[ovs-discuss] 40 Gbit Traffic Test Problem

[Gowrishankar Muthukrishnan]

On 10/25/18 4:35 PM, Mehmet Yaren wrote:
>
> Hi Ian,
>
>  
>
> We are using XL710 as below
>
>  
>
> argela at 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:
>
In your trex, are you sending multistream traffic (with L3/L4 header
variations) ?

Check if it helps throttling more packets into datapath by phy PMD (RSS).

Thanks.

-- 
Regards,
Gowrishankar M
gmuthukr at redhat.com

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.openvswitch.org/pipermail/ovs-discuss/attachments/20181026/6043d630/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: pEpkey.asc
Type: application/pgp-keys
Size: 1782 bytes
Desc: not available
URL: <http://mail.openvswitch.org/pipermail/ovs-discuss/attachments/20181026/6043d630/attachment.bin>