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.

Intel Arc support

Featured Replies

1 hour ago, FadedSun said:

Message Sent with link to files.

Hello, can you please share also for me please

  • 1 month later...
  • Replies 402
  • Views 145.1k
  • Created
  • Last Reply

Top Posters In This Topic

Most Popular Posts

  • Kernel 6.0.1 is stable since today, chances are high that the next Unraid release includes it.

  • DeadDevil6210
    DeadDevil6210

    Oke so this weekend i had some free time to spend and checked out support on unraid 6.12 rc2. I started of fresh by reinstalling my flash drive and putting back a copy of the config folder on to it. 

  • What are your vm settings for the guys getting passthrough working? Any special steps taken? I keep getting code 43 after installing the latest beta drivers. 

Posted Images

On 1/14/2025 at 8:18 AM, FadedSun said:

I'll be honest, I'm not sure. I've heard that ZFS support lagging behind on kernels is likely why Unraid 7 stuck with 6.6.

I'm just passing along warnings I've read while researching getting my iGPU working, I don't want to break your unraid with my build.

 

Maybe someone else can chime in with more info.

 

I can send you the files to try if you'd like, but I'd make extra sure you have backups; particularly of the cache drive.

When I had a failed kernel build I just had to put the old kernel back to fix my Unraid, but I can't promise all failure modes will be that simple.

I just picked up a B570, do you mind messaging me the same files? I have a test server that I can use without the danger of corrupting my main Unraid machine.

Thanks in advance!

I tried kernel by @FadedSun, it's working, but I have some issue with my lan driver
My mobo using some weird network card, I need to use RTL8126 PCI Drivers, but unfortunatelly by some reason after I installed new kernel, I am unable to install that driver

Any idea how this driver can be added to my system?

6 hours ago, J05u said:

I tried kernel by @FadedSun, it's working, but I have some issue with my lan driver
My mobo using some weird network card, I need to use RTL8126 PCI Drivers, but unfortunatelly by some reason after I installed new kernel, I am unable to install that driver

Any idea how this driver can be added to my system?

I found this.

https://www.reddit.com/r/ASRock/comments/1hf8r5o/comment/m2da9xa/?share_id=IGiSW9e3eiRjBDZGI56X6&utm_medium=android_app&utm_name=androidcss&utm_source=share&utm_term=1

 

The drivers for that card say

Quote

5G Ethernet LINUX driver r8126 for kernel up to 6.8

 

I'm not sure if that post is implying you can use those drivers in 6.12 or if you should use one of the other drivers.

Hello there, just upgraded to Unraid 7.0, bought Arc 310, seems to HW transcode fine with Plex, but i'm not able to keep any stable GPU statistics in dashboard for GPU 2 (GPU1 being integrated igpu)

image.png.31abe3b0d6954df595354876625f3a97.png 

GPU stats shows "Vendor command returned unparseable data"
image.png.c18036c0e122d7290b5385bfc8bb741b.png
where igpu top seems OK image.thumb.png.524fe2ff8f162b436437ea14a2783081.png

 

Plex extra parameters are: --device=/dev/dri/renderD129 which matches
/dev/dri/by-path# intel_gpu_top -L
card1                    Intel Dg2 (Gen12)                 pci:vendor=8086,device=56A6,card=0
└─renderD129            
card0                    Intel Sandybridge (Gen6)          pci:vendor=8086,device=010A,card=0
└─renderD128

 

Any ideas why "vendor command returned unparseable data" ?

On 1/14/2025 at 2:18 PM, FadedSun said:

I'll be honest, I'm not sure. I've heard that ZFS support lagging behind on kernels is likely why Unraid 7 stuck with 6.6.

I'm just passing along warnings I've read while researching getting my iGPU working, I don't want to break your unraid with my build.

 

Maybe someone else can chime in with more info.

 

I can send you the files to try if you'd like, but I'd make extra sure you have backups; particularly of the cache drive.

When I had a failed kernel build I just had to put the old kernel back to fix my Unraid, but I can't promise all failure modes will be that simple.

Would you be so kind to share, picked up Battlemage GPU which isn't working due to kernel issues

  • 3 months later...

Hello everyone,

I've had the Sparkle Arc A310 Eco for a while now and I have to say, the fan is killing me. Even at idle it is always speeding up and down. For space reasons, the server is in my office and I can hardly stand it without headphones.

I have already read something about driver updates and that you can somehow set the fan curve via VM, but these were not Unraid environments and I am unsure how and whether this works. Does anyone have experience under Unraid how to fix the problem? I am currently on 7.0.1 because I had problems with 7.1.3.

  • Community Expert
1 hour ago, Hadda said:

Hello everyone,

I've had the Sparkle Arc A310 Eco for a while now and I have to say, the fan is killing me. Even at idle it is always speeding up and down. For space reasons, the server is in my office and I can hardly stand it without headphones.

