Skip to content
View in the app

A better way to browse. Learn more.

Unraid

A full-screen app on your home screen with push notifications, badges and more.

To install this app on iOS and iPadOS
  1. Tap the Share icon in Safari
  2. Scroll the menu and tap Add to Home Screen.
  3. Tap Add in the top-right corner.
To install this app on Android
  1. Tap the 3-dot menu (⋮) in the top-right corner of the browser.
  2. Tap Add to Home screen or Install app.
  3. Confirm by tapping Install.

[Plugin] Intel-GVT-g

Featured Replies

7 hours ago, alturismo said:

its most likely when the RDO resolution is > 1080

when i tested all through, that was max to get it decent long running, once i went higher (as client had higher res) then i ran sonner or faster in oom error ... thats why i recommend parsec, in your case with ios may not the best option ;) so sunshine / moonlight would be a alternative, but as you mentioned also not so nice on ios.

I ended up just not using GPU acceleration and honestly it's not bad at all using windows RDP.

I think essentially how optimized windows RDP is > how much GPU power I get using an alternative client is the case for me. I do just basic heavy browsing, some mkvtoolnix muxing, YouTube / video watching, etc.

  • Replies 784
  • Views 225.5k
  • Created
  • Last Reply

Top Posters In This Topic

Most Popular Posts

  • Tutorial (Windows 10 VM)   First of all we will create a new VM with the latest Q35 "Machine" type (you can also use an existing VM but make sure that it actually set to to the latest Q35 "M

  • Exactly.   This won't work by now, please don't try it any further, this is a dead end for now. I will report back if I got some news for you.

  • Don't mind me. I had hoped to quickly and effortlessly set up Intel iGPU support for a Windows 7 VM (Windows 10 is not an option). The Unraid server is just a hobby anyway for which I don't want t

Posted Images

1 hour ago, Linguafoeda said:

I think essentially how optimized windows RDP is > how much GPU power I get using an alternative client is the case for me. I do just basic heavy browsing, some mkvtoolnix muxing, YouTube / video watching, etc.

yep, RDP is also my favour ...

but remote like guac its not working so nice, therefore parsec is my way to go.

@home in LAN RDP is also 1st choice here ;)

  • 1 month later...

Hi

Apologies I know I've commented on a few different driver posts. So I have experimented with NVIDIA and now I'm trying out IGPU driver. https://forums.unraid.net/topic/98978-plugin-nvidia-driver/page/207/

This is a screenshot from a DOCKER container using IGPU:

image.png image.png

So still OPUS transcoding audio and yes you were right there are hard encoded subtitles. IGPU is working significantly less hard than the CPU with the NVIDIA old GT750ti GPU (in docker).

I do however have a question.

So I am also testing environments, and below is a Win11 VM (unlike the docker above):

image.pngimage.png

You can see that the VM the GPU is working less hard and the CPU is working harder than the docker.

image.png

I did see the intel come up under GPU in the task manager performance; I also saw the 2nd display in hardware devices.

When I open Youtube I didn't really notice anything.

Also now if I Ctrl+Alt+Del in the VNC - my VNC screen just turns black. Also start menu doesn't open so its not liking it at all.

I'm a little lost tbh

I still want to compare all my options using a GPU or IGPU. I want both to be compatible in VM and Docker Environments.

Could you advise what I should do to try and correct the iGPU VM issues encountered?

Cheers Chris )

Edited by SmokeyColes

Update:

After removing the assignment, I can reuse VM again.

So I ended the VM, and reapplied the assignment.

image.png

I think I broke it... 🙃

UUID failed to delete or out of VRAM. I will reboot and try again.

Edited by SmokeyColes

  • Author
27 minutes ago, SmokeyColes said:

So still OPUS transcoding audio and yes you were right there are hard encoded subtitles. IGPU is working significantly less hard than the CPU with the NVIDIA old GT750ti GPU (in docker).

Because you are now transcoding to 21Mbps, this is another transcode, you can't compare 4K Dolby Vision with 4K transcoding and different bitrates.

29 minutes ago, SmokeyColes said:

You can see that the VM the GPU is working less hard and the CPU is working harder than the docker.

Yes and no, this is a virtualized environment and I don't know what resources you've assigned to the VM and also GVT-g is not the same as the if you where using the iGPU directly.

These are not equal comparisons, also you test Linux against Windows and these two systems work way differently, I don't even know what Plex is using on Windows and if they are using something else than a modified version from FFMPEG.

However, what is your goal here with these tests?

I really don't get the purpose here, sorry for asking but I don't see anything wrong.

