Index | Thread | Search

From:
Kirill A. Korinsky <kirill@korins.ky>
Subject:
Re: powersave CPU policy
To:
"Lorenz (xha)" <me@xha.li>
Cc:
Philip Guenther <guenther@gmail.com>, OpenBSD tech <tech@openbsd.org>
Date:
Mon, 03 Jun 2024 15:09:58 +0100

Download raw body.

Thread
On Mon, 03 Jun 2024 13:27:19 +0100,
"Lorenz (xha)" <me@xha.li> wrote:
> 
> On Mon, Jun 03, 2024 at 01:18:38PM +0100, Kirill A. Korinsky wrote:
> > I've tested that this logic works by inverting it (the lightest C-State) and
> > confirm how fast battery is drawn on my machine.
> 
> just out of curiosity, do you have some numbers?
> 

As a base line I use my usual workflow which include:
 - emacs for mail and write code;
 - chrome for slack and telegram;
 - some browsering with chrome.

Such protocol doesn't allow to measure anything clean, but allows to get
some numbers.

"auto" policy allows to run until suspend (on 8% battery) for about 2 hours.

"powersave" policy without enforcing the deepst C-state runs about 3 hours.

"powersave" with enforce the deepest C-state runs about 3.5 hours.

A hacked version with inverted logic of selection C-state runs ~1.5 hours.

This numbers from Huawei Matebook X 2020 which has not that new batter:

  hw.sensors.acpibat0.amphour3=2.35 Ah (remaining capacity), OK
  hw.sensors.acpibat0.amphour4=3.62 Ah (design capacity)

Bettter testing requires dedicated hardware which I lack right now.

So, if anyone willing to test it, I'll be appricieted.

-- 
wbr, Kirill