I have already read something about driver updates and that you can somehow set the fan curve via VM, but these were not Unraid environments and I am unsure how and whether this works. Does anyone have experience under Unraid how to fix the problem? I am currently on 7.0.1 because I had problems with 7.1.3.


Best you can do is get a JST to PWM style cable and plug the gpu ran into your motherboard. It's not possible to control the gpu fan speed from unraid.

11 hours ago, MowMdown said:


Best you can do is get a JST to PWM style cable and plug the gpu ran into your motherboard. It's not possible to control the gpu fan speed from unraid.

Thank you for your answer. Since I'm not an expert, can you perhaps explain a little about the benefits of connecting the fan directly to the mainboard? And what should be done then?

Edited by Hadda

  • Community Expert

It will allow you to control the fan speed and set it to a single setting so it's not ramping up and down constantly.

  • Community Expert
54 minutes ago, MowMdown said:

It will allow you to control the fan speed and set it to a single setting so it's not ramping up and down constantly.

This is a very good solution for the A310 (of course not ideal, the card should just work properly), as likely the vast majority of use cases for the card is only transcoding. Setting the fan speed to a constant very low speed should be more than enough cooling while generating little noise and using minimal power.

If you are chewing through LLM or playing higher end games, a different card would be your best solution.

8 hours ago, MowMdown said:

It will allow you to control the fan speed and set it to a single setting so it's not ramping up and down constantly.

Thanks for the advice.

Is it perhaps also possible and sensible to replace the fan? Are there any suitable fans from Noctua? These fans are actually always very quiet.

  • Community Expert
1 hour ago, Hadda said:

Thanks for the advice.

Is it perhaps also possible and sensible to replace the fan? Are there any suitable fans from Noctua? These fans are actually always very quiet.

That would probably require more work than simply plugging the GPUs existing fan into the motherboard.

You'd have to remove the cover and the fan, then find a way to attach a small 40mm fan to it and still plug it into the motherboard for control. It would not be a 1:1 fan swap.

On 6/16/2025 at 10:29 PM, Hadda said:

Hello everyone,

I've had the Sparkle Arc A310 Eco for a while now and I have to say, the fan is killing me. Even at idle it is always speeding up and down. For space reasons, the server is in my office and I can hardly stand it without headphones.

I have already read something about driver updates and that you can somehow set the fan curve via VM, but these were not Unraid environments and I am unsure how and whether this works. Does anyone have experience under Unraid how to fix the problem? I am currently on 7.0.1 because I had problems with 7.1.3.

it works

You need to create vm with passthrough of gpu, install arc drivers, setup curve. After you just need just start this vm on boot, launch driver and curve will stay until next server reboot. But I ended up returning my a310, not worth the hassle. I still have A380 from Sparkle which I am considering to change (Fan curve is set as I told), on top of it I have pci-e fan bracket with 2 fans blowing to GPU, so far there was no reason for gpu fan to kick in.
My curve is far from being ideal, GPU literally passivly cooled until 50 degrees if I remember right

Just avoid Sparkle next time, at least 310 and 380, never heard something from 570 or 580 series

  • 3 months later...
On 12/1/2025 at 20:21, FadedSun said:

Tengo mi iGPU Ultra 7 265K funcionando con la versión Unraid 7 con un kernel personalizado 6.12.9.

Supongo que esto también funcionaría con las GPU Arc.

imagen.png.6d3e66db49acf5813eb95051f372fd55.png

 

Se transcodifica con mi contenedor Jellyfin en versión estable, pero necesita la versión beta de Emby.

imagen.png.734ef57701d4c24c6c79c28d35c530f0.png

 

imagen.png.5b9f5a9d6a79009628ac76a218dcf8fa.png

 

Utilicé esta guía https://gitlab.com/stefan99353/unraid-custom-kernel/-/blob/main/README.md

pero usó el kernel 6.12.9 y la versión Unraid 7.

Me llevó un tiempo compilar todo, pero parece que todo funciona.

 

Si alguien quiere los archivos actualizados para probar, hágamelo saber.

SIN GARANTÍA ni promesas, por supuesto, y creo que no podrás usarlo si usas ZFS.

Ahora que está compilado, es bastante fácil probarlo: simplemente coloque los archivos en su unidad USB de arranque reemplazando lo que está allí y arranque.

Primero guarde los archivos originales y podrá volver a la versión anterior si hay algún problema.

Hello good, could you pass me the files?

I want to try Intel arc pro b50

Thank you

  • 2 weeks later...

I tried passing through my arc pro b50 to a windows VM and I am getting error 43... which makes me believe we have to wait for that kernel update?! If not any suggestions?

  • 1 month later...

B50 Pro here.... Its an interesting card but its also a little disappointing still as far as unraid support.

