comments

  • mkaibear - Monday, November 25, 2019 - link

    Intel had revenue of 19.2bn last quarter. The highest it's ever been for them.


    Claiming that Intel is destroyed is laughable.

    They're hurting at the moment, but then they were hurting in the Athlon era as well... and that didn't go so badly for them in the end.

    For reference, AMDs revenue for the same period was 1.8bn. yes, Intel, despite all their problems, earned *ten times* what AMD did.

    (Reference: https://www.anandtech.com/show/15045/amd-q3-fy-201...

    Claiming Intel are destroyed is just fanboyism at its worst.
  • Xyler94 - Monday, November 25, 2019 - link

    Unless Intel can get something out sooner rather than later, people are migrating to AMD because they are pushing things forward. 64 cores of Epyc fury is hitting them in the Server Space, which is where Intel is most scared of. They don't care that you or I buy an Intel chip or an AMD one, they care if Microsoft or Apple buys either or.

    Intel isn't destroyed, but they will be hurting for a while, as AMD is showing no signs of slowing down, and Intel has to beat what AMD makes next, not AMD today.
  • mkaibear - Monday, November 25, 2019 - link

    Again, Intel have record earnings this last quarter. As in over the last 3 months. As in after two years of AMD kicking their backside in the server space they're still making record amounts of revenue.

    Intel aren't stupid, they're one of the most ruthless companies in the sector. They can throw five times as much as AMD's *total profit* in R&D and still make five times as much profit as AMD does.
  • Xyler94 - Monday, November 25, 2019 - link

    Record breaking earnings mean nothing in the grand scheme of things.

    For as much as you gloat about Intel's RND, AMD is the one who's on top in 2 of the 4 markets (Laptops, Desktops, HEDT and Servers), some would argue 3. Doesn't matter how much money you can throw at a problem, it matters if you can solve it. AMD solved the problem, Intel hasn't, and it's a frantic state at Intel to make something happen, either get 10nm working better or changing their uArch in 14nm.

    Right now, the only reason to consider a XEON over an Epyc would be for AVX-512 only workloads. Because otherwise, ServerTheHome has shown that Epyc dominates, especially the 7742 64 core part.
  • SwackandSwalls - Monday, November 25, 2019 - link

    Those record breaking earnings (i.e. capital) mean a lot, and saying otherwise displays a large and intentional ignorance on how important capital is to the microprocessor industry. Intel can use that money to hire more both hardware and software talent, fund more research, build more fabs, outspend AMD in marketing, and on and on. If Intel had huge cash reserves but was putting up large losses every quarter then I'd be on board with your "grand scheme of things" comment. In reality they are massively profitable, selling more 14nm chips than they can produce, and have enough cash to not only learn from AMD's successes but also invest in following suit.
  • Xyler94 - Monday, November 25, 2019 - link

    Again, Hire all you want, throw as much money as you want. That doesn't matter if there's no results.

    AMD with literally tenths of Intel's funding can beat them, and have found better ways to make processors to increase core counts without sacrificing efficiency. Intel also needs to spend a lot of money on researching the node itself, AMD doesn't, so not all of Intel's RnD goes to making the CPU, lots of it goes into making the node itself.

    So while Intel may make more, they have to spend way more, especially since CPUs aren't the only thing Intel makes (They make flash chips, 3D XPoint, Networking chipsets, and many other products, all vying for that sweet RND cash)

    So while Intel makes more, they also spend more. Revenue is a great figure to look at on paper, but it doesn't amount to anything unless the spending is done wisely. AMD surely has shown that it doesn't take Intel levels of cash to become a market leader and capitalize on someone who's grown complacent.
  • milkywayer - Monday, November 25, 2019 - link

    "record breaking numbers mean a lot".

    So what happened then, why is AMDs offering more power and cost efficient at a much much lower price?
  • Korguz - Monday, November 25, 2019 - link

    mkaibear/SwackandSwalls, and point is ?? intel has all that money, yet.. been milking the SAME architecture for how many years ? as Xyler94 already said.. to keep throwing money at a problem, and it STILL doesnt get fixed, is NOT a good thing. AMD may not have the money that your beloved intel does, but guess what, they have been able to do MORE with what they do have, so tell me who is spending wiser ?? also.. how much of that 19.2 billion has intel had to dump into their fabs??
  • imaheadcase - Monday, November 25, 2019 - link

    I really hope you are not comparing Intell vs amd based on a just a CPU..that is illogical.
  • TEAMSWITCHER - Monday, November 25, 2019 - link

    Intel isn't "hurting" now... Desktop processors are not what most people want.

Log in

Don't have an account? Sign up now