More I/O For AM5: PCIe 5, Additional PCIe Lanes, & More Displays

AMD’s other big expenditure using socket AM5’s additional pins is on I/O support. While AM4 already supported a fair bit of I/O, including 24 PCIe lanes, 3 displays, and 4 Superspeed USB ports, there was still room for improvement. So for AM5, AMD has increased the amount of I/O and the flexibility offered with the platform.

The biggest change here is that the AM5 socket now provides for 28 lanes of PCIe, a net gain of 4 lanes. More significantly still, PCIe 5.0 is now supported (at least on the Ryzen 7000 “Raphael” processors), doubling the bandwidth of all of those PCIe lanes to a max of 4GB/sec/lane. Which gives the chip a maximum cumulative PCIe bandwidth of 112GB/sec in each direction.

In practice, those additional lanes are intended for NVMe drives, giving AM5 a second x4 connection to drive a second NVMe drive; though we have seen some motherboard designs where vendors are stealing the second x4 for a PCIe 5.0 x4 slot. Past that, things look a lot like AM4, with 16 PCIe lanes to directly drive one or more PCIe slots, and then 4 lanes for hooking up the chipset.

Meanwhile, the updated socket also offers enough pins for the CPU to drive 4 Superspeed USB 3.x ports, and a USB 2 port. The USB 2 port is new for this generation, and meanwhile 3 of those USB 3 ports now also support the USB Type-C connector, unlike AM4 which could only natively drive Type-A ports. As a result, AM5 CPUs can drive a total of 3 Superspeed Type-C ports, a fourth Superspeed Type-A port, and then the aforementioned USB 2 port.

There has been one regression, however, and that is SATA support. Whereas AM4 CPUs could drive a mix of NVMe and SATA drives (up to 2 SATA + a PCIe x2 for NVMe), AM5 is purely PCIe. So there is no native SATA support on the CPU, and supplying that will come from the chipset.

To visualize this, we’ll use part of the AM5 chipset diagram. We’ll go more into the specifics of the chipsets in a bit, but lays out what is wired to the CPU, and what will need to be wired to the chipset. Of note there, the current chipsets only use PCIe 4.0 connectivity to the Ryzen CPU, so the current generation of chipsets will not be making full use of the bandwidth capabilities of the CPU itself.

Which with the addition of PCIe 5.0 support to the platform, is going to be a recurring theme. While AMD has baked in 5.0 support into the Raphael CPUs, it’s up to motherboard vendors to actually make it so. Compared to PCIe 4.0, 5.0 has much tighter signal integrity requirements (the signaling frequency has been doubled), which at least at this time, makes PCIe 5.0 expensive to implement. A very well-designed motherboard is required with impeccable traces, and on top of that the overall short throw of PCIe 5.0 means that retimers/redrivers become necessary rather quickly. So while AM5 can support PCIe 5.0 throughout, the reality is that we’re still going to see a lot of PCIe 4.0 in use even in higher-end motherboards.

As for the necessity of PCIe 5.0 overall, thus far AMD is primarily focused on what it means for NVMe drive speeds. The first generation of PCIe 5.0-enabled consumer SSDs are expected to land a bit later this year, and they should be able to hit sequential burst transfer rates above the limits of PCIe 4.0 (~7GB/sec).

Past that, NVIDIA’s newly announced Ada Lovelace architecture GeForce RTX 40 series video cards do not support PCIe 5.0. So while we’re awaiting AMD to announce their RDNA 3-based product lineup later this year, regardless of what AMD does, the bulk of video cards sold next year are not going to use PCIe 5.0. So there is a bit less pressure on motherboard manufactures (and motherboard buyers) to get boards that support PCIe 5.0 to anything beyond a couple of M.2 slots.

Finally, in conjunction with the USB I/O changes, AM5 also introduces some display I/O changes. Whereas AM4 could directly drive up to 3 displays, AM5 brings this to 4. Specifically, AM5 offers one dedicated display output (which will generally be allocated to HDMI), while the other 3 display outputs are available over those 3 USB Type-C ports as DisplayPort alt mode. It’ll be up to motherboard manufacturers if they want to expose any of these USB-C root ports as physical USB-C ports or as DisplayPorts, but so far from the motherboard designs we’ve seen, the former is more common (though certainly not universal).

Anticipating a shift to more USB Type-C displays, AMD is also implementing what they call “hybrid graphics” support on AM5. Unlike previous products where this referred to linking up the integrated graphics with a discrete GPU in CrossFire mode, this time around it refers to being able to being able to use the mobo/iGPU’s display outputs to drive a monitor while using a dGPU to render content. This is largely lifted from AMD’s laptop technologies, where similar techniques are used to allow the dGPU to be powered down when it’s not in use. In the case of desktop processors, this just means every display output will work, regardless of whether it’s plugged into ports coming from the CPU or a discrete video card.

It’s also worth noting that AM5 is bringing a few other, more minor updates to other comms protocols. Among these is support for MIPI’s (relatively) new I3C chip-to-chip signaling standard, which will ultimately be supplanting the long-used I2C standard. As well, AM5 doubles the number of I2C/I3C ports available, bringing the total to 4 ports. The platform also adds a second (enhanced) Serial Peripheral Interface (eSPI/SPI) port, and on the audio front, adds support for the Digital Mic and MIPI’s Soundwire standard for low-cost audio peripherals.

