performance tuning on raspberry pi 1st generation

I’m submitting a …

• bug report
• feature request
• question about the decisions made in the repository

Do you want to request a feature or report a bug?

performance bug

What is the current behavior?

CherryPy consumes about 15-20% CPU on a single-core raspberry pi 1st generation when idle, ie. no requests being processed at all. To reduce this, I tried various configuration adjustments, but to no avail. How can I reduce this ?

If the current behavior is a bug, please provide the steps to reproduce and if possible a screenshots and logs of the problem. If you can, show us your code.

>>> import cherrypy
>>> cherrypy.quickstart()
[29/Apr/2021:14:14:22] ENGINE Listening for SIGTERM.
[29/Apr/2021:14:14:22] ENGINE Listening for SIGHUP.
[29/Apr/2021:14:14:22] ENGINE Listening for SIGUSR1.
[29/Apr/2021:14:14:22] ENGINE Bus STARTING
[29/Apr/2021:14:14:22] ENGINE Started monitor thread 'Autoreloader'.
[29/Apr/2021:14:14:23] ENGINE Serving on http://127.0.0.1
[29/Apr/2021:14:14:23] ENGINE Bus STARTED

top output:

Tasks:  85 total,   1 running,  84 sleeping,   0 stopped,   0 zombie
%Cpu(s): 23.9 us,  8.9 sy,  0.0 ni, 66.6 id,  0.0 wa,  0.0 hi,  0.7 si,  0.0 st
MiB Mem :    430.1 total,    152.9 free,     76.2 used,    201.0 buff/cache
MiB Swap:    100.0 total,    100.0 free,      0.0 used.    297.3 avail Mem 

  PID USER      PR  NI    VIRT    RES    SHR S  %CPU  %MEM     TIME+ COMMAND                                                                                             
10674 root      20   0  131484  21240   8188 S  18.5   4.8   0:45.63 python3                                                                                             

What is the expected behavior?

when idle, no noticeable CPU consumption even on a single-core raspberry pi (or generally ‘low-performance devices’)

What is the motivation / use case for changing the behavior?

useability on low-performance devices

Please tell us about your environment:

• Cheroot version: ?
• CherryPy version: 18.6.0
• Python version: 3.9.2
• OS: Raspbian GNU/Linux 10 (buster)
• Browser: [all ]

Other information (e.g. detailed explanation, stacktraces, related issues, suggestions how to fix, links for us to have context, e.g. stackoverflow, gitter, etc.)

have tried to reduce server.thread_pool to 1, unfortunately no CPU reduction

[webknjaz]

You might want to try CherryPy<9.0.0, as this is where the engine was switched from wsgiserver to cheroot: https://github.com/cherrypy/cherrypy/blob/main/CHANGES.rst#v900

It shouldn’t matter. The “switch” was just moving code from one place to another w/o changes. It was an organizational change, not a functional one. Of course, over time it got refactored but I don’t believe it’d influence stuff this much. OTOH, it may be a good idea to check if the use of cheroot alone produces a similar amount of load or not. The invocation examples are here: https://cheroot.readthedocs.io/en/latest/pkg/cheroot.cli.html.

[Blindfreddy]

Haven’t had time to use py-spy but thought I’d reduce the threadpool from default 10 to 1: sadly no change, but at least now it’s clear that the worker threads aren’t the root cause. Next I’ll turn off the Autoreloader thread. Keep you posted.