32 minutes ago, SmokeyColes said:

I want both to be compatible in VM and Docker Environments.

Just a word of warning, GVT-g is notorious to crash your whole Server at times when you exceed the VRAM <- on most motherboards you can't set it to something higher to 32 or 64MB and even that is not guaranteed to work properly because the GVT-g implementation is hit or miss at best from most Motherboard manufacturers.

However you can accomplish that with GVT-g what you want to do here, at least if we are talking a CPU from 5th to 10th gen Core series.

For Nvidia you won't be able to do that, at least not easily.

36 minutes ago, SmokeyColes said:

Could you advise what I should do to try and correct the iGPU VM issues encountered?

Not really because it can also depend on the installed iGPU driver.

Sadly I have to say that GVT-g isn't supported very well anymore.

23 minutes ago, SmokeyColes said:

So I ended the VM, and reapplied the assignment.

You have to edit the VM XML manually and remove the mediated devices.

Ok I will look for the mediated devices.

Ok so the end goal is I want a Plex server capable of transcoding. Without a gpu or igpu it states the hardware is not sufficient to transcode; it’s an error which pops up on the client.

The problem is I have no PCI ports.

As far as I can work out I have two options.

  1. A GPU either intel or Nvidia connected by thunderbolt as a egpu.

  2. Using IGPU as I have a series 10 intel cpu.

Truthfully I’ve no idea what I’m doing but I am learning a little about it being a lot of effort to get something set up. I’m inexperienced with dockers. That said I’m getting better by doing this hence the tests with VMs and dockers.

If you have any ideas I’d greatly appreciate it. My motherboard is fairly OK. It has 3x PCIx16 which I use for 3x sas controllers to run 3 lots of 8 drives.

I think it would also be confined with room if I considered upgrading one controller to something which supports 16 hdds. I’m also reluctant as that bit of the server works well.

I’m thinking the docker and the intel driver seems to be my best option. Thunderbolt still feels a bit problematic when it comes to gpu’s in there. But I have proven I can get it working in both VMs and Docker. This is a big learning curve. Your comments replies are helping.

Thank you

Chris

Edited by SmokeyColes

  • 4 months later...

I'm attempting to use GPU passthrough on some VMs. I posted a year or so ago here how updating from 7.0.1 broke my working VMs. Launching them woudl cause unraid to hang up and I would need to reboot.

I decided to update to 7.2.4 today to see if the issue remained and my VMs are broken again. I have attached diagnostics this time. Gemini identified this as an issue in my syslog-previous when starting the VM.

Mar 22 15:08:47 Tower usb_manager: Info: rc.usb_manager vm_action batocera prepare begin -

Mar 22 15:08:47 Tower acpid: input device has been disconnected, fd 11

Mar 22 15:08:47 Tower kernel: i915 0000:00:02.0: MDEV: Unregistering

Mar 22 15:08:47 Tower kernel: BUG: kernel NULL pointer dereference, address: 0000000000000090

Mar 22 15:08:47 Tower kernel: #PF: supervisor write access in kernel mode

Mar 22 15:08:47 Tower kernel: #PF: error_code(0x0002) - not-present page

Mar 22 15:08:47 Tower kernel: PGD 0 P4D 0

Mar 22 15:08:47 Tower kernel: Oops: Oops: 0002 [#1] PREEMPT SMP PTI

Mar 22 15:08:47 Tower kernel: CPU: 5 UID: 0 PID: 15809 Comm: rpc-libvirtd Tainted: P U O 6.12.54-Unraid #1

Mar 22 15:08:47 Tower kernel: Tainted: [P]=PROPRIETARY_MODULE, [U]=USER, [O]=OOT_MODULE

Mar 22 15:08:47 Tower kernel: Hardware name: HP HP EliteDesk 800 G4 SFF/83E1, BIOS Q01 Ver. 02.30.00 12/30/2024

Anyone know how to fix this?

tower-diagnostics-20260322-1531.zip

  • 1 month later...

Any have idea on the issue above. Now on Unraid 7.2.5. I am still unable to start my VMs using integrated GPU

Any ideas on the above? I updated to Unraid 7.2.5. I still can not start VMs using intel iGPU

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

Account

Navigation

Search

Search

Configure browser push notifications

Chrome (Android)
  1. Tap the lock icon next to the address bar.
  2. Tap Permissions → Notifications.
  3. Adjust your preference.
Chrome (Desktop)
  1. Click the padlock icon in the address bar.
  2. Select Site settings.
  3. Find Notifications and adjust your preference.