search

jxai / lean-conky-config

Lean Conky config that just works
MIT License
128 stars 31 forks source link

Add basic AMD support via amdgpu_top #32

Open ChaosInfinited opened 5 months ago

ChaosInfinited commented 5 months ago

Hi, I've came up with following solution for AMD GPUs, it's not the prettiest but it works for me, it requires lua-json and amdgpu_top, if you decide that it's not good enough, maybe some of the code will be useful for someone? :) Will gladly adjust it further.

image

I've tried using using similar approach as you did for nvidia.

ChaosInfinited commented 5 months ago

Actually I've just noticed that the graphs aren't working as intended, not sure if I've missed something for my amd implementation, or it's the same with Nvidia..? There's no historical data on them..

And top it's working right either work now, should be an easy fix, will push it soon.

ChaosInfinited commented 5 months ago

Fixed the gpu top, but I'm still unable to fix the graphs, if I try rendering them directly lcc.tpl.amd = [[${lua_graph "echo {%= math.random(5,50) %}"}]] then it renders correctly, so I'm assuming that conky's loosing a point of reference/history when the template's being rendered via lua..? lcc.tpl.amd = [[${lua amd {%= interv %} {%= top_n %}}]].. can't tell if it's an issue with my code or Conky itself :thinking:

I think it's an issue with Conky, network graphs are behaving in the same way.