Server idling at 130W - Intel A380 pinned at 100% Doing Nothing & Not showing up

[hsingh314]

My server is idling at 130W with all the drives spun up, it dips to 80W, but that still seems high. I suspect my Intel A380 is running at 100% even though it doesn't show up on the GPU Statistics, but does show up in System Devices & in Powertop. I understand that there aren't drivers for Intel Arc Gpus at this moment, but if I cannot use it, is there a way that I can make it use less power (short of pulling the card, which isn't possible right now)? 

[SimonF]

4 minutes ago, hsingh314 said:

My server is idling at 130W with all the drives spun up, it dips to 80W, but that still seems high. I suspect my Intel A380 is running at 100% even though it doesn't show up on the GPU Statistics, but does show up in System Devices & in Powertop. I understand that there aren't drivers for Intel Arc Gpus at this moment, but if I cannot use it, is there a way that I can make it use less power (short of pulling the card, which isn't possible right now)? 

Which os vers are you running as you need 6.12 for arc. what does dmesg | grep i915 as you will need a force probe

[hsingh314]

I am running 6.11.5. I am aware about 6.12, but my understanding its not in the 'stable' category (which I am not sure if that matters, but this is my first time doing something like this). 

 

I have attached the output from 'dmesg | grep i915' and I have no idea what any of this means. 

 

I understand that I might not be able to use the intel arc gpu until I update the OS (which I will when its stable), but is there a way that I can reduce its power consumption, instead of it idling at 100%?

[SimonF]

The driver is only seeing the igpu. You need a later kernel which is 6.12

[hsingh314]

Is it safe to, or should I wait for the stable version to be released?

[SimonF]

7 minutes ago, hsingh314 said:

Is it safe to, or should I wait for the stable version to be released?

Its gettimg close to stable so you might want to wait. 

[hsingh314]

Cool, I'll wait. Thank you. 

[hsingh314]

So I updated to 6.12 (yay) - but the driver still doesn't see the a380? Is there something else I need to do. I've seen people mention 'probing' but I am not sure how to do that. 

 

On 5/11/2023 at 5:35 PM, SimonF said:

The driver is only seeing the igpu. You need a later kernel which is 6.12

 

[SimonF]

11 minutes ago, hsingh314 said:

So I updated to 6.12 (yay) - but the driver still doesn't see the a380? Is there something else I need to do. I've seen people mention 'probing' but I am not sure how to do that. 

 

 

You need to have a file ignore guc bit but we need to find code for your card.

 

 

 

root@computenode:~# cat /boot/config/modprobe.d/i915.conf 
options i915 force_probe=56a0
options i915 enable_guc=7
root@computenode:~# 

 

run this command and in the messages it should safe the value for force probe.(Mine doesnt not show as already force probed or bound to vfio)

 

create the file as about with the correct value and reboot. or if you have intel_gpu_top installed it should do the modprobe for you without the file.

 

root@computenode:/tmp# dmesg | grep i915
[   50.400020] i915 0000:00:02.0: [drm] Incompatible option enable_guc=7 - undocumented flag
[   50.400647] i915 0000:00:02.0: [drm] VT-d active for gfx access
[   50.400677] i915 0000:00:02.0: [drm] Using Transparent Hugepages
[   50.401279] i915 0000:00:02.0: vgaarb: changed VGA decodes: olddecodes=io+mem,decodes=none:owns=io+mem
[   50.401727] mei_hdcp 0000:00:16.0-b638ab7e-94e2-4ea2-a552-d1c54b627f04: bound 0000:00:02.0 (ops i915_hdcp_component_ops [i915])
[   50.403380] i915 0000:00:02.0: [drm] Finished loading DMC firmware i915/adls_dmc_ver2_01.bin (v2.1)
[   50.516061] i915 0000:00:02.0: [drm] GuC firmware i915/tgl_guc_70.bin version 70.5.1
[   50.516066] i915 0000:00:02.0: [drm] HuC firmware i915/tgl_huc.bin version 7.9.3
[   50.518499] i915 0000:00:02.0: [drm] HuC authenticated
[   50.518697] i915 0000:00:02.0: [drm] GuC submission enabled
[   50.518698] i915 0000:00:02.0: [drm] GuC SLPC enabled
[   50.518989] i915 0000:00:02.0: [drm] GuC RC: enabled
[   50.519673] mei_pxp 0000:00:16.0-fbf6fcf1-96cf-4e2e-a6a6-1bab8cbe36b1: bound 0000:00:02.0 (ops i915_pxp_tee_component_ops [i915])
[   50.519738] i915 0000:00:02.0: [drm] Protected Xe Path (PXP) protected content support initialized
[   51.024746] [drm] Initialized i915 1.6.0 20201103 for 0000:00:02.0 on minor 0
[   51.025749] i915 0000:00:02.0: [drm] Cannot find any crtc or sizes
[   51.025887] i915 0000:00:02.0: [drm] Cannot find any crtc or sizes

