22 minutes ago, murakume said:

Plugin seems to install fine, but when my gpu is selected it causes the fan to ramp up and down. When this is happening it also shows the reported memory speed going up and down, but it's hard to just get a screenshot of that. NVTOP doesn't cause the same behavior and behavior existed with or without NVTOP being installed. Running current open source drivers on a 5060ti. The card does have a zero RPM idle out of the box and I haven't tried messing with fan curves manually.

Thanks for the help and lemme know if there's more information needed

Does it do it when you just run nvidia-smi -q -x as that is the command the GPUstat is running.

4 minutes ago, SimonF said:

Does it do it when you just run nvidia-smi -q -x as that is the command the GPUstat is running.

Yes, fans briefly spin up then go back to idle. This behavior is also present when first selecting the gpu in the options page, I just assumed it was initial polling to figure out values to create the UI with or something like that.

Edit: just in case it makes a difference: after intitially selecting the gpu it only spins up once and thats it. It isn't until I'm on the dashboard and the UI is working that the fans and memory speed start oscillating.

Edited February 22Feb 22 by murakume

Hello, i just installed it, not seeing anything, any info on a intel igpu

cat /tmp/gpudata0000\:00\:02.0

{"period":{"duration":1028.53972,"unit":"ms"},"frequency":{"requested":0,"actual":0,"unit":"MHz"},"interrupts":{"count":0,"unit":"irq\/s"},"rc6":{"value":99.999984,"unit":"%"},"power":{"GPU":0,"Package":10.310478,"unit":"W"},"imc-bandwidth":{"reads":513.225966,"writes":124.511836,"unit":"MiB\/s"},"engines":{"Render\/3D":{"busy":0,"sema":0,"wait":0,"unit":"%"},"Blitter":{"busy":0,"sema":0,"wait":0,"unit":"%"},"Video":{"busy":0,"sema":0,"wait":0,"unit":"%"},"VideoEnhance":{"busy":0,"sema":0,"wait":0,"unit":"%"}},"clients":[]}

cpu is intel 8600T

!!EDIT: i unselected it and select it again, working ok... thnx

Edited March 2Mar 2 by kicsrules

On 2/22/2026 at 9:47 AM, murakume said:

Yes, fans briefly spin up then go back to idle. This behavior is also present when first selecting the gpu in the options page, I just assumed it was initial polling to figure out values to create the UI with or something like that.

Edit: just in case it makes a difference: after intitially selecting the gpu it only spins up once and thats it. It isn't until I'm on the dashboard and the UI is working that the fans and memory speed start oscillating.

Might be bad etiquette to quote myself, however for closures sake: enabling persistence mode fixed fan and clock speeds jumping around with this plugin active. Noticed it was reporting P0, but when NVTOP was open at the same time it would flip to P8 and settle down until NVTOP was closed again. Honestly not sure if it was a me thing (misconfiguration or whatever else), probably was, but it is fixed now.

Any info when Battlemage cards will be supported? I can see applications are using the card but no stats. If i use NVTOP inside unraid terminal i can see stats of the card.

Any chance this plugin will be updated for Unraid 7.3 Beta's to test new 6.18 kernel?

Thanks

On 3/25/2026 at 1:58 PM, T_Matz said:

Any info when Battlemage cards will be supported? I can see applications are using the card but no stats. If i use NVTOP inside unraid terminal i can see stats of the card.

Agreed, Just installed a B70 Pro. Waiting for an update. Thanks!

Agreed, Just installed B70 Pro. Waiting for an update. Thanks!

Here is a breakdown of my hardware and what I am seeing:

• Unraid 7.2.2

• GPU: Intel Arc Pro B70 (Battlemage G31)

• Motherboard: ASUS TUF Gaming Z690-Plus WIFI

• PCI ID: 03:00.0

• Dashboard Error: "Vendor command returned no data."

Seems it is loading i915 drivers vs XE?

unraid-server-diagnostics-20260419-1636.zip

Edited April 19Apr 19 by jtaz16

8 hours ago, jtaz16 said:

Agreed, Just installed B70 Pro. Waiting for an update. Thanks!

Here is a breakdown of my hardware and what I am seeing:

• Unraid 7.2.2

• GPU: Intel Arc Pro B70 (Battlemage G31)

• Motherboard: ASUS TUF Gaming Z690-Plus WIFI

• PCI ID: 03:00.0

• Dashboard Error: "Vendor command returned no data."

