Hi Misc, I was wondering if a kind soul could educate me about the results of my trivial iperf test (tcp protocol only).
I have two computers. My desktop is running OpenBSD 5.9 amd 64 stable while my daughters' MAC is running 10.7.5 32-bit Lion. Running iperf server on OpenBSD desktop (OpenBSD download test) I see on the MAC client speed of 739 Mbits/s. Running iperf server on the MAC (OpenBSD upload test) I get the speed of 545 Mbits/s. Both computers have 1 Gigbit Intel NICs. Firewalls were swithed off. I see similar discrepancy running the tests between DragonFly file server and MAC through the same ASUS 1Gigabit switch. According to iperf DragonFly can download things with the speed in excess of 935Mbits/s. While DragonFly is uploading things to MAC OS with the speed 750Mbits/s. I am really starting to be suspicious about iperf. Namely running speed tests between two OpenBSD machines one of which has Broadcom NIC I see almost symmetric speeds of about 565 Mbits/s which I would expect from a consumer grade 1Gigabit gear (same ASUS dumb switch used above). However iperf on the MAC consistently shows some stellar numbers. At work we run a similar test and on the decent Cisco gear MAC consistently shows tcp speeds over 930 Mbits/s. Is iperf lying? Predrag
