Supermicro C9Z390-CG

The Supermicro C9Z390-CG and the more expensive C9Z390-CGW seemingly have the same PCB layout which could indicate that both PCB boards are identical in terms of specification, but with this model having less in the way of controllers etc. Aesthetically the CG drops the use of a rear panel cover with metallic grey power delivery heatsinks and a black/grey chipset heatsink. The other key difference is that the C9Z390-CG completely drops RGB support with no integrated LEDs or headers onboard at all. On the storage side of things, the C9Z390-CG has two PCIe 3.0 x4 M.2 slots and also includes a total of six SATA ports.

The C9Z390-CG has a total of four RAM slots with support for DDR4-3866 and up to 64 GB in total. On the PCIe front, the board has two full-length PCIe 3.0 slots which operate at ether x16/x4 with the bottom PCIe 3.0 x4 slot populated at the bottom and/or x16/x8 and the bottom slot is disabled when the second full-slot is in use. Separate to the these is a total of three PCIe 3.0 x1 slots.

Compared to the more feature-rich and more expensive C9Z390-CGW (on the previous page), the C9Z390-CG has the same layout minus one of the dual LAN and instead opts for a single Intel I219V Gigabit LAN. The USB ports included on the rear panel consist of three USB 3.1 Gen2 Type-A, one USB 3.1 Gen2 Type-C and two USB 3.0 Type-A ports. The onboard audio which consists of five 3.5 mm audio jacks and an optical S/PDIF output is controlled by a Realtek ALC1220 HD audio codec and the C9Z390-CG also has a trio of video outputs with two DisplayPort 1.2 and a single HDMI 1.4 port.

The Supermicro C9Z390-CG pricing as it stands is currently unavailable. The target market is gamers on a budget and offers a more cost-effective and solid feature set without passing the cost of flashy RGB and unnessacary visually pleasing covers with the primary focus being on the performance.

Supermicro C9Z390-CGW Supermicro C9Z390-CG-IW
Comments Locked

79 Comments

View All Comments

  • Valantar - Wednesday, October 10, 2018 - link

    That would be pretty shocking, yeah, but the sheer size of that lump of metal still has me a bit worried. Guess that's what you get when you try to squeeze power delivery for a CPU that (likely) pulls >300W when overclocked into an ITX board (and refuse to use riser boards like before, for some reason).
  • FXi - Monday, October 8, 2018 - link

    The power feed also changed with z390 I believe at least in the Asus models it did. The power feed of the 370 was "enough" to drive the newer 9700/9900 but there is a difference there that may impact enthusiasts. I don't think it enough to warrant an upgrade but something to consider.
    Also people should remember that while it is still a bit of a ways off, wifi is going to change to Wifi6 or 802.11ax starting now and probably seeing much of the changeover during 2019/2020 depending on adoption choices. And there is also pci-e 4.0 to consider next year probably that should be thought about before people do "marginal" upgrades from 370 era chipsets.
  • FXi - Monday, October 8, 2018 - link

    Silly thing posted in edit window. Sorry power delivery and other points covered by you. Would have edited if I could have found that option
  • DanNeely - Monday, October 8, 2018 - link

    Other things to look forward to in the next few generations are: Less-hacky USB3.1 implementations (eg this articles speculation that a 10g port will need to eat 2 HSIO lanes instead of 1, and still needing an extra chip to support USB-C). Spectre/Meltdown fixes in hardware. A reduced DMI bottleneck between the CPU and chipset (either just from upgrading the link to PCIe4/5, moving some of the peripheral IO onto the CPU, or both.
  • Valantar - Wednesday, October 10, 2018 - link

    Considering that the maximum theoretical bandwidth of PCIe 3.0 x1 is 984.6MB/s, you _need_ two PCIe lanes (and thus two HSIO lanes) for a USB 3.1G2 (1.25GB/s) controller unless you want to significantly bottleneck it. That's not "hacky", that's reality, even if this leaves a lot of bandwidth "on the table" if this only powers a single port (which it rarely does, though, and given that a full load on two ports at one time is unlikely, running two 1.25GB/s ports off two .99GB/s lanes is a good solution).

    Moving DMI to PCIe 4.0 will be good, though, particularly for multiple NVMe SSDs and >GbE networking.
  • DanNeely - Wednesday, October 10, 2018 - link

    Splitting the traffic over 2 HSIO lanes is a hack because it'd require something to split/combine the traffic between the chipset and usbport. That in turn has me wondering if the speculation about the implementation being done that way is correct, or if the Z390 has 6 HSIO lanes that can run 10Gbps instead of the 8 that the rest top out at for PCIe3
  • repoman27 - Thursday, October 18, 2018 - link

    The implementation is absolutely not done that way. HSIO lanes are simply differential signaling pairs connected to a PCIe switch or various controllers via a mux. The PCH has a 6-port USB 3.1 Gen 2 xHCI, which can only feed 6 HSIO muxes. The back end of that xHCI is connected to an on-die PCIe switch which in turn is connected to the DMI interface. That DMI 3.0 x4 interface is already massively oversubscribed, but it is at least equivalent to a PCIe 3.0 x4 link, which is the most bandwidth that can be allotted to a single PCH connected device.
  • Srikzquest - Monday, October 8, 2018 - link

    HDMI 2.0 is available in Asus and Gigabyte's ITX boards as well.
  • gavbon - Tuesday, October 9, 2018 - link

    Thank you Srikzquest; updated the tables, obviously missed this yesterday :) - Thanks again
  • HickorySwitch - Monday, October 8, 2018 - link

    Correction:
    https://www.asus.com/us/Commercial-Servers-Worksta...
    It says under "Specifications" that the board sports HDMI 2.0[b?]

Log in

Don't have an account? Sign up now