Socket AM5: The New Platform for Consumer AMD AM5 Chipsets: X670 and B650, Built by ASMedia
POST A COMMENT

205 Comments

View All Comments

  • Tomatotech - Friday, September 30, 2022 - link

    Nice idea but you’re swimming against the flow of history. The trend is always to more tightly integrate various components into smaller and smaller packages. Apple have moved to onboard RAM in the same package as the CPU which has bought significant bandwidth advantages and seems to have boosted iGPU to the level of low-end dGPUs.

    The main takeaway from your metaphor of the 650w dGPU with a 55w mainboard and 100-200w CPU is that high-end dGPUs are now effectively separate computers in their own right - especially as a decent one can be well over 50% of the cost of the whole PC - and are being constrained by having to fit into the PC in terms of physical space, power supply capacity, and cooling capacity.

    It’s a shrinking market on both the low end and high end for home use of dGPU, given these innovations and constraints and I don’t know where it’s going to go from here.

    Since I got optic fibre, I’ve started renting cloud based high-end dGPU and it has been amazing albeit the software interface has been frustrating at times. With symmetric gigabit service and 1-3ms ping, it’s like having it under my desk. I worked out that for unlimited hours and given the cost of electricity, it would take 10 years for my cloud rental costs to match the cost of buying and running a home high end dGPU.

    Not everyone has optic fibre of course but globally it’s rolling out year by year so the trend is clear again.
    Reply
  • Castillan - Wednesday, September 28, 2022 - link

    "

    clang version 10.0.0
    clang version 7.0.1 (ssh://git@github.com/flang-compiler/flang-driver.git
    24bd54da5c41af04838bbe7b68f830840d47fc03)

    -Ofast -fomit-frame-pointer
    -march=x86-64
    -mtune=core-avx2
    -mfma -mavx -mavx2
    "

    ...and then later the article says:

    "The performance increase can be explained by a number of variables, including the switch from DDR4 to DDR5 memory, a large increase in clock speed, as well as the inclusion of the AVX-512 instruction set, albeit using two 256-bit pumps."

    The problem here being that those arguments to Clang will NOT enable AVX-512. Only AVX2 will be enabled. I verified this on an AVX512 system.

    To enable AVX512, at least at the most basic level, you'll want to use "-mavx512f ". There's also a whole stack of other AVX512 capabilities, which are enabled with "-mavx512dq -mavx512bw -mavx512vbmi -mavx512vbmi2 -mavx512vl" but some may not be supported. It won't hurt to include those on the command line though, until you try to compile something that makes use of those specific features, and then you'll see a failure if the platform doesn't support those extensions.
    Reply
  • Ryan Smith - Friday, September 30, 2022 - link

    Correct. AVX-512 is not in play here. That is an error in analysis on our part. Thanks! Reply
  • pman6 - Thursday, September 29, 2022 - link

    intel supports 8k60 AV1 decode.

    Does ryzen 7000 support 8k60 ??
    Reply
  • GeoffreyA - Monday, October 3, 2022 - link

    The Radeon Technology Group is getting 16K ready. Reply
  • yhselp - Thursday, September 29, 2022 - link

    I'd love to see you investigate memory scaling on the Zen 4 core. Reply
  • Myrandex - Thursday, September 29, 2022 - link

    The table on page four mentions "Quad Channel (128-bit bus)" for memory support. Does that mean we could have a 4 memory slot solution, with one memory module per channel, with four channel support? This way to drastically increase memory bandwidth all while maintaining those fast DDR5 frequencies? Reply
  • Ryan Smith - Friday, September 30, 2022 - link

    No. That configuration would be no different than a 2 DIMM setup in terms of bandwidth or capacity. Slotted memory is all configured DIMMs; as in Dual Inline Memory Module. Reply
  • GeoffreyA - Friday, September 30, 2022 - link

    All in all, excellent work, AMD, on the 7950X. Undoubtedly shocking performance. Even that dubious AVX-512 benchmark where Intel used to win, Zen 4 has taken command of it. However, lower your prices, AMD, and don't be so greedy. Little by little, you are becoming Intel. Don't be evil.

    Thanks, Ryan and Gavin, for the review and all the hard work. Much appreciated. Have a great week.
    Reply
  • Footman36 - Friday, September 30, 2022 - link

    Yawn. I really don't see what the big fuss is about. I currently run 5600X and was interested to see how the 7600X compared and while it does look like a true uplift in performance over the 5600X, I would have to factor in cost of new motherboard and DDR5 ram! On top of that, the comparison is not exactly apples to apples in the testing. 7600X has a turbo speed of 5.3, 5600X 4.6. 7600X runs with 5200 DDR5 and 5600X 3200 DDR4, 7600X has TDP 105W, 5600X 65W. If you take a look at the final page where the 7950X is tested in ECO mode which effectively supplies 65W instead of 105W you lose 18% performance. If we try to do apples to apples and use eco mode with 7600X, to get apples to apples with 65W of 5600W, then lower boost to 4.6ghz then the performance of the 2 cpu's looks very similar. Perhaps not the way I should be analyzing the results, but just my observation.... Reply

Log in

Don't have an account? Sign up now