On Mon, Sep 08, 2014 at 12:51:11PM -0700, Ethan Jackson wrote: > > Is there any point to testing OVS in a configuration that we won't > > distribute? > > This is a subject in which there could legitimately be many points of > view, but I'll take a stab at explaining how I think about it. > > To answer your question: In general no, but with DPDK yes. > > If you look at how stock OVS is typically deployed, a vendor (Redhat, > Debian, etc), builds it in a portable fashion and distributes binaries > to whoever wants them. In this case, I completely agree that these > sorts of compiler optimizations don't make sense. > > With DPDK I think the situation is slightly different. The DPDK > libraries are specifically designed for people to heavily tune the > applications built on top of them, and I suspect the people interested > in OVS on DPDK at this point are the same. To the extent that OVS on > DPDK is deployed, it will be in very specific circumstances where > performance is critical and these sorts of compile time optimizations > are typical. > > Anyways that's my thinking on it. If you prefer, I could make it an > option, but for the dev script I don't think it really matters that > much. Keep in mind that real numbers we publish are based on official > builds which don't use this script.
It's a philosophical question really, and you gave me the philosophical answer that I wanted. Thanks, that makes sense. At any rate I don't use this script and this script is your baby, so don't let me block anything in any case. _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