[hsingh314]

40 minutes ago, SimonF said:

You need to have a file ignore guc bit but we need to find code for your card.

 

 

 

root@computenode:~# cat /boot/config/modprobe.d/i915.conf 
options i915 force_probe=56a0
options i915 enable_guc=7
root@computenode:~# 

 

run this command and in the messages it should safe the value for force probe.(Mine doesnt not show as already force probed or bound to vfio)

 

create the file as about with the correct value and reboot. or if you have intel_gpu_top installed it should do the modprobe for you without the file.

 

root@computenode:/tmp# dmesg | grep i915
[   50.400020] i915 0000:00:02.0: [drm] Incompatible option enable_guc=7 - undocumented flag
[   50.400647] i915 0000:00:02.0: [drm] VT-d active for gfx access
[   50.400677] i915 0000:00:02.0: [drm] Using Transparent Hugepages
[   50.401279] i915 0000:00:02.0: vgaarb: changed VGA decodes: olddecodes=io+mem,decodes=none:owns=io+mem
[   50.401727] mei_hdcp 0000:00:16.0-b638ab7e-94e2-4ea2-a552-d1c54b627f04: bound 0000:00:02.0 (ops i915_hdcp_component_ops [i915])
[   50.403380] i915 0000:00:02.0: [drm] Finished loading DMC firmware i915/adls_dmc_ver2_01.bin (v2.1)
[   50.516061] i915 0000:00:02.0: [drm] GuC firmware i915/tgl_guc_70.bin version 70.5.1
[   50.516066] i915 0000:00:02.0: [drm] HuC firmware i915/tgl_huc.bin version 7.9.3
[   50.518499] i915 0000:00:02.0: [drm] HuC authenticated
[   50.518697] i915 0000:00:02.0: [drm] GuC submission enabled
[   50.518698] i915 0000:00:02.0: [drm] GuC SLPC enabled
[   50.518989] i915 0000:00:02.0: [drm] GuC RC: enabled
[   50.519673] mei_pxp 0000:00:16.0-fbf6fcf1-96cf-4e2e-a6a6-1bab8cbe36b1: bound 0000:00:02.0 (ops i915_pxp_tee_component_ops [i915])
[   50.519738] i915 0000:00:02.0: [drm] Protected Xe Path (PXP) protected content support initialized
[   51.024746] [drm] Initialized i915 1.6.0 20201103 for 0000:00:02.0 on minor 0
[   51.025749] i915 0000:00:02.0: [drm] Cannot find any crtc or sizes
[   51.025887] i915 0000:00:02.0: [drm] Cannot find any crtc or sizes

Ugh - I am not sure what I am doing wrong. Also I do have intel gpu top installed but the a380 doesn't show up there either. 

[SimonF]

4 minutes ago, hsingh314 said:

Ugh - I am not sure what I am doing wrong. Also I do have intel gpu top installed but the a380 doesn't show up there either. 

 

your i915.conf should be the following 56a0 is for ARC770 yours is 56a5 as shown ^ 

options i915 force_probe=56a5
 

remove blacklist

[hsingh314]

Ok, 

 

I figured it out. I just typed in exactly what you had, not realizing that was the output of the .conf file & I had to edit the .conf file to have those parameters (lol - I am dumb). I got it work, or at least Unraid to detect the card. Thanks!