On 05/23/14 at 01:15am, Thomas Graf wrote: > Test: * Avg pps as reported by pktgen for a single pktgen thread > * qemu, vcpu=2, mem=2GB, megaflows disabled > > Orig [0] ZC [1] mmap [2] shared-ofpbuf [3] > 64B (linear) 624K 564K 962K 900K > 1.5K (linear) 612K 629K 800K 891K > 9K (linear) 476K 474K 508K 513K > 9K (4 frags) 365K 411K 503K 631K > [...] > > Numbers on ring buffer size: > ring-size: 512K 1M 2M 4M > pps: 596K 914K 926K 900K
Another option is to reduce the max frame size from 16K to 4K or even 2K (translating to a ring size of 256KB respectively 128KB) and handle messages >4K through a regular recvmsg() again. Tests indicate that the performance for these exceeding messages is on the same level as they are now. I would personally consider this option if we can't come up with a solution to the fairness issue using shared rings. _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
