On 18 July 2014 10:53, Joe Stringer <joestrin...@nicira.com> wrote: > Basically I'd be much more comfortable with the kernel generating its >> own flow key the first time it dumps a flow, then caching that. How >> much overhead does that have versus caching the userspace-provided >> flow key? >> > > By the same metric, roughly 15% increase in flows supported. This is > locking to write the cache back to the flow, but zero contention on that > lock (I used an independent lock and a single revalidator thread for the > test). >
I think I may have made this ambiguous. I mean, this Patch v3 has approx. 33% improvement against master. Constructing during dump has approximately 15% improvement against master. _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