I can force it off of the i915 and onto the XE driver just fine with unraid 7.2.0. I can never get any video out on it though. It does see the rebar and does detect it, just not initializing the device correctly. I cant get it to pass through to a VM. Maybe until everything is working under Unraid native I have to exempt the pcie device from the OS trying init it thus leaving it as a dummy pcie device for a VM to do the init for instead.


some nice lspci output though:

45:00.0 VGA compatible controller: Intel Corporation Battlemage G21 [Intel Graphics] (prog-if 00 [VGA controller])

Subsystem: Intel Corporation Device 1114

Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-

Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-

Latency: 0, Cache Line Size: 64 bytes

Interrupt: pin ? routed to IRQ 213

Region 0: Memory at 9d000000 (64-bit, non-prefetchable) [size=16M]

Region 2: Memory at 80000000 (64-bit, prefetchable) [size=256M]

Expansion ROM at 000c0000 [disabled] [size=128K]

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 (v2) Endpoint, IntMsgNum 0

DevCap: MaxPayload 256 bytes, PhantFunc 0, Latency L0s unlimited, L1 unlimited

ExtTag+ AttnBtn- AttnInd- PwrInd- RBE+ FLReset+ SlotPowerLimit 0W TEE-IO-

DevCtl: CorrErr- NonFatalErr- FatalErr- UnsupReq-

RlxdOrd+ ExtTag+ PhantFunc- AuxPwr- NoSnoop+ FLReset-

MaxPayload 256 bytes, MaxReadReq 512 bytes

DevSta: CorrErr- NonFatalErr- FatalErr- UnsupReq- AuxPwr- TransPend-

LnkCap: Port #0, Speed 2.5GT/s, Width x1, ASPM L0s L1, Exit Latency L0s <64ns, L1 <1us

ClockPM- Surprise- LLActRep- BwNot- ASPMOptComp+

LnkCtl: ASPM Disabled; RCB 64 bytes, LnkDisable- CommClk-

ExtSynch- ClockPM- AutWidDis- BWInt- AutBWInt- FltModeDis-

LnkSta: Speed 2.5GT/s, Width x1

TrErr- Train- SlotClk- DLActive- BWMgmt- ABWMgmt-

DevCap2: Completion Timeout: Range B, TimeoutDis+ NROPrPrP- LTR+

10BitTagComp+ 10BitTagReq+ OBFF Not Supported, ExtFmt+ EETLPPrefix-

EmergencyPowerReduction Not Supported, EmergencyPowerReductionInit-

FRS- TPHComp- ExtTPHComp-

AtomicOpsCap: 32bit- 64bit- 128bitCAS-

DevCtl2: Completion Timeout: 50us to 50ms, TimeoutDis-

AtomicOpsCtl: ReqEn-

IDOReq- IDOCompl- LTR- EmergencyPowerReductionReq-

10BitTagReq- OBFF Disabled, EETLPPrefixBlk-

LnkCap2: Supported Link Speeds: 2.5GT/s, Crosslink- Retimer- 2Retimers- DRS-

LnkCtl2: Target Link Speed: 2.5GT/s, EnterCompliance- SpeedDis-

Transmit Margin: Normal Operating Range, EnterModifiedCompliance- ComplianceSOS-

Compliance Preset/De-emphasis: -6dB de-emphasis, 0dB preshoot

LnkSta2: Current De-emphasis Level: -6dB, EqualizationComplete- EqualizationPhase1-

EqualizationPhase2- EqualizationPhase3- LinkEqualizationRequest-

Retimer- 2Retimers- CrosslinkRes: unsupported, FltMode-

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

Address: 00000000fee12000 Data: 0025

Masking: 00000000 Pending: 00000000

Capabilities: [d0] Power Management version 3

Flags: PMEClk- DSI- D1- D2- AuxCurrent=0mA PME(D0+,D1-,D2-,D3hot+,D3cold-)

Status: D0 NoSoftRst+ PME-Enable- DSel=0 DScale=0 PME-

Capabilities: [100 v1] Alternative Routing-ID Interpretation (ARI)

ARICap: MFVC- ACS-, Next Function: 0

ARICtl: MFVC- ACS-, Function Group: 0

Capabilities: [110 v1] Null

Capabilities: [200 v1] Address Translation Service (ATS)

ATSCap: Invalidate Queue Depth: 00

ATSCtl: Enable-, Smallest Translation Unit: 00

Capabilities: [420 v1] Physical Resizable BAR

BAR 2: current size: 256MB, supported: 256MB 512MB 1GB 2GB 4GB 8GB 16GB

Capabilities: [400 v1] Latency Tolerance Reporting

Max snoop latency: 0ns

Max no snoop latency: 0ns

Kernel driver in use: xe

Kernel modules: xe


firmware files are pretty up to date too in the 7.2 release
dmesg | grep -e xe | grep firmware

