Hi Eli, Vlad
Do you have any performance data that you can share? specifically,
assuming I use connectx such that I have already checksum and LSO
offloads working for the datagram mode and 2K MTU.
What would be the advantage added by LRO? would it be only to the CPU
utilization or there's also better BW?
Or.
_______________________________________________
general mailing list
[email protected]
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general
To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
