{"slug": "intel-arc-pro-b60-stuck-at-pcie-gen1-x1-and-21gb-vram-in-proxmox", "title": "Intel Arc Pro B60 stuck at PCIe Gen1 x1 and ~21GB VRAM in Proxmox", "summary": "Intel Arc Pro B60 graphics cards are stuck at PCIe Gen1 x1 and report about 21GB of VRAM instead of the advertised 24GB in a Proxmox system, according to a user on the Proxmox forum. The user, running an AMD EPYC 7K62 with a Supermicro H12SSL-i motherboard, found that both cards (device ID 8086:e211) are recognized by the xe driver but fail to be detected by Intel's tools, and BIOS settings like Resizable BAR and Above 4G Decoding did not resolve the link issue.", "body_md": "I need help with a couple recently purchased Intel Arc Pro B60s in my Proxmox system (EPYC 7K62, Supermicro H12SSL-i, 128GB ECC DDR4). I bought the ARKN variants from Newegg ([product page](https://www.newegg.com/arkn-8357-00128-arc-pro-b60-24gb-graphics/p/N82E16814983001)) and started noticing issues pretty much right away.\n\nMy original plan was to pass them straight through to a VM for local inference using VFIO, but none of the Intel tools I tried, clinfo and the Level Zero tools, were picking them up as GPUs. Because of that I started digging into what the cards looked like from the host side.\n\nThat’s where things started looking weird. The PCIe link negotiation for both cards only shows Gen1 x1 at 2.5GT/s. I double checked the usual BIOS settings, Resizable BAR, Above 4G Decoding, SR-IOV support, and made sure all the x16 slots are forced to **[x16]** instead of **[Auto]**. None of that changed the link behavior.\n\nI also noticed the card is being reported with about **21GB of VRAM** instead of the **24GB** listed on the product page. Funny enough, there’s a reviewer on the Newegg page saying they couldn’t get 24GB out of it either, along with this [YouTube video](https://www.youtube.com/watch?v=0clT6fWeWMI).\n\nAt this point I’m not really sure where to go next. If anyone has experience getting these cards working properly in a setup like this, I’d appreciate the advice.\n\nHere are some relevant details from what I’ve found so far:\n\n```\nroot@proxmox:~# ./gpu_debug.sh\n========== SYSTEM ==========\n\n### COMMAND:\nuname -r\n### OUTPUT:\n6.17.13-2-pve\n\n### COMMAND:\npveversion\n### OUTPUT:\npve-manager/9.1.6/71482d1833ded40a (running kernel: 6.17.13-2-pve)\n\n### COMMAND:\ncat /proc/cmdline\n### OUTPUT:\nBOOT_IMAGE=/boot/vmlinuz-6.17.13-2-pve root=/dev/mapper/pve-root ro quiet amd_iommu=on iommu=pt pcie_aspm=off pcie_acs_override=downstream,multifunction pci=realloc\n\n### COMMAND:\nlscpu | grep 'Model name'\n### OUTPUT:\nModel name:                              AMD EPYC 7K62 48-Core Processor\n\n### COMMAND:\ndmidecode -t baseboard | grep -E 'Manufacturer|Product|Version'\n### OUTPUT:\n        Manufacturer: Supermicro\n        Product Name: H12SSL-i\n        Version: 1.10\n\n========== GPU DEVICES ==========\n\n### COMMAND:\nlspci -nn | grep -E 'VGA|3D'\n### OUTPUT:\n43:00.0 VGA compatible controller [0300]: ASPEED Technology, Inc. ASPEED Graphics Family [1a03:2000] (rev 41)\n83:00.0 VGA compatible controller [0300]: Intel Corporation Battlemage G21 [Intel Graphics] [8086:e211]\nc1:00.0 VGA compatible controller [0300]: NVIDIA Corporation GB205 [GeForce RTX 5070] [10de:2f04] (rev a1)\nc4:00.0 VGA compatible controller [0300]: Intel Corporation Battlemage G21 [Intel Graphics] [8086:e211]\n\n### COMMAND:\nlspci -nnk -d 8086:e211\n### OUTPUT:\n83:00.0 VGA compatible controller [0300]: Intel Corporation Battlemage G21 [Intel Graphics] [8086:e211]\n        Subsystem: Device [207e:a023]\n        Kernel driver in use: xe\n        Kernel modules: xe\nc4:00.0 VGA compatible controller [0300]: Intel Corporation Battlemage G21 [Intel Graphics] [8086:e211]\n        Subsystem: Device [207e:a023]\n        Kernel driver in use: xe\n        Kernel modules: xe\n\n### COMMAND:\nlspci -tv\n### OUTPUT:\n-+-[0000:00]-+-00.0  Advanced Micro Devices, Inc. [AMD] Starship/Matisse Root Complex\n |           +-00.2  Advanced Micro Devices, Inc. [AMD] Starship/Matisse IOMMU\n |           +-01.0  Advanced Micro Devices, Inc. [AMD] Starship/Matisse PCIe Dummy Host Bridge\n |           +-02.0  Advanced Micro Devices, Inc. [AMD] Starship/Matisse PCIe Dummy Host Bridge\n |           +-03.0  Advanced Micro Devices, Inc. [AMD] Starship/Matisse PCIe Dummy Host Bridge\n |           +-03.3-[01]----00.0  Sandisk Corp WD Black SN770 / PC SN740 256GB / PC SN560 (DRAM-less) NVMe SSD\n |           +-03.4-[02]----00.0  Samsung Electronics Co Ltd NVMe SSD Controller S4LV008[Pascal]\n |           +-04.0  Advanced Micro Devices, Inc. [AMD] Starship/Matisse PCIe Dummy Host Bridge\n |           +-05.0  Advanced Micro Devices, Inc. [AMD] Starship/Matisse PCIe Dummy Host Bridge\n |           +-07.0  Advanced Micro Devices, Inc. [AMD] Starship/Matisse PCIe Dummy Host Bridge\n |           +-07.1-[03]--+-00.0  Advanced Micro Devices, Inc. [AMD] Starship/Matisse PCIe Dummy Function\n |           |            \\-00.2  Advanced Micro Devices, Inc. [AMD] Starship/Matisse PTDMA\n |           +-08.0  Advanced Micro Devices, Inc. [AMD] Starship/Matisse PCIe Dummy Host Bridge\n |           +-08.1-[04]--+-00.0  Advanced Micro Devices, Inc. [AMD] Starship/Matisse Reserved SPP\n |           |            +-00.2  Advanced Micro Devices, Inc. [AMD] Starship/Matisse PTDMA\n |           |            \\-00.3  Advanced Micro Devices, Inc. [AMD] Starship USB 3.0 Host Controller\n |           +-14.0  Advanced Micro Devices, Inc. [AMD] FCH SMBus Controller\n |           +-14.3  Advanced Micro Devices, Inc. [AMD] FCH LPC Bridge\n |           +-18.0  Advanced Micro Devices, Inc. [AMD] Starship Device 24; Function 0\n |           +-18.1  Advanced Micro Devices, Inc. [AMD] Starship Device 24; Function 1\n |           +-18.2  Advanced Micro Devices, Inc. [AMD] Starship Device 24; Function 2\n |           +-18.3  Advanced Micro Devices, Inc. [AMD] Starship Device 24; Function 3\n |           +-18.4  Advanced Micro Devices, Inc. [AMD] Starship Device 24; Function 4\n |           +-18.5  Advanced Micro Devices, Inc. [AMD] Starship Device 24; Function 