I can't apply my configuration about lact

[TeaHistoria]

Here my ROG Advantage Specs with RX 6800M

OS: Arch Linux x86_64
Host: ROG Strix G513QY_G513QY 1.0
Kernel: 6.2.8-zen1-1-zen
Uptime: 1 hour, 21 mins
Packages: 1563 (pacman), 6 (flatpak)
Shell: bash 5.1.16
Resolution: 2560x1440
DE: GNOME 43.3
WM: Mutter
WM Theme: Adwaita
Theme: Adwaita-dark [GTK2/3]
Icons: Numix-Circle [GTK2/3]
Terminal: gnome-terminal
CPU: AMD Ryzen 9 5900HX with Radeon Graphics (16) @ 3.300GHz
GPU: AMD ATI Radeon Vega Series / Radeon Vega Mobile Series
GPU: AMD ATI Radeon RX 6700/6700 XT/6750 XT / 6800M/6850M XT
Memory: 5803MiB / 15391MiB

[TeaHistoria]

Maybe something is wrong with my configuration but I've tried on and off kernel parameters and nothing happened.

/etc/modprobe.d/amdgpu.conf

blacklist radeon
options radeon si_support=0
options radeon cik_support=0
options amdgpu si_support=1
options amdgpu cik_support=1
options amdgpu ppfeaturemask=0xffffffff

/etc/default/grub

GRUB_CMDLINE_LINUX_DEFAULT="loglevel=3 quiet radeon.si_support=0 radeon.cik_support=0 amdgpu.si_support=1 amdgpu.cik_support=1 amdgpu.ppfeaturemask=0xffffffff"

[ilya-zlobintsev]

Did you enable and start the daemon? You can check it with systemctl status lactd
LACT gives you a warning on startup about not being able to change settings if you don't have the daemon running.

[TeaHistoria]

yeah i enable it

lactd.service - AMDGPU Control Daemon
Loaded: loaded (/usr/lib/systemd/system/lactd.service; enabled; preset: disabled)
Active: active (running) since Tue 2023-04-04 20:32:32 WIB; 6h left
Main PID: 829 (lact)
Tasks: 3 (limit: 18423)
Memory: 12.5M
CPU: 60ms
CGroup: /system.slice/lactd.service
└─829 lact daemon

Apr 04 20:32:32 anarki systemd[1]: Started AMDGPU Control Daemon.
Apr 04 20:32:32 anarki lact[829]: 2023-04-04T13:32:32.090574Z INFO lact_daemon::socket: listening on "/var/run/lactd.sock"

[TeaHistoria]

maybe there really was an error in my arch linux installation, when i ran the command

cat /sys/class/drm/card1/device/pp_sclk_od
0

the output is "0" instead

I just realized 6 hours ago. maybe I'll reinstall the OS again when I have free time and try the LACT program and report the results in this comment

[ilya-zlobintsev]

Could you set log_level to debug in /etc/lact/config.yaml, then try applying settings again and check the logs in

sudo journalctl -u lactd -e | grep PowerCap

Looks like laptop GPUs don't allow editing some settings that LACT tries to apply

[ilya-zlobintsev]

maybe there really was an error in my arch linux installation, when i ran the command

cat /sys/class/drm/card1/device/pp_sclk_od 0

the output is "0" instead

I just realized 6 hours ago. maybe I'll reinstall the OS again when I have free time and try the LACT program and report the results in this comment

This file is not used by LACT, it's normal for it to output 0. It's unlikely to be a problem with your installation, and more to do with specific gpu type quirks (laptop GPU in your example, which LACT was not heavily tested with)

[TeaHistoria]

but from the first time I bought this laptop 2 years ago, I had a problem with the vram mclk stuck at 1000mhz and never found a solution

[TeaHistoria]

This file is not used by LACT, it's normal for it to output 0. It's unlikely to be a problem with your installation, and more to do with specific gpu type quirks (laptop GPU in your example, which LACT was not heavily tested with)

people who run this command get the result like below

sudo cat /sys/class/drm/card0/device/pp_od_clk_voltage

OD_SCLK:
0: 300MHz 750mV
1: 588MHz 765mV
2: 980MHz 987mV
3: 1100MHz 950mV
4: 1100MHz 950mV
5: 1100MHz 950mV
6: 1100MHz 950mV
7: 1100MHz 950mV
OD_MCLK:
0: 300MHz 750mV
1: 1000MHz 800mV
2: 1970MHz 950mV
OD_RANGE:
SCLK: 300MHz 2000MHz
MCLK: 300MHz 2250MHz
VDDC: 750mV 1150mV

But the results I get are blank

[TeaHistoria]

[tea@anarki | ~]
==> cat /sys/class/drm/card0/device/pp_dpm_mclk
0: 96Mhz *
1: 456Mhz
2: 675Mhz
3: 1000Mhz
[tea@anarki | ~]
==> sudo cat /sys/class/drm/card0/device/pp_od_clk_voltage
[sudo] password for tea:

[tea@anarki | ~]
==> cat /sys/class/drm/card0/device/pp_od_clk_voltage

[tea@anarki | ~]
==> cat /sys/class/drm/card1/device/pp_od_clk_voltage

[tea@anarki | ~]
==> cat /sys/class/drm/card0/device/pp_dpm_mclk
0: 96Mhz *
1: 456Mhz
2: 675Mhz
3: 1000Mhz
[tea@anarki | ~]
==> cat /sys/class/drm/card1/device/pp_dpm_mclk
0: 1600Mhz
1: 400Mhz *
[tea@anarki | ~]

[ilya-zlobintsev]

But the results I get are blank

This is expected on integrated and laptop GPUs, and is not the reason you can't apply settings. The error comes from not being able to set the power limit (which may not be configurable on a laptop)

[TeaHistoria]

maybe, this is because the newly introduced sPPT settings don't have the right driver, so what I'm looking for on the Internet are very few posts with the keyword "Linux sPPT"

Indeed, this is a bit strange where the CPU and GPU power consumption are combined into the sPPT Platform. When I set the CPU limit to be bigger than the GPU (APU sPPT), the power is focused on the CPU, so the power on the GPU is left over from the limit.

APU sPPT = 60W
Platforms sPPT = 100W

if programs/games that require a lot of CPU are usually like this
Platforms sPPT (Total power) = 100W
CPU = 60W
GPUs = 40W

and when a program/game is bigger to GPU performance.. then
Platforms sPPT (Total power) = 100W
CPU = 10W
GPUs = 90W

actually for what AMD to make laptop settings like this?

[ilya-zlobintsev]

LACT will now avoid setting the power limit unless it was changed by the user, this will not make the setting available on your system, but it should make it possible to change other settings.