On 05/06/2024 08:32, zhou via USRP-users wrote:
Hi All,
I am using MCX516A-CCAT for X410 USRP. It has three network ports, two
for 100Gb QSFP and one for 1Gb ethernet. They are directly connected
to host. Surprisingly, I find much bigger latency on the 100Gb link
than the 1Gb link when ping them. I didn't notice this before.
Then I checked X310. Its latency is also pretty big compared to the
1Gb port: rtt min/avg/max/mdev = 0.341/0.539/0.793/0.187 ms
Why is the latency in 100Gb bigger than 1Gb port?
~$ uhd_find_devices
[INFO] [UHD] linux; GNU C++ version 11.4.0; Boost_107400; DPDK_21.11;
UHD_4.5.0.0-0-g471af98f
--------------------------------------------------
-- UHD Device 0
--------------------------------------------------
Device Address:
serial: 3289B23
addr: 192.168.20.2
claimed: False
fpga: CG_400
mgmt_addr: 192.168.10.2
mgmt_addr: 192.168.20.2
mgmt_addr: 192.168.6.66
name: ni-x4xx-3289B23
product: x410
type: x4xx
~$ ping 192.168.10.2
PING 192.168.10.2 (192.168.10.2) 56(84) bytes of data.
64 bytes from 192.168.10.2: icmp_seq=1 ttl=64 time=0.998 ms
64 bytes from 192.168.10.2: icmp_seq=2 ttl=64 time=0.888 ms
64 bytes from 192.168.10.2: icmp_seq=3 ttl=64 time=0.886 ms
64 bytes from 192.168.10.2: icmp_seq=4 ttl=64 time=0.894 ms
^C
--- 192.168.10.2 ping statistics ---
4 packets transmitted, 4 received, 0% packet loss, time 3036ms
rtt min/avg/max/mdev = 0.886/0.916/0.998/0.047 ms
~$ ping 192.168.6.66
PING 192.168.6.66 (192.168.6.66) 56(84) bytes of data.
64 bytes from 192.168.6.66: icmp_seq=1 ttl=64 time=0.180 ms
64 bytes from 192.168.6.66: icmp_seq=2 ttl=64 time=0.143 ms
64 bytes from 192.168.6.66: icmp_seq=3 ttl=64 time=0.115 ms
64 bytes from 192.168.6.66: icmp_seq=4 ttl=64 time=0.119 ms
^C
--- 192.168.6.66 ping statistics ---
4 packets transmitted, 4 received, 0% packet loss, time 3080ms
rtt min/avg/max/mdev = 0.115/0.139/0.180/0.025 ms
Thanks,
H.
Probably because the RJ45 (1G) port is more directly routed to the CPU
part of the RFSoC chip, since it is intended by
the RFSoC architecture for management of the embedded Linux
component. The 100G ports are, in this application,
intended largely for radio sample traffic, and management functions,
like ICMP, are more circuitously-routed through
the FPGA and into the Linux side of the house. Suffice it to say,
that sample traffic suffers no such long latency.
_______________________________________________
USRP-users mailing list -- usrp-users@lists.ettus.com
To unsubscribe send an email to usrp-users-le...@lists.ettus.com