[ 14.118916] xe 0000:45:00.0: [drm] Using GuC firmware from xe/bmg_guc_70.bin version 70.49.4

[ 14.130618] xe 0000:45:00.0: [drm] Using GuC firmware from xe/bmg_guc_70.bin version 70.49.4

[ 14.145231] xe 0000:45:00.0: [drm] Using HuC firmware from xe/bmg_huc.bin version 8.2.10

[ 14.369444] xe 0000:45:00.0: [drm] Finished loading DMC firmware i915/bmg_dmc.bin (v2.6)


Interesting... it is loading a i915 blob and also complains about rebar despite it being enabled on my x399 platform...

dmesg | grep -Ei 'xe|i915' | tail -n 100

[ 0.000000] NX (Execute Disable) protection: active

[ 0.000000] efi: Remove mem55: MMIO range=[0xef800000-0xef8fffff] (1MB) from e820 map

[ 0.000000] e820: remove [mem 0xef800000-0xef8fffff] reserved

[ 0.000000] efi: Not removing mem56: MMIO range=[0xefb00000-0xefb00fff] (4KB) from e820 map

[ 0.000000] efi: Remove mem57: MMIO range=[0xefc00000-0xefc7ffff] (0MB) from e820 map

[ 0.000000] e820: remove [mem 0xefc00000-0xefc7ffff] reserved

[ 0.000000] MTRR map: 6 entries (3 fixed + 3 variable; max 20), built from 9 variable MTRRs

[ 0.000000] IOAPIC[1]: apic_id 129, version 33, address 0xefb00000, GSI 24-55

[ 0.000000] [mem 0xb7a01000-0xefafffff] available for PCI devices

[ 0.123953] ... fixed-purpose events: 0

[ 13.549908] RAPL PMU: API unit is 2^-32 Joules, 1 fixed counters, 163840 ms ovfl timer

[ 13.933822] mpt3sas_cm0: handle(0xe) sas_address(0x4433221102000000) port_type(0x1)

[ 14.103461] xe 0000:45:00.0: vgaarb: deactivate vga console

[ 14.103778] xe 0000:45:00.0: [drm] Found BATTLEMAGE (device ID e212) display version 14.01 stepping B0

[ 14.118916] xe 0000:45:00.0: [drm] Using GuC firmware from xe/bmg_guc_70.bin version 70.49.4

[ 14.130618] xe 0000:45:00.0: [drm] Using GuC firmware from xe/bmg_guc_70.bin version 70.49.4

[ 14.145231] xe 0000:45:00.0: [drm] Using HuC firmware from xe/bmg_huc.bin version 8.2.10

[ 14.174013] xe 0000:45:00.0: vgaarb: VGA decodes changed: olddecodes=io+mem,decodes=none:owns=io+mem

[ 14.175393] xe 0000:45:00.0: [drm] Attempting to resize bar from 256MiB -> 16384MiB

[ 14.175408] xe 0000:45:00.0: BAR 2 [mem 0x80000000-0x8fffffff 64bit pref]: releasing

[ 14.175484] pcieport 0000:44:01.0: bridge window [mem 0x00000000 64bit pref] to [bus 45] requires relaxed alignment rules

[ 14.175512] pcieport 0000:43:00.0: bridge window [mem 0x00000000 64bit pref] to [bus 44-46] requires relaxed alignment rules

[ 14.175566] xe 0000:45:00.0: BAR 2 [mem size 0x400000000 64bit pref]: can't assign; no space

[ 14.175573] xe 0000:45:00.0: BAR 2 [mem size 0x400000000 64bit pref]: failed to assign

[ 14.175689] xe 0000:45:00.0: [drm] Failed to resize BAR2 to 16384M (-ENOSPC). Consider enabling 'Resizable BAR' support in your BIOS

[ 14.175702] xe 0000:45:00.0: BAR 2 [mem 0x80000000-0x8fffffff 64bit pref]: assigned

[ 14.175755] xe 0000:45:00.0: [drm] VISIBLE VRAM: 0x0000000080000000, 0x0000000010000000

[ 14.176116] xe 0000:45:00.0: [drm] Small BAR device

[ 14.176122] xe 0000:45:00.0: [drm] VRAM[0, 0]: Actual physical size 0x0000000400000000, usable size exclude stolen 0x00000003fb000000, CPU accessible size 0x0000000010000000

[ 14.176135] xe 0000:45:00.0: [drm] VRAM[0, 0]: DPA range: [0x0000000000000000-400000000], io range: [0x0000000080000000-90000000]

[ 14.176146] xe 0000:45:00.0: [drm] VRAM: 0x0000000400000000 is larger than resource 0x0000000010000000

[ 14.176154] xe 0000:45:00.0: [drm] Total VRAM: 0x0000000080000000, 0x0000000400000000

