Closed b3m17 closed 1 year ago
hmm no, as you have amd gpu, the nvidia packages are not expected to be installed, so the script should be detecting it... This is a bug...
I'm getting this. Do I need the AMD-Pro drivers? Not showing in Waybar just CPU & Memory
AMD-Pro drivers
It doesn't work for me either. I also tried with AMD-Pro drivers with the same result. I then returned to the open amd drivers.
I can fix this tomorrow morning. — looks like the AMD gpu_info is probably grabbing from the wrong location.
I don’t have a AMD GPU to verify but I can do a script refresh tomorrow.
Sorry -- Family emergency yesterday.. Working on this now
Sorry -- Family emergency yesterday.. Working on this now
Hey it's okay, man. Take your time. I hope everything turns out alright for you and your family.
Sorry -- Family emergency yesterday.. Working on this now
Relax brother, take care of yourself and family...
I'm almost done! This was more work than I thought lol -- but im about to submit my PR
**DONE
I'm almost done! This was more work than I thought lol -- but im about to submit my PR
**DONE
I just updated to your commit and get this result:
Do you have an AMD integrated?
Running into a lot of road blocks on the ability to get a grepable output for the AMD integrated -- may be the same for the intel gpu -- haven't tested yet
Do you have an AMD integrated?
Running into a lot of road blocks on the ability to get a grepable output for the AMD integrated -- may be the same for the intel gpu -- haven't tested yet
Yes I have an integrated AMD on an laptop serface 4 AMD
❯ lspci -v | grep -A 12 "VGA controller"
03:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] Renoir (rev e3) (prog-if 00 [VGA controller])
Subsystem: Microsoft Corporation Renoir
Flags: bus master, fast devsel, latency 0, IRQ 47
Memory at 2b0000000 (64-bit, prefetchable) [size=256M]
Memory at 2c0000000 (64-bit, prefetchable) [size=2M]
I/O ports at 1000 [size=256]
Memory at f1300000 (32-bit, non-prefetchable) [size=512K]
Capabilities:
03:00.1 Audio device: Advanced Micro Devices, Inc. [AMD/ATI] Renoir Radeon High Definition Audio Controller Subsystem: Advanced Micro Devices, Inc. [AMD/ATI] Renoir Radeon High Definition Audio Controller
https://github.com/Umio-Yasuno/amdgpu_top
im working on integrating this package into the script as its the only greppable format cli tool for amd gpu's that i've found
amdgpu_top -J and -d >>> puts in json format or dumps. Need a little time to dial it in!!
Here's the result on my end
Do you have an AMD integrated?
Running into a lot of road blocks on the ability to get a grepable output for the AMD integrated -- may be the same for the intel gpu -- haven't tested yet
I also get the same result on the desktop with discrete AMD GPU: If you want I can act as a guinea pig with the desktop. Maybe I'll sell the laptop today :)
Can you go to my repo and clone and run the install script so I don't have to open a PR until it is ready?
I have updated the script to use amdgpu_top and formatted the output similar to the nvidia-smi output.
REPO: https://github.com/T-Crypt/hyprdots
Note: as I test you will see commits so you may need to do a git pull after verifying you have the most recent commits i have made.
for first use make sure to run ./gpuinfo.sh and it will install the needed packages.
for first use make sure to run ./gpuinfo.sh and it will install the needed packages.
I get this:
Do you get an ouput with running amdgpu_top -d
That doesn't look like the updated script, output should be:
These commands should net these results:
Okay working on adding debugging statements to see where this is getting caught up.
Okay try now from my commit - 9 minutes ago.
** Now im getting some odd error where the gpu status isn't even showing.
@T-Crypt tried your gpuinfo.sh with 6800XT & it now show in waybar but no temp or power But running amdgpu_top -d it shows it? I'll also check on my laptop when I get to work with integrated AMD
Yeah I need to refine the script today / tomorrow to get this working. I've got mostly there -- When I actually echo the $gpu_info_amd -- it seems to produce the output for the script but for some reason messes up the whole waybar gpu module as I believe the output is not formatted correctly in "json" like the gpuinfo.jsonc has it formatted.
