Hi Ian,
We are using XL710 as below
argela@dsfc-ovs:~$ lspci
...
08:00.0 Ethernet controller: Intel Corporation Ethernet Controller XL710
for 40GbE QSFP+ (rev 02)
08:00.1 Ethernet controller: Intel Corporation Ethernet Controller XL710
for 40GbE QSFP+ (rev 02) ...
Our test setup is simple phy to phy test. We are using Trex to send traffic
into our DPDK based OvS with XL710. It is a basic loopback setup. We can
see ~10Gbps at most on OvS even though our CPUs usage are below than %25.
OvS drops remaining traffic after ~10Gb. In addition, even if we use just
one CPU in OvS, we see the same traffic ~10Gb. Why does more CPU count not
increase our traffic handling? Our configuration is as follows:
argela@dsfc-ovs:~$ sudo ovs-ofctl dump-flows br0 cookie=0x0,
duration=135.441s, table=0, n_packets=38181256, n_bytes=6251887265,
priority=50,ip,in_port=dpdk0,nw_dst=10.0.0.0/16
actions=mod_dl_dst:3c:fd:fe:a8:23:e0,output:dpdk1
cookie=0x0, duration=128.913s, table=0, n_packets=14099677,
n_bytes=8482650942, priority=50,ip,in_port=dpdk1,nw_dst=10.0.0.0/16
actions=mod_dl_dst:3c:fd:fe:a8:23:e1,output:dpdk0
argela@dsfc-ovs:~$ sudo ovs-vsctl show
0ce5f427-33c6-4d79-a67c-b2a1c588c3e9
Bridge "br0"
Port "dpdk0"
Interface "dpdk0"
type: dpdk
options: {dpdk-devargs="0000:08:00.0", n_rxq="4"}
Port "br0"
Interface "br0"
type: internal
Port "dpdk1"
Interface "dpdk1"
type: dpdk
options: {dpdk-devargs="0000:08:00.1", n_rxq="4"}
argela@dsfc-ovs:~$ sudo ovs-appctl dpif-netdev/pmd-rxq-show pmd thread
numa_id 0 core_id 0:
isolated : true
port: dpdk0 queue-id: 0 pmd usage: 25 %
pmd thread numa_id 1 core_id 1:
isolated : true
port: dpdk0 queue-id: 1 pmd usage: 25 %
pmd thread numa_id 0 core_id 2:
isolated : true
port: dpdk0 queue-id: 2 pmd usage: 23 %
pmd thread numa_id 1 core_id 3:
isolated : true
port: dpdk0 queue-id: 3 pmd usage: 26 %
pmd thread numa_id 0 core_id 4:
isolated : true
port: dpdk1 queue-id: 0 pmd usage: 9 %
pmd thread numa_id 1 core_id 5:
isolated : true
port: dpdk1 queue-id: 1 pmd usage: 10 %
pmd thread numa_id 0 core_id 6:
isolated : true
port: dpdk1 queue-id: 2 pmd usage: 9 %
pmd thread numa_id 1 core_id 7:
isolated : true
port: dpdk1 queue-id: 3 pmd usage: 10 %
Regards,
Mehmet.
Ian Stokes <ian.sto...@intel.com>, 24 Eki 2018 Çar, 17:23 tarihinde şunu
yazdı:
> On 10/23/2018 1:02 PM, mehmetyaren wrote:
> > Hi,
> >
> > We want to send 40 Gbit traffic with t-rex traffic generator, we are
> > using 40 Gbit supported DPDK NIC, but when we generate traffic as 40
> > Gbit, it can not transmit all traffics just 10 Gbit traffic is
> > transmitted by DPDK NIC. We are using Open vSwitch 2.10.0 version and
> > DPDK 17.11.2 version.
>
> What type of NICs are being used in your setup?
>
> >
> > We made some configuration to increasing core number for dpdk ports with
> > below commands;
> >
> > ovs-vsctl set interface dpdk0 options:n_rxq=4
> > other_config:pmd-rxq-affinity="0:0,1:1,2:2,3:3"
> >
> > ovs-vsctl set interface dpdk1 options:n_rxq=4
> > other_config:pmd-rxq-affinity="0:4,1:5,2:6,3:7"
> >
> > And we have seen all pmd capacities are not used by dpdk ports as seen
> > below results;
> >
> > sudo ovs-appctl dpif-netdev/pmd-rxq-show pmd thread numa_id 0 core_id 0:
> >
> > isolated : true
> >
> > port: dpdk0 queue-id: 0 pmd usage: 19 %
> >
> > pmd thread numa_id 1 core_id 1:
> >
> > isolated : true
> >
> > port: dpdk0 queue-id: 1 pmd usage: 20 %
> >
> > pmd thread numa_id 0 core_id 2:
> >
> > isolated : true
> >
> > port: dpdk0 queue-id: 2 pmd usage: 20 %
> >
> > pmd thread numa_id 1 core_id 3:
> >
> > isolated : true
> >
> > port: dpdk0 queue-id: 3 pmd usage: 21 %
> >
> > pmd thread numa_id 0 core_id 4:
> >
> > isolated : true
> >
> > port: dpdk1 queue-id: 0 pmd usage: 13 %
> >
> > pmd thread numa_id 1 core_id 5:
> >
> > isolated : true
> >
> > port: dpdk1 queue-id: 1 pmd usage: 13 %
> >
> > pmd thread numa_id 0 core_id 6:
> >
> > isolated : true
> >
> > port: dpdk1 queue-id: 2 pmd usage: 13 %
> >
> > pmd thread numa_id 1 core_id 7:
> >
> > isolated : true
> >
> > port: dpdk1 queue-id: 3 pmd usage: 15 %
> >
>
> Can you provide more information regarding the test setup? I assume it's
> simple hair pin tests (phy to phy).
>
> What flow rules are you using?
>
> Regards
> Ian
>
> > Can anyone help me ?
> >
> > Mehmet.
> >
> > Sent from Mail <https://go.microsoft.com/fwlink/?LinkId=550986> for
> > Windows 10
> >
> >
> >
> > _______________________________________________
> > discuss mailing list
> > disc...@openvswitch.org
> > https://mail.openvswitch.org/mailman/listinfo/ovs-discuss
> >
>
>
_______________________________________________
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss
