Hello,
I am switching routers: An R7800 on OpenWrt 23.05.4 to an XR500 on OpenWrt 24.10.1.
Naturally I wanted to test some WLAN speeds, so I installed iperf3. Not long after I wondered how to run it permanently on the router so I added this in System/Startup
iperf3 -s &
Later I saw that I used a different line on the old one:
iperf3 -s -D
Does it actually make a difference how I start the server? Or would you advise something different?
I would not run iperf on the router at all, iperf needs a lot of CPU cycles so will skew the result.
But if you really want then running it with & will put it in the background of the terminal and running it with -D will run it as daemon which means it is detached from the terminal.
So I would run it -D daemonized but in practice there will not be much difference, but as said I do not use it at all on the router, best throughput measurement is between two connected PC's 
Thanks! Hmm, top showed only zeroes... Luci's Load was like this:
...and only after starting it, before it was only yawning as well 
Well, having it as daemon on the router saves the effort of starting it manually. Yes, I am that lazy 
But, yes, will test later between endpoints as well (just now I realized needing iperf on android, too).
PS: Basically I will just use for testing wifi (and funny thing here: Partially occupied CH3 was faster than the empty CH13)
I use "magic iperf" for android not sure if that is still available
Let me know the difference between 23.05 and 24.10 (CPU's are the same for R7800 and XR500, I think)
Apart from the flash size (128 MB vs 256 MB and the LED GPIOs, both sharing 99.8% of the *.dtsi), there's zero difference between the r7800 and xr500 on the hardware side, same SOC (IPQ8065), same wireless (QCA9984), same RAM size. There shouldn't (can't) be any performance difference, if your testing suggests otherwise, I would assume a mistake in the benchmark setup.
If you're looking for a speedup, you'll need to look at different -more modern, wifi 6- hardware (SOCs).
iperf/ iperf3 are also part of https://play.google.com/store/apps/details?id=net.he.networktools, but it can also be installed via https://play.google.com/store/apps/details?id=com.termux.
You are right I even have net.he.networktools installed 
Yeah, still funny: the mobile also showed 5MB/s on old v23 and CH3 and 9.x on new v24 on CH13 (same room, 2m distance). Perhaps I should switch channels and test again...
Exactly, I got the new one so cheap so I was keen to try: Same hardware, just double flash size. Seemed like a nobrainer, what could go wrong?
Right, but "mistake" is such a harsh word, I'd rather use "difference". After all I am aware that I have 2 different versions, using different channels.
Well, I was sort of... Don't we all yearn for some speedup? I was considering it but then there was hardly anything suitable:
Would like two other features: speedier LAN and detachable antennas... Alas, perhaps next year.
Thanks, didnt know those he.net tools (weird preset though).
Curious, even changing channels does not improve the lower performance of the old v23:
v23 ch3
[ ID] Interval Transfer Bitrate Retr
[ 5] 0.00-10.03 sec 39.8 MBytes 33.3 Mbits/sec 636 sender
[ 5] 0.00-10.02 sec 39.2 MBytes 32.8 Mbits/sec receiver
v23 ch13
[ ID] Interval Transfer Bitrate Retr
[ 5] 0.00-10.12 sec 51.9 MBytes 43.0 Mbits/sec 0 sender
[ 5] 0.00-10.02 sec 49.4 MBytes 41.3 Mbits/sec receiver
v24 ch3
[ ID] Interval Transfer Bitrate Retr
[ 5] 0.00-10.01 sec 61.2 MBytes 51.3 Mbits/sec 77 sender
[ 5] 0.00-10.01 sec 59.1 MBytes 49.6 Mbits/sec receiver
v24 ch13
[ ID] Interval Transfer Bitrate Retr
[ 5] 0.00-10.02 sec 83.9 MBytes 70.2 Mbits/sec 0 sender
[ 5] 0.00-10.02 sec 80.6 MBytes 67.5 Mbits/sec receiver
edit: Also switched the lines in startup file, v24 still faster (as expected). I guess, after changing the "old" r7800 to v24, I will check the R7800 again (as for now, I ran out of other ideas).