I apologize for the issues guys, working hard on getting this to work for AMD gpu's and integrated as well -- seems nvidia was straight forward were as AMD wants me to jump through hoops for the needed stats.
Thank you bro, I think amd should be straight forward without the amd_top, you could try these files. Please lemme know if you need anything from me
Thank you bro, I think amd should be straight forward without the amd_top, you could try these files. Please lemme know if you need anything from me
I looked into this yesterday and thats what I was going to try this morning -- It just has to have a lot of conversion to the right metrics and converting the output to the same format as the nvidia -- the only one that gives a straight result no conversion needed is the
cat /sys/class/drm/card0/device/gpu_busy_percent
cool, there is also this nvtop... I've not tried this, but you can check it out if you find it useful... it shows gpu info for all nvidia/intel/amd cards.
There is also python libraries available from gpu vendors that can be used in a python script to fetch the gpu info.. i might mess around with it later when i get some time!
I updated the script
https://github.com/T-Crypt/hyprdots/blob/main/Configs/.config/hypr/scripts/gpuinfo.sh
my next option is like you said python libraries:
I've got it working with Python!!
Now I just need to modify the original script to call for this in:
amd_info=$(python3 ~/.config/hypr/scripts/amdgpu.py)
**UPDATE: can one of you try with your AMD gpu to the updated commits?
No icon in waybar after reboot. What you see is the CPU temperature (screenshot). Now the shell shows signs of life :) When I launch the script from the terminal it always reinstalls python-pyamdgpuinfo then prints the output (screenshot)
Okay I removed the install function and placed python-pyamdgpuinfo in custom_hypr.lst so it doesn't try to reinstall over and over.
Should be installed when you run ./install.sh and then that way the package is already installed on first boot and not when the script is triggered.
Works! Thank you :)
Ill fix the MHz / GHz format today so that is resolved!!
Finally got this working -- Makes me happy!
Ill fix the MHz / GHz format today so that is resolved!!
Finally got this working -- Makes me happy!
good job !!!
If it's not a problem, I would like to point out that there are problems with Intel GPUs too:
Yeah -- that's the next hurdle for today's resolve!!
Ill work on this now, it may require python as well to integrate the results.
I noticed it now that there are both Mhz and Ghz units :)
ok! I am available in case of testing.
I noticed it now that there are both Mhz and Ghz units :)
I updated this in the pull request -- just need to update to the latest commit :)
Can you do me a favor and use this command and paste your output? I dont have an intel device to test on with arch or linux --- Just WSL :( my main is a nvidia and AMD apu
only on your intel igpu device!
run this in terminal
sensors
I noticed that it doesn't reveal the intel gpu temp. I did not notice that. I installed arch linux on this laptop yesterday To you. Anyway, here's the output:
Now I have to go out. Later :)
I need someone to run this with a intel integrated GPU
yes | sudo sensors-detect
and then paste the output of the command below in here:
sensors
I ran "yes | sudo sensors-detect" but it doesn't reveal the igpu temp. I'll look around to see if I can find a solution for this damn sensor. :)
I think its because its handled with a ACPI sensor..
this one is a little more complicated -- may just need to use the coretemp
CPU and GPU temperatures are the same thing, they both live on the same piece of silicon. (quote https://bbs.archlinux.org/viewtopic.php?id=223852) Is flawless
CPU and GPU temperatures are the same thing, they both live on the same piece of silicon. (quote https://bbs.archlinux.org/viewtopic.php?id=223852) Is flawless
I think it is fair to say that for Intel igpu's the coretemp of the CPU package ID 0: will most likely be 5°C higher or lower than the igpu sensor temp. Of course, only to a degree as the proximity would necessarily even things out, but still they can heat up differently.
For now you can at least monitor the whole package temp with this updated script in my test branch! You want to download this script and run this on your intel machine to see what result you get?
https://github.com/T-Crypt/hyprdots/blob/test/Configs/.config/hypr/scripts/gpuinfo.sh
ok I'll try now
Is there supposed to be a package I have to install first in order for this to work?