5\n |           +-18.6  Advanced Micro Devices, Inc. [AMD] Starship Device 24; Function 6\n |           \\-18.7  Advanced Micro Devices, Inc. [AMD] Starship Device 24; Function 7\n +-[0000:40]-+-00.0  Advanced Micro Devices, Inc. [AMD] Starship/Matisse Root Complex\n |           +-00.2  Advanced Micro Devices, Inc. [AMD] Starship/Matisse IOMMU\n |           +-01.0  Advanced Micro Devices, Inc. [AMD] Starship/Matisse PCIe Dummy Host Bridge\n |           +-02.0  Advanced Micro Devices, Inc. [AMD] Starship/Matisse PCIe Dummy Host Bridge\n |           +-03.0  Advanced Micro Devices, Inc. [AMD] Starship/Matisse PCIe Dummy Host Bridge\n |           +-03.3-[41]----00.0  ASMedia Technology Inc. ASM1042A USB 3.0 Host Controller\n |           +-03.4-[42-43]----00.0-[43]----00.0  ASPEED Technology, Inc. ASPEED Graphics Family\n |           +-03.5-[44]----00.0  ASMedia Technology Inc. ASM1042A USB 3.0 Host Controller\n |           +-03.6-[45]--+-00.0  Broadcom Inc. and subsidiaries NetXtreme BCM5720 Gigabit Ethernet PCIe\n |           |            \\-00.1  Broadcom Inc. and subsidiaries NetXtreme BCM5720 Gigabit Ethernet PCIe\n |           +-04.0  Advanced Micro Devices, Inc. [AMD] Starship/Matisse PCIe Dummy Host Bridge\n |           +-05.0  Advanced Micro Devices, Inc. [AMD] Starship/Matisse PCIe Dummy Host Bridge\n |           +-07.0  Advanced Micro Devices, Inc. [AMD] Starship/Matisse PCIe Dummy Host Bridge\n |           +-07.1-[46]--+-00.0  Advanced Micro Devices, Inc. [AMD] Starship/Matisse PCIe Dummy Function\n |           |            \\-00.2  Advanced Micro Devices, Inc. [AMD] Starship/Matisse PTDMA\n |           +-08.0  Advanced Micro Devices, Inc. [AMD] Starship/Matisse PCIe Dummy Host Bridge\n |           +-08.1-[47]--+-00.0  Advanced Micro Devices, Inc. [AMD] Starship/Matisse Reserved SPP\n |           |            +-00.1  Advanced Micro Devices, Inc. [AMD] Starship/Matisse Cryptographic Coprocessor PSPCPP\n |           |            +-00.2  Advanced Micro Devices, Inc. [AMD] Starship/Matisse PTDMA\n |           |            \\-00.3  Advanced Micro Devices, Inc. [AMD] Starship USB 3.0 Host Controller\n |           +-08.2-[48]----00.0  Advanced Micro Devices, Inc. [AMD] FCH SATA Controller [AHCI mode]\n |           \\-08.3-[49]----00.0  Advanced Micro Devices, Inc. [AMD] FCH SATA Controller [AHCI mode]\n +-[0000:80]-+-00.0  Advanced Micro Devices, Inc. [AMD] Starship/Matisse Root Complex\n |           +-00.2  Advanced Micro Devices, Inc. [AMD] Starship/Matisse IOMMU\n |           +-01.0  Advanced Micro Devices, Inc. [AMD] Starship/Matisse PCIe Dummy Host Bridge\n |           +-01.1-[81-84]----00.0-[82-84]--+-01.0-[83]----00.0  Intel Corporation Battlemage G21 [Intel Graphics]\n |           |                               \\-02.0-[84]----00.0  Intel Corporation Device e2f7\n |           +-02.0  Advanced Micro Devices, Inc. [AMD] Starship/Matisse PCIe Dummy Host Bridge\n |           +-03.0  Advanced Micro Devices, Inc. [AMD] Starship/Matisse PCIe Dummy Host Bridge\n |           +-04.0  Advanced Micro