Fedora 28 Will Hopefully Enable Intel PSR To Further Conserve Laptop Power

[phoronix]

Phoronix: Fedora 28 Will Hopefully Enable Intel PSR To Further Conserve Laptop Power

Red Hat developer Hans de Goede has recently been on a mission to improve Linux battery life on Fedora. Now that SATA link power management is better handled and other tweaks, his latest target is on getting Intel's Panel Self Refresh (PSR) support enabled...

http://www.phoronix.com/scan.php?page=news_item&px=Fedora-28-Aims-For-PSR

[starshipeleven]

I suppose this does not work for AMD apus with eDP screens.

[molecule-eye]

The ALPM SATA fix is a real blessing. I'll certainly try out PSR.

[numacross]

I wish them good luck since feature-parity with Windows in terms of power management is always a good thing. Unfortunately even my "Linux friendly" Dell XPS 13 requires a long kernel command line in order to enable every bit of PM. After that a few tweaks with powertop and it's idling at ~3.5W with wireless on and minimal brightness on the QHD screen.

The problem with PSR is that it can work partially, as is the case with my laptop, and require different values than 1 for the module configuration (https://patchwork.kernel.org/patch/8182841/) - mine needs 2 in order to be stable on 4.14.

[brent]

I wish them good luck since feature-parity with Windows in terms of power management is always a good thing. Unfortunately even my "Linux friendly" Dell XPS 13 requires a long kernel command line in order to enable every bit of PM. After that a few tweaks with powertop and it's idling at ~3.5W with wireless on and minimal brightness on the QHD screen.

Sounds good, can you share those tweaks? My XPS 13 9350 with FHD display still idles at around 3.9W. I presume it could be quite a bit more efficient.

[numacross]

Sounds good, can you share those tweaks? My XPS 13 9350 with FHD display still idles at around 3.9W. I presume it could be quite a bit more efficient.

Apart from this https://wiki.archlinux.org/index.php..._13_%289350%29 I've disabled the card reader in BIOS since it actually took power when not used (probably because of lacking ASPM support in the kernel).

[microcode]

hansdegoede For some reason LiveJournal won't show me the required captcha to post my comment, so I'll do it here instead.

I notice you chose to refer to the Battery reported discharge rate field instead of the Total power consumed (unfortunately mislabeled as power rather than energy) field. The battery/power supply reported wattage is probably the wrong number to look at, since what you're really interested in is the amount of energy.

Powertop takes measurements in 20-second increments. Using my laptop as an example, powertop reports it drew 73.7J over 20 seconds (so average power draw was 73.7 / 20 = ~3.69W), but reported a discharge rate of 3.82W (equivalent to 76.4J over 20 seconds).

It may be a better idea to take the Total power consumed number instead of the Battery reported discharge rate, since the latter is just whatever the battery currently reports, which is noisy compared to the energy measurement. You could even take down a couple of them, and divide by any multiple of 20.

Cheers, and happy hacking.

[Templar82]

It would be great to see Fedora and Linux in general make some improvements in this area.
For a 'Lighter' operating system than windows, its kinda a shame that power consumption is usually higher.

[hansdegoede]

Update: Thank you everyone for all the test-reports I've received. The response has been quite overwhelming, with well over 50 test-reports received sofar. The results are all over the place, some people see no changes, some people report the aprox. 0.5W saving my own test show and many people also report display problems, sometimes combined with a significant increase in power-consumption. I need to take a closer look at all the results, but right now I believe that the best way forward with this is (unfortunately) a whitelist matching on a combination of panel-id (from edid) and dmi data, so that we can at least enable this on popular models (any model with at least one user willing to contribute).

[starshipeleven]

hansdegoede any chance this could be done also for laptops with AMD APUs? I understand that they are not really common, but for example Lenovo Ideapad 310 have AMD apus and eDP screens (I have one and I swapped the screen to an IPS so I know for sure this is the case), so do the HP Envy laptops with AMD apus (again I opened some).