Seems it is loading i915 drivers vs XE?

unraid-server-diagnostics-20260419-1636.zip

You need to set the card to be ignored by the i915 driver and enabled in the XE on 7.2.x =! disables

@T_Matz @Dazog @jtaz16

I will be looking to see if https://crates.io/crates/qmassa

provide enought output as a json file to import into GPUs stats. NVTop is interactive and does not provide a json output but qmassa does. I hope I may have a test version before the end of April subject to other workloads.

8 hours ago, SimonF said:

You need to set the card to be ignored by the i915 driver and enabled in the XE on 7.2.x =! disables

Thanks for the info. I have those drivers set now, waiting 17hrs after my party check. My cache drives reached read only status and locked up. Had to perform new config...

On 4/20/2026 at 12:52 AM, SimonF said:

You need to set the card to be ignored by the i915 driver and enabled in the XE on 7.2.x =! disables

I adjusted my xe drivers. Seems it is still not init on the B70 Pro..

Performed restart. Also now running on 7.2.4

I do also have a igpu. Do i need to specify that igpu for i915 as well?

Edited April 21Apr 21 by jtaz16

2 hours ago, jtaz16 said:

I adjusted my xe drivers. Seems it is still not init on the B70 Pro..

Performed restart. Also now running on 7.2.4

I do also have a igpu. Do i need to specify that igpu for i915 as well?

Go into gpu stats settings, remove the GPUs, apply then add the back in and apply to see if that fixes.

Also run lspci -k to check the gpu is using xe.

11 minutes ago, SimonF said:

Go into gpu stats settings, remove the GPUs, apply then add the back in and apply to see if that fixes.

Also run lspci -k to check the gpu is using xe.

Removed, saved and added the GPUs back. No change.


Here is lspci:

Server:~# lspci -v -s 03:00.0

03:00.0 VGA compatible controller: Intel Corporation Battlemage G31 [Intel Graphics] (prog-if 00 [VGA controller])

Subsystem: Intel Corporation Device 1701

Flags: bus master, fast devsel, latency 0, IOMMU group 23

Memory at 7004000000 (64-bit, prefetchable) [size=16M]

Memory at 6000000000 (64-bit, prefetchable) [size=32G]

Expansion ROM at 84600000 [disabled] [size=2M]

Capabilities: [40] Vendor Specific Information: Intel Capabilities v1

CapA: Peg60Dis- Peg12Dis- Peg11Dis- Peg10Dis- PeLWUDis- DmiWidth=x4

EccDis- ForceEccEn- VTdDis- DmiG2Dis- PegG2Dis- DDRMaxSize=Unlimited

1NDis- CDDis- DDPCDis- X2APICEn- PDCDis- IGDis- CDID=0 CRID=0

DDROCCAP+ OCEn- DDRWrtVrefEn+ DDR3LEn+

CapB: ImguDis- OCbySSKUCap- OCbySSKUEn- SMTCap- CacheSzCap 0x0

SoftBinCap- DDR3MaxFreqWithRef100=Disabled PegG3Dis-

PkgTyp- AddGfxEn- AddGfxCap- PegX16Dis- DmiG3Dis- GmmDis-

DDR3MaxFreq=2932MHz LPDDR3En-

Capabilities: [70] Express Endpoint, IntMsgNum 0

Capabilities: [ac] MSI: Enable- Count=1/1 Maskable+ 64bit+

Capabilities: [d0] Power Management version 3

Capabilities: [100] Alternative Routing-ID Interpretation (ARI)

Capabilities: [110] Null

Capabilities: [200] Address Translation Service (ATS)

Capabilities: [420] Physical Resizable BAR

Capabilities: [220] Virtual Resizable BAR

Capabilities: [320] Single Root I/O Virtualization (SR-IOV)

Capabilities: [400] Latency Tolerance Reporting

lspci -k:
03:00.0 VGA compatible controller: Intel Corporation Battlemage G31 [Intel Graphics]

Subsystem: Intel Corporation Device 1701

2 hours ago, jtaz16 said:

Removed, saved and added the GPUs back. No change.


