Hi, Ken, I don’t have a ZCU216 either, but the specs say it has four SFP28 ports. I don’t know whether the CASPER 100 GbE core can gang those four ports into a single 100 GbE interface, but others on this list should be able to confirm that one way or the other. At the other end of the connection, the 100 GbE NIC likely has a QSFP28 port and likely supports 100GBASE-CR4. One can buy a breakout cable with a single QSFP connector on one end and four SFP28 connectors on the other end, but AFAIK there are no NICs which will present the four lanes of a “CR4” (or “SR4" etc) port as four separate interfaces. To my knowledge, the only devices that do that are switches.
I think you’ll need a small switch (check eBay) that will connect from one port to the ZCU216 via one of those breakout cables and from another port to the NIC with a QSFP28-to-QSFP28 cable. The switch port that connects to the ZCU216 will run in either 100 GbE mode (if the CASPER 100 GbE core can split the 100 GbE interface over the four SQF28 ports) or 4x25 GbE mode. The switch port that connects to the NIC will run in 100 GbE mode (as will the NIC itself). It would be good to hear from folks with more direct experience with the ZCU216 before buying anything, but I think it’s safe to say that the network config your vendor showed you is for a 25 GbE connection. Their switch port is “25G” and their NIC is “25000baseCR” (aka 25GBASE-CR). Hope this helps, Dave > On Jul 30, 2023, at 15:43, Ken Semanov <shapkiqua...@gmail.com> wrote: > > The ZCU216 has a 4 x 25.7 cage for a CAUI-4. In order to receive jumbo > frames at a receiving NIC that only has 100GbE ports, what kind of hardware > cable work would be appropriate? > > I have attempted to work closely with a hardware manufacturer, but I feel > they are unreliable in their "advice". They sent detailed results of a > connection test of their own cable, allegedly proving that their hardware is > appropriate for my use case. But they clearly implemented a single 25GBASE-CR > connection over a single port of a network switch, which then addressed a > single device on the receiving side (enp1s0f0) . Their receiving NIC was set > to 4x25 port configuration, making it even less relevant. I post this > question here as a last resort, not a first resort for this ongoing problem. > > Lacking a ZCU216 and its FPGA, the tech support specialists used a network > switch as a substitute. Here is the list of the ports on their switch from > their results. (truncated for space) > > ``` > FS#show interfaces status > Interface Status Vlan Duplex Speed > Type > > ---------------------------------------- -------- ---- ------- --------- > ------ > > TenGigabitEthernet 0/1 down 1 Unknown Unknown > fiber > > . > > . > > . > > > TenGigabitEthernet 0/30 down 1 Unknown Unknown > fiber > > TFGigabitEthernet 0/21 up 1 Full 25G > fiber > > TFGigabitEthernet 0/22 down 1 Unknown Unknown > fiber > > . > . > . > ``` > > Ethtool for their single network device shows 25GBASE-CR. > ``` > Settings for enp1s0f0: > > Supported ports: [ FIBRE ] > > Supported pause frame use: Symmetric > > Supports auto-negotiation: Yes > > Supported FEC modes: None RS BASER > > Advertised link modes: 10000baseT/Full > > 25000baseCR/Full > > Advertised pause frame use: No > > Advertised auto-negotiation: Yes > > Advertised FEC modes: None RS BASER > > Speed: 25000Mb/s > > Duplex: Full > > ``` > > But 25BASE-CR is not CAUI-4. I conclude this test is therefore useless for > this network configuration. Your thoughts? > > > > > -- > You received this message because you are subscribed to the Google Groups > "casper@lists.berkeley.edu" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to casper+unsubscr...@lists.berkeley.edu > <mailto:casper+unsubscr...@lists.berkeley.edu>. > To view this discussion on the web visit > https://groups.google.com/a/lists.berkeley.edu/d/msgid/casper/f8097049-e5c6-4a30-89b1-832aed2dda5an%40lists.berkeley.edu > > <https://groups.google.com/a/lists.berkeley.edu/d/msgid/casper/f8097049-e5c6-4a30-89b1-832aed2dda5an%40lists.berkeley.edu?utm_medium=email&utm_source=footer>. -- You received this message because you are subscribed to the Google Groups "casper@lists.berkeley.edu" group. To unsubscribe from this group and stop receiving emails from it, send an email to casper+unsubscr...@lists.berkeley.edu. To view this discussion on the web visit https://groups.google.com/a/lists.berkeley.edu/d/msgid/casper/E3AF5924-C746-464B-8748-51C29D212A50%40berkeley.edu.