[ 14.176161] xe 0000:45:00.0: [drm] Available VRAM: 0x0000000080000000, 0x00000003fb000000

[ 14.354298] xe 0000:45:00.0: [drm] ccs1 fused off

[ 14.354308] xe 0000:45:00.0: [drm] ccs2 fused off

[ 14.354314] xe 0000:45:00.0: [drm] ccs3 fused off

[ 14.369444] xe 0000:45:00.0: [drm] Finished loading DMC firmware i915/bmg_dmc.bin (v2.6)

[ 14.387307] xe 0000:45:00.0: [drm] vcs1 fused off

[ 14.387315] xe 0000:45:00.0: [drm] vcs3 fused off

[ 14.387321] xe 0000:45:00.0: [drm] vcs4 fused off

[ 14.387326] xe 0000:45:00.0: [drm] vcs5 fused off

[ 14.387331] xe 0000:45:00.0: [drm] vcs6 fused off

[ 14.387338] xe 0000:45:00.0: [drm] vcs7 fused off

[ 14.387344] xe 0000:45:00.0: [drm] vecs2 fused off

[ 14.387350] xe 0000:45:00.0: [drm] vecs3 fused off

[ 14.387359] xe 0000:45:00.0: [drm] gsccs disabled due to lack of FW

[ 14.450622] [drm] Initialized xe 1.1.0 for 0000:45:00.0 on minor 0

[ 14.490056] xe 0000:45:00.0: [drm] Cannot find any crtc or sizes

[ 14.570040] xe 0000:45:00.0: [drm] Cannot find any crtc or sizes

[ 14.610052] xe 0000:45:00.0: [drm] Cannot find any crtc or sizes

Edited by Athirne
added firmware grep results.

  • Community Expert

Needs kernel 6.14, unraid is on 6.12. Once unraid moves to the next LTS kernel it should have everything working with maybe the exception of the SR-IOV stuff.

It would be really awesome if we could get an Unraid next release with the current LTS kernel in the next few weeks, so that we can start tinkering around with the b50 and b60. Any place where we can ask (in a friendly manner) if that would be possible? Or do we have to switch to a custom kernel in the meantime? @thor2002ro has released some in the past, just none for 7.20 (yet)... Is anyone willing to jump the ship and use his last release (7.1.2-7.1.4 support) to test it?

  • Community Expert
39 minutes ago, styx.techno said:

It would be really awesome if we could get an Unraid next release with the current LTS kernel in the next few weeks,

Unraid 7.2 IS using the current LTS Linux kernel, so you would need a custom kernel if you cannot wait for the next LTS kernel (which as far as I know currently has no predicted date).

1 hour ago, itimpi said:

Unraid 7.2 IS using the current LTS Linux kernel, so you would need a custom kernel if you cannot wait for the next LTS kernel (which as far as I know currently has no predicted date).

My bad, misread a chart with the current overview of the kernel. custom it is then...

2 hours ago, styx.techno said:

It would be really awesome if we could get an Unraid next release with the current LTS kernel in the next few weeks, so that we can start tinkering around with the b50 and b60. Any place where we can ask (in a friendly manner) if that would be possible? Or do we have to switch to a custom kernel in the meantime? @thor2002ro has released some in the past, just none for 7.20 (yet)... Is anyone willing to jump the ship and use his last release (7.1.2-7.1.4 support) to test it?

You can run 6.17 on 7.1.x releases of Unraid through self compile of the kernel with some patch work (as pointed out by you Styx). I have done this before with this GPU without sucess.


I think I might be having another issue since even under that kernel under Unraid I am still getting Rebar allocation errors even though Rebar + 4G is enabled on my x399 system. I am not sure if another component or something is missing hence the experimentation I did under 7.2 to see if all of those updated files would fix something (updated MESA, firmware files, something else?). No surprise 6.17 wont boot with 7.2 without plenty of patch rework.
I should try USB booting the system using arch or something with the latest bits to see if that fixes it. Its either a incompatibility with something in my threadripper based server (it has been a dependable workhorse till now) or maybe I need to just throw it in an eGPU enclosure, run Windows, and let the Intel driver install any firmware updates to the card. Since the b50 I have was a pre-order it wouldn't shock me if there weren't plenty of onboard firmware bugs that have gotten sorted out in the months since the card was produced. (remember it often takes 1-2 months of card production before you get it thanks to launch inventory build up and freight shipping)
I know the Alchemist series needed firmware updates almost out of the box to even POST in a lot of systems.

SR-IOV should be in a somewhat functional state in 6.18 (either this or next will be LTS since the LF does these release semi-regularly required by projects like Ubuntu for their LTS release set for this spring), but I wouldn't be shocked if it wasn't stable till 6.19 or shortly after. Lvl1 techs has a pretty good summery about the progress for Alchemist and Battle mage enablement is just a little behind. Intel needs to stop firing the engineers working on making their stuff work. Its bad enough their fab issues have resulted in their engineers getting early silicon to start working on with half the lead time they used to have.

