I have a Dell XPS 15 7590, and I’ve noticed that, after the laptop is woken up from suspend, the power consumption is twice as high than after a normal boot.
I used powertop to measure this. After boot, the measured discharge rate is ~10W per hour (which is not great but OK, given that the laptop has a large OLED screen). If I suspend the laptop (in “deep” mode), after resume powertop measures a discharge rate of ~20W per hour. This doesn’t happen if I use the “s2idle” mode for suspension (when resuming the power consumption is the same of startup).
Unfortunately, using suspension in s2idle mode is far from ideal, because the laptop discharges quite quickly (~5.5W per hour) compared to the deep mode. On the other hand, the most efficient suspension mode gives me high battery drain after resume.
I’ve made some attempts to discover what is the source of higher battery usage. It’s like suspension in deep mode erases some configuration applied at startup which saves battery usage. Things I’ve tried:
restarting tlp
re-switching off the Nvidia GPU (with the script provided by optimus-switch)
But the battery drain is still there. I’m afraid I don’t have many other ideas, as I’m not expert in this. Any suggestion is appreciated.
Thanks for your suggestion but… VLC is not even installed on my machine. I have installed htop but I don’t know exactly what to look at. Apparently there is no process using a lot of CPU – I also looked in the system monitor… but the discharge rate is insanely high (~27W).
As you didn’t provide much information, have you tried other kernels yet? (5.4 LTS and 5.10 LTS come to mind)
Also, an inxi --admin --verbosity=7 --filter --no-host --width would be the minimum required information to help you further if even the kernel suggestion above doesn’t provide any solace… (Personally Identifiable Information like serial numbers and MAC addresses will be filtered out by the above command)
I have tried both kernels 510 and 54, and the behaviour is the same. I have tried also to reset the wifi (unload and reload the iwlwifi module), but that doesn’t change anything. Unfortunately, powertop reports a (very) different discharge rate pre- and post- deep suspension, but the details about which process is using energy looks roughly the same. Wake-up from suspend elicits some power usage that powertop is not able to detect. And I don’t know if there are other apps that can do that. To the best of my knowledge, it seems that all apps of this kind look at CPU usage, which doesn’t help in my case.
Just to give an idea of how power consuption changes after suspension. You can guess the point at which the laptop was suspended (with deep suspension, let me reiterate, with s2idle suspension this doesn’t happen)
This is just to confirm that it’s not powertop getting crazy and reporting a discharge that’s not happening.
I seem to have solved by doing some cleaning in the /etc/default/grub file. I removed some kernel options that I’d had since the beginning and found out were (at best) unnecessary if not damaging, and replaced with the suggestions found here. So now the line with the kernel options in my /etc/default/grub is as follows:
Energy usage has improved (~5W when the laptop “rests”), and resuming after deep suspension and hibernation doesn’t result in increased power usage. Hurray!