On Tue, May 20, 2008 at 6:12 AM, Tziporet Koren <[EMAIL PROTECTED]> wrote: > Chris Worley wrote: >> >> In using netcat in UDP mode over IPoIB, I loose 25%-40% of the >> packets. Is that expected? >> > > What is the netcap test?
Start a listener on one node (both nodes running RHEL5.1/OFED1.3): [EMAIL PROTECTED] ~]# nc -v -v -u -l 61984 | dd of=/dev/null bs=1024k Start a sender on another, which completes: [EMAIL PROTECTED] mnt]# dd if=/dev/zero bs=1024k count=10000 | nc -v -v -u 36.102.28.91 61984 10000+0 records in 10000+0 records out 10485760000 bytes (10 GB) copied, 48.0578 seconds, 218 MB/s But the sender just hangs until ^C: Connection from 36.102.28.94 port 61984 [udp/*] accepted 0+6078248 records in 0+6078248 records out 6239293440 bytes (6.2 GB) copied, 179.339 seconds, 34.8 MB/s Chris > > Tziporet > _______________________________________________ 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