19 minutes ago, Athirne said:

You can run 6.17 on 7.1.x releases of Unraid through self compile of the kernel with some patch work (as pointed out by you Styx). I have done this before with this GPU without sucess.


I think I might be having another issue since even under that kernel under Unraid I am still getting Rebar allocation errors even though Rebar + 4G is enabled on my x399 system. I am not sure if another component or something is missing hence the experimentation I did under 7.2 to see if all of those updated files would fix something (updated MESA, firmware files, something else?). No surprise 6.17 wont boot with 7.2 without plenty of patch rework.
I should try USB booting the system using arch or something with the latest bits to see if that fixes it. Its either a incompatibility with something in my threadripper based server (it has been a dependable workhorse till now) or maybe I need to just throw it in an eGPU enclosure, run Windows, and let the Intel driver install any firmware updates to the card. Since the b50 I have was a pre-order it wouldn't shock me if there weren't plenty of onboard firmware bugs that have gotten sorted out in the months since the card was produced. (remember it often takes 1-2 months of card production before you get it thanks to launch inventory build up and freight shipping)
I know the Alchemist series needed firmware updates almost out of the box to even POST in a lot of systems.

SR-IOV should be in a somewhat functional state in 6.18 (either this or next will be LTS since the LF does these release semi-regularly required by projects like Ubuntu for their LTS release set for this spring), but I wouldn't be shocked if it wasn't stable till 6.19 or shortly after. Lvl1 techs has a pretty good summery about the progress for Alchemist and Battle mage enablement is just a little behind. Intel needs to stop firing the engineers working on making their stuff work. Its bad enough their fab issues have resulted in their engineers getting early silicon to start working on with half the lead time they used to have.

The firmware issue is a good hint though! I will throw it in my windows machine later to do that. Thank you!

The proxmox guys have done it already and made the b50 run in a modified stable kernel, maybe we will see some bugs/workarounds from their side that we wight encounter later on, honest recommendation to keep an eye out for that. I will dig into running Unraid with a custom kernel, any resources or links that are helpful?

  • 4 weeks later...
On 10/31/2025 at 1:38 AM, Athirne said:

B50 Pro here.... Its an interesting card but its also a little disappointing still as far as unraid support.

I can force it off of the i915 and onto the XE driver just fine with unraid 7.2.0. I can never get any video out on it though. It does see the rebar and does detect it, just not initializing the device correctly. I cant get it to pass through to a VM. Maybe until everything is working under Unraid native I have to exempt the pcie device from the OS trying init it thus leaving it as a dummy pcie device for a VM to do the init for instead.


some nice lspci output though:

45:00.0 VGA compatible controller: Intel Corporation Battlemage G21 [Intel Graphics] (prog-if 00 [VGA controller])

Subsystem: Intel Corporation Device 1114

Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-

Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-

Latency: 0, Cache Line Size: 64 bytes

Interrupt: pin ? routed to IRQ 213

Region 0: Memory at 9d000000 (64-bit, non-prefetchable) [size=16M]

Region 2: Memory at 80000000 (64-bit, prefetchable) [size=256M]

Expansion ROM at 000c0000 [disabled] [size=128K]

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 (v2) Endpoint, IntMsgNum 0

DevCap: MaxPayload 256 bytes, PhantFunc 0, Latency L0s unlimited, L1 unlimited

ExtTag+ AttnBtn- AttnInd- PwrInd- RBE+ FLReset+ SlotPowerLimit 0W TEE-IO-

DevCtl: CorrErr- NonFatalErr- FatalErr- UnsupReq-

RlxdOrd+ ExtTag+ PhantFunc- AuxPwr- NoSnoop+ FLReset-

MaxPayload 256 bytes, MaxReadReq 512 bytes

DevSta: CorrErr- NonFatalErr- FatalErr- UnsupReq- AuxPwr- TransPend-

LnkCap: Port #0, Speed 2.5GT/s, Width x1, ASPM L0s L1, Exit Latency L0s <64ns, L1 <1us