Here is lspci:

Server:~# lspci -v -s 03:00.0

03:00.0 VGA compatible controller: Intel Corporation Battlemage G31 [Intel Graphics] (prog-if 00 [VGA controller])

Subsystem: Intel Corporation Device 1701

Flags: bus master, fast devsel, latency 0, IOMMU group 23

Memory at 7004000000 (64-bit, prefetchable) [size=16M]

Memory at 6000000000 (64-bit, prefetchable) [size=32G]

Expansion ROM at 84600000 [disabled] [size=2M]

Capabilities: [40] Vendor Specific Information: Intel Capabilities v1

CapA: Peg60Dis- Peg12Dis- Peg11Dis- Peg10Dis- PeLWUDis- DmiWidth=x4

EccDis- ForceEccEn- VTdDis- DmiG2Dis- PegG2Dis- DDRMaxSize=Unlimited

1NDis- CDDis- DDPCDis- X2APICEn- PDCDis- IGDis- CDID=0 CRID=0

DDROCCAP+ OCEn- DDRWrtVrefEn+ DDR3LEn+

CapB: ImguDis- OCbySSKUCap- OCbySSKUEn- SMTCap- CacheSzCap 0x0

SoftBinCap- DDR3MaxFreqWithRef100=Disabled PegG3Dis-

PkgTyp- AddGfxEn- AddGfxCap- PegX16Dis- DmiG3Dis- GmmDis-

DDR3MaxFreq=2932MHz LPDDR3En-

Capabilities: [70] Express Endpoint, IntMsgNum 0

Capabilities: [ac] MSI: Enable- Count=1/1 Maskable+ 64bit+

Capabilities: [d0] Power Management version 3

Capabilities: [100] Alternative Routing-ID Interpretation (ARI)

Capabilities: [110] Null

Capabilities: [200] Address Translation Service (ATS)

Capabilities: [420] Physical Resizable BAR

Capabilities: [220] Virtual Resizable BAR

Capabilities: [320] Single Root I/O Virtualization (SR-IOV)

Capabilities: [400] Latency Tolerance Reporting

lspci -k:
03:00.0 VGA compatible controller: Intel Corporation Battlemage G31 [Intel Graphics]

Subsystem: Intel Corporation Device 1701

The GPU does not seem to gave any drivers.

I would expect.

04:00.0 VGA compatible controller: Intel Corporation Battlemage G21 [Arc Pro B50]

Subsystem: Intel Corporation Device 1114

Kernel driver in use: xe

Kernel modules: xe

What does dmesg | grep xe show.

9 hours ago, SimonF said:

The GPU does not seem to gave any drivers.

I would expect.

04:00.0 VGA compatible controller: Intel Corporation Battlemage G21 [Arc Pro B50]

Subsystem: Intel Corporation Device 1114

Kernel driver in use: xe

Kernel modules: xe

What does dmesg | grep xe show.

I have been trying to inject the drivers manually so there was much less information before.

~# dmesg | grep xe

[ 0.000000] Command line: BOOT_IMAGE=/bzimage initrd=/bzroot xe.force_probe=e223 i915.force_probe=!e223

[ 0.000000] NX (Execute Disable) protection: active

[ 0.000605] MTRR map: 6 entries (3 fixed + 3 variable; max 23), built from 10 variable MTRRs

[ 0.084545] Kernel command line: BOOT_IMAGE=/bzimage initrd=/bzroot xe.force_probe=e223 i915.force_probe=!e223

[ 0.178346] ... fixed-purpose events: 4

[ 0.007288] ... fixed-purpose events: 3

[ 0.360304] pci 0000:00:1f.4: BAR 4 [io 0xefa0-0xefbf]

[ 33.572134] RAPL PMU: API unit is 2^-32 Joules, 2 fixed counters, 655360 ms ovfl timer

[ 34.912059] Bluetooth: hci0: Fseq executed: 00.00.02.41

[ 542.264034] filemap_add_folio+0x69/0xe0

[ 542.264065] btrfs_readahead+0xe0/0x140

[ 542.264102] kthread+0xec/0x100

[ 545.242486] filemap_add_folio+0x69/0xe0

[ 545.242517] btrfs_readahead+0xe0/0x140

[ 545.242555] kthread+0xec/0x100

[ 556.644924] filemap_add_folio+0x69/0xe0

[ 556.644955] btrfs_readahead+0xe0/0x140

[ 556.645000] kthread+0xec/0x100

[ 557.518624] filemap_add_folio+0x69/0xe0

[ 557.518655] btrfs_readahead+0xe0/0x140

[ 557.518692] kthread+0xec/0x100

[ 561.956079] filemap_add_folio+0x69/0xe0

[ 561.956110] btrfs_readahead+0xe0/0x140

[ 561.956138] kthread+0xec/0x100

2 hours ago, jtaz16 said:

I have been trying to inject the drivers manually so there was much less information before.

