Performance testing with iperf3...

[SnakeByte]

Now that I can compile binaries for my RT-AC68U, I've been running some network performance tests.

Here's a test running iperf3 (much better than iperf2!) on the device itself testing through the loopback interface:

Test Complete. Summary Results:
[ ID] Interval Transfer Bandwidth Retr
[ 4] 0.00-10.00 sec 1.59 GBytes 1.37 Gbits/sec 128 sender
[ 4] 0.00-10.00 sec 1.59 GBytes 1.37 Gbits/sec receiver
CPU Utilization: local/sender 76.1% (1.3%u/74.9%s), remote/receiver 34.9% (1.2%u/33.8%s)

And here's the results, testing from a Ubuntu 12.0.4 LTS Hyper-V guest (running on Windows 8) to the router:

Test Complete. Summary Results:
[ ID] Interval Transfer Bandwidth Retr
[ 4] 0.00-10.00 sec 617 MBytes 518 Mbits/sec 0 sender
[ 4] 0.00-10.00 sec 616 MBytes 517 Mbits/sec receiver
CPU Utilization: local/sender 4.3% (0.1%u/4.2%s), remote/receiver 18.8% (0.2%u/18.6%s)

Strange that the BW has effectively been cut in half. I'm going to try with a VMWare Workstation 10.0.1 version of that ubuntu guest and see if there is any performance change. I'm also going to test from a mac.

I did some testing through an openvpn tunnel as well and was happy to see that my internet bandwidth was fully utilized. (20Mbit/2Mbit)