ClockPM- Surprise- LLActRep- BwNot- ASPMOptComp+

LnkCtl: ASPM Disabled; RCB 64 bytes, LnkDisable- CommClk-

ExtSynch- ClockPM- AutWidDis- BWInt- AutBWInt- FltModeDis-

LnkSta: Speed 2.5GT/s, Width x1

TrErr- Train- SlotClk- DLActive- BWMgmt- ABWMgmt-

DevCap2: Completion Timeout: Range B, TimeoutDis+ NROPrPrP- LTR+

10BitTagComp+ 10BitTagReq+ OBFF Not Supported, ExtFmt+ EETLPPrefix-

EmergencyPowerReduction Not Supported, EmergencyPowerReductionInit-

FRS- TPHComp- ExtTPHComp-

AtomicOpsCap: 32bit- 64bit- 128bitCAS-

DevCtl2: Completion Timeout: 50us to 50ms, TimeoutDis-

AtomicOpsCtl: ReqEn-

IDOReq- IDOCompl- LTR- EmergencyPowerReductionReq-

10BitTagReq- OBFF Disabled, EETLPPrefixBlk-

LnkCap2: Supported Link Speeds: 2.5GT/s, Crosslink- Retimer- 2Retimers- DRS-

LnkCtl2: Target Link Speed: 2.5GT/s, EnterCompliance- SpeedDis-

Transmit Margin: Normal Operating Range, EnterModifiedCompliance- ComplianceSOS-

Compliance Preset/De-emphasis: -6dB de-emphasis, 0dB preshoot

LnkSta2: Current De-emphasis Level: -6dB, EqualizationComplete- EqualizationPhase1-

EqualizationPhase2- EqualizationPhase3- LinkEqualizationRequest-

Retimer- 2Retimers- CrosslinkRes: unsupported, FltMode-

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

Address: 00000000fee12000 Data: 0025

Masking: 00000000 Pending: 00000000

Capabilities: [d0] Power Management version 3

Flags: PMEClk- DSI- D1- D2- AuxCurrent=0mA PME(D0+,D1-,D2-,D3hot+,D3cold-)

Status: D0 NoSoftRst+ PME-Enable- DSel=0 DScale=0 PME-

Capabilities: [100 v1] Alternative Routing-ID Interpretation (ARI)

ARICap: MFVC- ACS-, Next Function: 0

ARICtl: MFVC- ACS-, Function Group: 0

Capabilities: [110 v1] Null

Capabilities: [200 v1] Address Translation Service (ATS)

ATSCap: Invalidate Queue Depth: 00

ATSCtl: Enable-, Smallest Translation Unit: 00

Capabilities: [420 v1] Physical Resizable BAR

BAR 2: current size: 256MB, supported: 256MB 512MB 1GB 2GB 4GB 8GB 16GB

Capabilities: [400 v1] Latency Tolerance Reporting

Max snoop latency: 0ns

Max no snoop latency: 0ns

Kernel driver in use: xe

Kernel modules: xe


firmware files are pretty up to date too in the 7.2 release
dmesg | grep -e xe | grep firmware

[ 14.118916] xe 0000:45:00.0: [drm] Using GuC firmware from xe/bmg_guc_70.bin version 70.49.4

[ 14.130618] xe 0000:45:00.0: [drm] Using GuC firmware from xe/bmg_guc_70.bin version 70.49.4

[ 14.145231] xe 0000:45:00.0: [drm] Using HuC firmware from xe/bmg_huc.bin version 8.2.10

[ 14.369444] xe 0000:45:00.0: [drm] Finished loading DMC firmware i915/bmg_dmc.bin (v2.6)


Interesting... it is loading a i915 blob and also complains about rebar despite it being enabled on my x399 platform...

dmesg | grep -Ei 'xe|i915' | tail -n 100

[ 0.000000] NX (Execute Disable) protection: active

[ 0.000000] efi: Remove mem55: MMIO range=[0xef800000-0xef8fffff] (1MB) from e820 map

[ 0.000000] e820: remove [mem 0xef800000-0xef8fffff] reserved

[ 0.000000] efi: Not removing mem56: MMIO range=[0xefb00000-0xefb00fff] (4KB) from e820 map

[ 0.000000] efi: Remove mem57: MMIO range=[0xefc00000-0xefc7ffff] (0MB) from e820 map

[ 0.000000] e820: remove [mem 0xefc00000-0xefc7ffff] reserved

[ 0.000000] MTRR map: 6 entries (3 fixed + 3 variable; max 20), built from 9 variable MTRRs

[ 0.000000] IOAPIC[1]: apic_id 129, version 33, address 0xefb00000, GSI 24-55

[ 0.000000] [mem 0xb7a01000-0xefafffff] available for PCI devices

[ 0.123953] ... fixed-purpose events: 0

[ 13.549908] RAPL PMU: API unit is 2^-32 Joules, 1 fixed counters, 163840 ms ovfl timer

[ 13.933822] mpt3sas_cm0: handle(0xe) sas_address(0x4433221102000000) port_type(0x1)

[ 14.103461] xe 0000:45:00.0: vgaarb: deactivate vga console

[ 14.103778] xe 0000:45:00.0: [drm] Found BATTLEMAGE (device ID e212) display version 14.01 stepping B0

[ 14.118916] xe 0000:45:00.0: [drm] Using GuC firmware from xe/bmg_guc_70.bin version 70.49.4

[ 14.130618] xe 0000:45:00.0: [drm] Using GuC firmware from xe/bmg_guc_70.bin version 70.49.4

