Performance and Power Demand RockPro64

[dukla2000]

I would be running it as a webserver at home and was hoping maybe someone was already doing that. If you were willing to do some measurements for me, I would be interested in power consumption in headless mode without hard drives or so, both in idle and full load. And of course, I would be very grateful.

Sorry can't help on the server side, but this is what I get:

Code:

uname -a
Linux Rpro64.dukla.net 4.4.202-1237-rockchip-ayufan-gfd4492386213 #1 SMP Sat Nov 23 13:55:47 UTC 2019 aarch64 aarch64 aarch64 GNU/Linux

I am measuring at the wall using a Brennenstuhl power meter. I happen to be @ 238V AC, and use an industrial 12V DC PSU that, to cause an argument, is 70% efficient. (I really have no idea at these low powers where I am on its curve, or indeed even what its curve is! It is not that old, maybe a year, and not that high powered, maybe 120W, so think 70% is a plausible guess.)

The lowest measure I get is 14.1W, so the RockPro64 is sucking say 9.9W. This reading is pretty constant whether I have the HDMI plugged in or not, whether I am at a boot prompt or running Openbox, whether I have my BT dongle plugged in or not, and whether the NVMe is plugged in or not. I am running Interactive power management so all 6 cores are at 408MHz.

I ran sbc-bench and from time to time watched the meter, as well as my gkrellm monitor.
All of these numbers are as per the Brennenstuhl:
With 1 A53 core fully loaded at 1512MHz the incremental power use is 0.28W
With 1 A72 core fully loaded at 1992MHz the incremental power use is 2.73W
With all 6 cores running 100% in general the incremental power use was 7.53W, I did see 1 spike at +9.53W.

These 'incremental power' observations were all during the 7-zip part of the benchmark, so not the max load that could be done with say neon instructions, but pretty much an average heavy load.

So overall highest ever meter reading was 23.63W so @ 70% my guess is max RockPro64 usage was 16.5W.

Hope this makes sense. And is useful.