I have the following result from a UDP throughput test as measured by iperf3
$ iperf3 -u -t 10 -c 192.168.1.2 -b 100M -V
iperf 3.6+
Linux pi-raspberry1 4.4.38-v7+ #938 SMP Thu Dec 15 15:22:21 GMT 2016 armv7l
Control connection MSS 1448
Setting UDP block size to 1448
Time: Thu, 04 Oct 2018 09:58:29 GMT
Connecting to host 192.168.1.2, port 5201
Cookie: h3wj4by4mtyobgtq42gsf62s3nfhymg6djry
[ 5] local 192.168.1.1 port 40988 connected to 192.168.1.2 port 5201
Starting Test: protocol: UDP, 1 streams, 1448 byte blocks, omitting 0 seconds, 10 second test, tos 0
[ ID] Interval Transfer Bitrate Total Datagrams
[ 5] 0.00-1.00 sec 3.34 MBytes 28.0 Mbits/sec 2421
[ 5] 1.00-2.00 sec 3.13 MBytes 26.2 Mbits/sec 2264
[ 5] 2.00-3.00 sec 2.15 MBytes 18.1 Mbits/sec 1560
[ 5] 3.00-4.00 sec 905 KBytes 7.41 Mbits/sec 640
[ 5] 4.00-5.00 sec 486 KBytes 3.98 Mbits/sec 344
[ 5] 5.00-6.00 sec 2.19 MBytes 18.4 Mbits/sec 1587
[ 5] 6.00-7.00 sec 2.55 MBytes 21.4 Mbits/sec 1848
[ 5] 7.00-8.00 sec 2.18 MBytes 18.3 Mbits/sec 1576
[ 5] 8.00-9.00 sec 2.72 MBytes 22.8 Mbits/sec 1968
[ 5] 9.00-10.00 sec 3.21 MBytes 26.9 Mbits/sec 2321
- - - - - - - - - - - - - - - - - - - - - - - - -
Test Complete. Summary Results:
[ ID] Interval Transfer Bitrate Jitter Lost/Total Datagrams
[ 5] 0.00-10.00 sec 22.8 MBytes 19.1 Mbits/sec 0.000 ms 0/16529 (0%) sender
[ 5] 0.00-10.77 sec 22.8 MBytes 17.8 Mbits/sec 22.003 ms 4/16529 (0.024%) receiver
CPU Utilization: local/sender 9.8% (1.5%u/8.2%s), remote/receiver 1.5% (0.0%u/1.5%s)
If UDP constantly sends data without caring if they arrived or not, why is it that the amount of MBytes sent differs each second as well as the amount of total Datagrams sent does?
Moreover I see that the same number of MBytes is seen by both the server and the client at the Summary Results, but the Bitrate is not the same, does this have to do with the bigger interval taken into account by the server (10.77 seconds instead of 10)?
Best Answer
Here's a iperf3 client RaspberryPi3B running current Raspbian to a iperf3 server MacbookPro8,2 running current Fedora 29 x86_64, across a cheap DLink DGS-1005A gigabit ethernet switch, same parameters as the pi-raspberry1 test in the question:
What's the provenance of the pi-raspberry1's software? Compare with the stock current Raspbian: pi-raspberry1's kernel is much older; pi-raspberry1's iperf3 version is newer. If you compiled the newer iperf3 did you do it via the Debian packaging system to make sure you compiled with the same compiler options and patches?
Note the difference in CPU: pi-raspberry1: 9.8% (1.5%u/8.2%s); vk5tu 19.3% (0.5%u/18.8%s). Did you check CPU usage during the test to ensure that the transmission was not CPU-bound? Also the USB bus on the RPi is shared between the ethernet, WLAN and USB ports, so that USB bus needs to have no other traffic whilst running the iperf tests. That pi-raspberry1 used three times as much system CPU than vk5tu whilst doing only 20% of the throughput suggests that either pi-raspberry1 is not recent RPi hardware or that there have been substantial improvements in the kernel from 4.4.38 (Dec 2016) to 4.14.79 (Nov 2018).
Have you repeated the test with a stock current system?
As you can see from the above, it is important to give a lot of information about the test environment when asking about performance issues. Best wishes.
Edit:
For your assistance, here's the same test with no output pacing, to stress the RPi3B: