You are not logged in.
Pages: 1
Hello guys,
I'm desperate I need help, I used to be able to overclock my GPU but for some reason I cannot anymore.
The ppfeaturemask is set.
I tried with lact and sysfs interface and none of them work for the memory clock. It works for the power cap but other than that when I check the mclk from hwmon or amdgpu_top or mangohud, they all report 1258Mhz.
When I print the output of pp_od_clk_voltage I can see the frequencies I entered but they don't take effect.
OD_SCLK_OFFSET:
300Mhz
OD_MCLK:
0: 97Mhz
1: 1320MHz
OD_VDDGFX_OFFSET:
0mV
OD_RANGE:
SCLK_OFFSET: -500Mhz 1000Mhz
MCLK: 97Mhz 1500Mhz
VDDGFX_OFFSET: -200mv 0mvI'm out of idea, I tried downgrading packages with no success, tried on a fresh install with no success neither.
Offline
I tried on Linux Mint 22.3 and it worked...
Offline
idk wth is going on i'm tired of this
Offline
You might look at the journal on each system (Mint and Arch) and look for clues in the differences.
I don't have your hardware, so I can't offer any other advice.
Nothing is too wonderful to be true, if it be consistent with the laws of nature -- Michael Faraday
The shortest way to ruin a country is to give power to demagogues.— Dionysius of Halicarnassus
---
How to Ask Questions the Smart Way
Offline
Ok thanks I'll do that.
Offline
Nothing relevant to me. What can it be if it's not the kernel or the firmware?
Offline
Sometimes it eventually applies the mem frequency after a while like a few hours idk why.
Offline
Hey, if you have smtg to add use the edit button instead of posting multiple times
is it wayland? how's the behavior on a x11 session?
check amdgpu_top when no browser or elcetron app is running; if it becomes adjustable
Does your monitor have Freesync/Adaptive Sync enabled? if yes try disabling it or lowering your refresh rate to 60Hz temporarily to see if the pp_od_clk_voltage settings suddenly "take."
run `sudo fuser -v /dev/dri/renderD128` when the clock is stuck see if smtg is holding...
Last edited by 5hridhyan (2026-05-04 17:47:32)
"Nothing matters" -a Nihilist
"Why bother thinking what matters?" -me
Offline
why is it a problem that i post multiple times ? it's wayland, thanks i'll try the things you suggested
Offline
Not a big deal, just mentioning it since the Arch guidelines suggest editing instead of multiple posts so threads stay easier to follow
https://wiki.archlinux.org/title/General_guidelines
"Nothing matters" -a Nihilist
"Why bother thinking what matters?" -me
Offline
What kernel version are you on? I am using CachyOS atm and noticed that if I boot 7.x I have the same issue, if I boot LTS 6.x kernel it works fine. I don't have more of a solution atm as I am trying to solve this as well. Here's the output of the LACT service when running kernel 7.0.5, but it's been happening since at least 7.0.3.
lactd.service - LACT GPU Control Daemon
Loaded: loaded (/usr/lib/systemd/system/lactd.service; enabled; preset: disabled)
Active: active (running) since Fri 2026-05-08 15:34:37 EDT; 21min ago
Invocation: 45c09cf1b6c347c8866d0f3e353518d4
Main PID: 1109 (lact)
Tasks: 3 (limit: 38328)
Memory: 130.9M (peak: 472.8M)
CPU: 3.443s
CGroup: /system.slice/lactd.service
└─1109 lact daemon
May 08 15:34:37 Studio-PC systemd[1]: Started LACT GPU Control Daemon.
May 08 15:34:38 Studio-PC lact[1109]: 2026-05-08T19:34:38.050362Z INFO lact_daemon::socket: listening on "/run/lactd.sock"
May 08 15:34:38 Studio-PC lact[1109]: 2026-05-08T19:34:38.054909Z INFO lact_daemon::server: TCP listener disabled
May 08 15:34:38 Studio-PC lact[1109]: 2026-05-08T19:34:38.106432Z INFO lact_daemon::server::handler: AMDGPU DRM initialized
May 08 15:34:38 Studio-PC lact[1109]: 2026-05-08T19:34:38.106606Z INFO lact_daemon::server::handler: initialized amdgpu controller for GPU 1002:687F-1462:3681-0000:09:00.0 at '/sys/class/drm/card1/device'
May 08 15:34:38 Studio-PC lact[1109]: 2026-05-08T19:34:38.106615Z INFO lact_daemon::server::handler: initialized 1 GPUs
May 08 15:34:38 Studio-PC lact[1109]: 2026-05-08T19:34:38.108107Z ERROR lact_daemon::server::gpu_controller::amd: custom clock settings are present but will be ignored, could not get clocks table: parse error: parse error: No sclk range found at line 15 at line 1
May 08 15:35:52 Studio-PC lact[1109]: 2026-05-08T19:35:52.715292Z INFO lact_daemon: got kernel drm subsystem event, reloading GPUs
May 08 15:35:52 Studio-PC lact[1109]: 2026-05-08T19:35:52.721838Z INFO lact_daemon::server::handler: initialized amdgpu controller for GPU 1002:687F-1462:3681-0000:09:00.0 at '/sys/class/drm/card1/device'
May 08 15:35:52 Studio-PC lact[1109]: 2026-05-08T19:35:52.721849Z INFO lact_daemon::server::handler: GPU list reloaded with 1 devices, reapplying configuration
May 08 15:35:52 Studio-PC lact[1109]: 2026-05-08T19:35:52.724379Z ERROR lact_daemon::server::gpu_controller::amd: custom clock settings are present but will be ignored, could not get clocks table: parse error: parse error: No sclk range found at line 15 at line 1
May 08 15:35:52 Studio-PC lact[1109]: 2026-05-08T19:35:52.724553Z INFO lact_daemon::server::handler: configuration applied
May 08 15:35:53 Studio-PC lact[1109]: 2026-05-08T19:35:53.004013Z INFO lact_daemon: got kernel drm subsystem event, reloading GPUs
May 08 15:35:53 Studio-PC lact[1109]: 2026-05-08T19:35:53.010562Z INFO lact_daemon::server::handler: initialized amdgpu controller for GPU 1002:687F-1462:3681-0000:09:00.0 at '/sys/class/drm/card1/device'
May 08 15:35:53 Studio-PC lact[1109]: 2026-05-08T19:35:53.010571Z INFO lact_daemon::server::handler: GPU list reloaded with 1 devices, reapplying configuration
May 08 15:35:53 Studio-PC lact[1109]: 2026-05-08T19:35:53.010671Z ERROR lact_daemon::server::gpu_controller::amd: custom clock settings are present but will be ignored, could not get clocks table: parse error: parse error: No sclk range found at line 15 at line 1
May 08 15:35:53 Studio-PC lact[1109]: 2026-05-08T19:35:53.011054Z INFO lact_daemon::server::handler: configuration applied
May 08 15:35:54 Studio-PC lact[1109]: 2026-05-08T19:35:54.215692Z INFO lact_daemon: got kernel drm subsystem event, reloading GPUs
May 08 15:35:54 Studio-PC lact[1109]: 2026-05-08T19:35:54.222098Z INFO lact_daemon::server::handler: initialized amdgpu controller for GPU 1002:687F-1462:3681-0000:09:00.0 at '/sys/class/drm/card1/device'
May 08 15:35:54 Studio-PC lact[1109]: 2026-05-08T19:35:54.222106Z INFO lact_daemon::server::handler: GPU list reloaded with 1 devices, reapplying configuration
May 08 15:35:54 Studio-PC lact[1109]: 2026-05-08T19:35:54.222318Z ERROR lact_daemon::server::gpu_controller::amd: custom clock settings are present but will be ignored, could not get clocks table: parse error: parse error: No sclk range found at line 15 at line 1
May 08 15:35:54 Studio-PC lact[1109]: 2026-05-08T19:35:54.222481Z INFO lact_daemon::server::handler: configuration applieduname -a output:
Linux Studio-PC 7.0.5-1-cachyos #1 SMP PREEMPT_DYNAMIC Fri, 08 May 2026 11:18:45 +0000 x86_64 GNU/LinuxLast edited by agaskins (2026-05-08 19:57:27)
Offline
Not a big deal, just mentioning it since the Arch guidelines suggest editing instead of multiple posts so threads stay easier to follow
Okay sorry I understand.
What kernel version are you on?
It's 7.x zen, but I had tried the 6.18 and it still occured. When I quit the wayland session and launch it again the mem clock is applied or when I turn off a monitor. Idk, it can't be libdrm because last time I updated was a while ago. It's weird.
Offline
Pages: 1