IIUC, this is basically just a python script that wraps vcgencmd
and displays it interactively or logs the values? I wonder what's the performance impact of doing that, especially for so many values that might be of no interest to the user.
I think a better approach would be to find out what sysfs interface vcgencmd is using for things, and then accessing those nodes directly. That eliminates the overhead of running it and parsing its output. Then, give users an option of which stats they want, so you don't waste time polling stuff they don't care about. Basically, they should be trying to replicate what turbostat
does.
The efficiency of monitoring tools can be really important, since generally want the tool to have minimal impact on either the system or whatever benchmark it happens to be running.
Finally, I hope it supports polling intervals above 1 Hz. A lot of these values can change significantly, in much less time than that. For my own performance measurements, I tend to use a polling interval of 0.1 sec (i.e. 10 Hz).