~# dmesg | grep xe

[ 0.000000] Command line: BOOT_IMAGE=/bzimage initrd=/bzroot xe.force_probe=e223 i915.force_probe=!e223

[ 0.000000] NX (Execute Disable) protection: active

[ 0.000605] MTRR map: 6 entries (3 fixed + 3 variable; max 23), built from 10 variable MTRRs

[ 0.084545] Kernel command line: BOOT_IMAGE=/bzimage initrd=/bzroot xe.force_probe=e223 i915.force_probe=!e223

[ 0.178346] ... fixed-purpose events: 4

[ 0.007288] ... fixed-purpose events: 3

[ 0.360304] pci 0000:00:1f.4: BAR 4 [io 0xefa0-0xefbf]

[ 33.572134] RAPL PMU: API unit is 2^-32 Joules, 2 fixed counters, 655360 ms ovfl timer

[ 34.912059] Bluetooth: hci0: Fseq executed: 00.00.02.41

[ 542.264034] filemap_add_folio+0x69/0xe0

[ 542.264065] btrfs_readahead+0xe0/0x140

[ 542.264102] kthread+0xec/0x100

[ 545.242486] filemap_add_folio+0x69/0xe0

[ 545.242517] btrfs_readahead+0xe0/0x140

[ 545.242555] kthread+0xec/0x100

[ 556.644924] filemap_add_folio+0x69/0xe0

[ 556.644955] btrfs_readahead+0xe0/0x140

[ 556.645000] kthread+0xec/0x100

[ 557.518624] filemap_add_folio+0x69/0xe0

[ 557.518655] btrfs_readahead+0xe0/0x140

[ 557.518692] kthread+0xec/0x100

[ 561.956079] filemap_add_folio+0x69/0xe0

[ 561.956110] btrfs_readahead+0xe0/0x140

[ 561.956138] kthread+0xec/0x100

If xe is not in use, then you can do modprobe -r xe follow by modprobe xe which should unload and reload the driver. If that does not work can you post diagnostics

you may need to jump to 7.3.0rc1 to get the 6.18 kernel which has xe included.

lspci -s 07:00.0 -v
07:00.0 VGA compatible controller: Intel Corporation Battlemage G21 [Arc Pro B60] (prog-if 00 [VGA controller])
        ....
        Kernel driver in use: xe
        Kernel modules: xe

Still doesnt display correctly for me tho. Unsure how to go about passing or blocking i915/xe drivers noted about

11 minutes ago, isaw said:

you may need to jump to 7.3.0rc1 to get the 6.18 kernel which has xe included.

lspci -s 07:00.0 -v
07:00.0 VGA compatible controller: Intel Corporation Battlemage G21 [Arc Pro B60] (prog-if 00 [VGA controller])
        ....
        Kernel driver in use: xe
        Kernel modules: xe

Still doesnt display correctly for me tho. Unsure how to go about passing or blocking i915/xe drivers noted about

XE shows limited info in GPU Stats at present, at most Fans/Temps and processes. I am working on integrating qmassa which should provide some better status as Intel don't provide any useful tools and most others don't provide a JSON output that can be used.

Give release 2026.05.08 a try for XE additional stats. This is the initial release and may need tweaks going forward. Question mark is nvtop.

Spaceinvaderone was installing and running Ollama which is most of the mem usage.

I was running plex with transcode and virtio3d running glxgears

After loading the plugin I recommend do CTLR + F5 to reload the dashboard.

After the update, my fan show N/A instead of 0 (it almost never spins up). Memory usage I think is new, at least that bar.

Edited May 8May 8 by Niklas

5 minutes ago, Niklas said:

After the update, my fan show N/A instead of 0 (it almost never spins up). Memory usage I think is new, at least that bar.

Do cntr + F5 to reload the page if not already. You might get more info if you switch to the XE driver.

1 minute ago, SimonF said:

Do cntr + F5 to reload the page if not already.

I usually use dev tools (F12) and do empty cache with hard reload. Did it now and at least the memory usage bar fill went away, fan and memory usage N/A but memory usage may not be available on Arc A380 but fan is.

Edit: The fan maybe needs to turn on at least once for it to not show N/A? I'll keep my eyes on it.

Edited May 8May 8 by Niklas

Memory usage works with the xe-driver (but some containers does not like xe so back to i915, maybe in the future).

Fan shows N/A on both. Worked on i915 on old release.

Edited May 8May 8 by Niklas

3 minutes ago, Niklas said:

Memory usage works with the xe-driver (but some containers does not like xe so back to i915, maybe in the future).

Fan shows N/A on both. Worked on i915 on old release.

If you run qmassa on the cmdline does it show fan speed? But will look into it.