[ 14.145231] xe 0000:45:00.0: [drm] Using HuC firmware from xe/bmg_huc.bin version 8.2.10

[ 14.174013] xe 0000:45:00.0: vgaarb: VGA decodes changed: olddecodes=io+mem,decodes=none:owns=io+mem

[ 14.175393] xe 0000:45:00.0: [drm] Attempting to resize bar from 256MiB -> 16384MiB

[ 14.175408] xe 0000:45:00.0: BAR 2 [mem 0x80000000-0x8fffffff 64bit pref]: releasing

[ 14.175484] pcieport 0000:44:01.0: bridge window [mem 0x00000000 64bit pref] to [bus 45] requires relaxed alignment rules

[ 14.175512] pcieport 0000:43:00.0: bridge window [mem 0x00000000 64bit pref] to [bus 44-46] requires relaxed alignment rules

[ 14.175566] xe 0000:45:00.0: BAR 2 [mem size 0x400000000 64bit pref]: can't assign; no space

[ 14.175573] xe 0000:45:00.0: BAR 2 [mem size 0x400000000 64bit pref]: failed to assign

[ 14.175689] xe 0000:45:00.0: [drm] Failed to resize BAR2 to 16384M (-ENOSPC). Consider enabling 'Resizable BAR' support in your BIOS

[ 14.175702] xe 0000:45:00.0: BAR 2 [mem 0x80000000-0x8fffffff 64bit pref]: assigned

[ 14.175755] xe 0000:45:00.0: [drm] VISIBLE VRAM: 0x0000000080000000, 0x0000000010000000

[ 14.176116] xe 0000:45:00.0: [drm] Small BAR device

[ 14.176122] xe 0000:45:00.0: [drm] VRAM[0, 0]: Actual physical size 0x0000000400000000, usable size exclude stolen 0x00000003fb000000, CPU accessible size 0x0000000010000000

[ 14.176135] xe 0000:45:00.0: [drm] VRAM[0, 0]: DPA range: [0x0000000000000000-400000000], io range: [0x0000000080000000-90000000]

[ 14.176146] xe 0000:45:00.0: [drm] VRAM: 0x0000000400000000 is larger than resource 0x0000000010000000

[ 14.176154] xe 0000:45:00.0: [drm] Total VRAM: 0x0000000080000000, 0x0000000400000000

[ 14.176161] xe 0000:45:00.0: [drm] Available VRAM: 0x0000000080000000, 0x00000003fb000000

[ 14.354298] xe 0000:45:00.0: [drm] ccs1 fused off

[ 14.354308] xe 0000:45:00.0: [drm] ccs2 fused off

[ 14.354314] xe 0000:45:00.0: [drm] ccs3 fused off

[ 14.369444] xe 0000:45:00.0: [drm] Finished loading DMC firmware i915/bmg_dmc.bin (v2.6)

[ 14.387307] xe 0000:45:00.0: [drm] vcs1 fused off

[ 14.387315] xe 0000:45:00.0: [drm] vcs3 fused off

[ 14.387321] xe 0000:45:00.0: [drm] vcs4 fused off

[ 14.387326] xe 0000:45:00.0: [drm] vcs5 fused off

[ 14.387331] xe 0000:45:00.0: [drm] vcs6 fused off

[ 14.387338] xe 0000:45:00.0: [drm] vcs7 fused off

[ 14.387344] xe 0000:45:00.0: [drm] vecs2 fused off

[ 14.387350] xe 0000:45:00.0: [drm] vecs3 fused off

[ 14.387359] xe 0000:45:00.0: [drm] gsccs disabled due to lack of FW

[ 14.450622] [drm] Initialized xe 1.1.0 for 0000:45:00.0 on minor 0

[ 14.490056] xe 0000:45:00.0: [drm] Cannot find any crtc or sizes

[ 14.570040] xe 0000:45:00.0: [drm] Cannot find any crtc or sizes

[ 14.610052] xe 0000:45:00.0: [drm] Cannot find any crtc or sizes

You dont install any drivers for VM passtrough, I run it Right now with win11 even with an Igpu VF for testing with lossles scaling (it sucks). I run it headless with Apollo//Sunshine as a streaming VM and it just runs Fine Like Doom 4 80ish fps and Resident evil 8 at 50ish fps 1440p

Unraid 7.2

14600k

MSI z670

  • 1 month later...

So a long time ago I set my a380 with thor's kernel 6.5.10 for jellyfin transcoding purposes, I am currently on version 6.12.4 Where does that leave me for future updates, I am kinda afraid to touch it currently? My plan is to wait for when unraid moves to LTS 6.18, but what exactly happens when updating then? I assume I have to move to a stock unraid kernel, but I am not sure where to find the one for unraid 6.12.4(I didn't save it unfortunately), Does this sound sensible? Some input would be really helpful

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.