Update from iperf3 version 3.11 to 3.13

[HughNhan]

To overcome iperf_udp: implement UDP connect retry mechanism #1144

[k-rister]

What is the status of this PR?

[HughNhan]

At the time, this update meant to pick up PR esnet/iperf#1402. It has been a while. I think we need to verify the latest and fast forward if it is good.

[k-rister]

What do we need to do in order to verify the latest?

[HughNhan]

The problem we had with iperf3 was UDP under congestion. iperf3 uses UDP for both control traffic and data traffic. So when the congestion happens, the conrol traffic got lost and all hell broke loose.
I am not sure if latest has fixed this specific problem or not, but to ensure no regression we should run multibench and compare UDP stream throughputs specifically bench id 21-26. If there are no new failures and/or lower throughput, we can fast forward.

[k-rister]

This has been stale for so long that I'm going to close it. If we want to revisit this then we need to rebase, reopen, and potentially do so with a more up to date version of iperf since we let this PR languish for so long.