Devices, Inc. [AMD] Starship/Matisse PCIe Dummy Host Bridge\n |           +-05.0  Advanced Micro Devices, Inc. [AMD] Starship/Matisse PCIe Dummy Host Bridge\n |           +-07.0  Advanced Micro Devices, Inc. [AMD] Starship/Matisse PCIe Dummy Host Bridge\n |           +-07.1-[85]--+-00.0  Advanced Micro Devices, Inc. [AMD] Starship/Matisse PCIe Dummy Function\n |           |            \\-00.2  Advanced Micro Devices, Inc. [AMD] Starship/Matisse PTDMA\n |           +-08.0  Advanced Micro Devices, Inc. [AMD] Starship/Matisse PCIe Dummy Host Bridge\n |           +-08.1-[86]--+-00.0  Advanced Micro Devices, Inc. [AMD] Starship/Matisse Reserved SPP\n |           |            \\-00.2  Advanced Micro Devices, Inc. [AMD] Starship/Matisse PTDMA\n |           +-08.2-[87]----00.0  Advanced Micro Devices, Inc. [AMD] FCH SATA Controller [AHCI mode]\n |           \\-08.3-[88]----00.0  Advanced Micro Devices, Inc. [AMD] FCH SATA Controller [AHCI mode]\n \\-[0000:c0]-+-00.0  Advanced Micro Devices, Inc. [AMD] Starship/Matisse Root Complex\n             +-00.2  Advanced Micro Devices, Inc. [AMD] Starship/Matisse IOMMU\n             +-01.0  Advanced Micro Devices, Inc. [AMD] Starship/Matisse PCIe Dummy Host Bridge\n             +-01.1-[c1]--+-00.0  NVIDIA Corporation GB205 [GeForce RTX 5070]\n             |            \\-00.1  NVIDIA Corporation Device 2f80\n             +-02.0  Advanced Micro Devices, Inc. [AMD] Starship/Matisse PCIe Dummy Host Bridge\n             +-03.0  Advanced Micro Devices, Inc. [AMD] Starship/Matisse PCIe Dummy Host Bridge\n             +-03.1-[c2-c5]----00.0-[c3-c5]--+-01.0-[c4]----00.0  Intel Corporation Battlemage G21 [Intel Graphics]\n             |                               \\-02.0-[c5]----00.0  Intel Corporation Device e2f7\n             +-04.0  Advanced Micro Devices, Inc. [AMD] Starship/Matisse PCIe Dummy Host Bridge\n             +-05.0  Advanced Micro Devices, Inc. [AMD] Starship/Matisse PCIe Dummy Host Bridge\n             +-07.0  Advanced Micro Devices, Inc. [AMD] Starship/Matisse PCIe Dummy Host Bridge\n             +-07.1-[c6]--+-00.0  Advanced Micro Devices, Inc. [AMD] Starship/Matisse PCIe Dummy Function\n             |            \\-00.2  Advanced Micro Devices, Inc. [AMD] Starship/Matisse PTDMA\n             +-08.0  Advanced Micro Devices, Inc. [AMD] Starship/Matisse PCIe Dummy Host Bridge\n             \\-08.1-[c7]--+-00.0  Advanced Micro Devices, Inc. [AMD] Starship/Matisse Reserved SPP\n                          \\-00.2  Advanced Micro Devices, Inc. [AMD] Starship/Matisse PTDMA\n\n========== PCIe LINK STATE ==========\n\n### COMMAND:\nlspci -vv -s 83:00.0 | grep -E 'LnkCap:|LnkSta:|LnkCap2:'\n### OUTPUT:\n                LnkCap: Port #0, Speed 2.5GT/s, Width x1, ASPM L0s L1, Exit Latency L0s <64ns, L1 <1us\n                LnkSta: Speed 2.5GT/s, Width x1\n                LnkCap2: Supported Link Speeds: 2.5GT/s, Crosslink- Retimer- 2Retimers- DRS-\n\n### COMMAND:\nlspci -vv -s c4:00.0 | grep -E 'LnkCap:|LnkSta:|LnkCap2:'\n### OUTPUT:\n                LnkCap: Port #0, Speed 2.5GT/s, Width x1, ASPM L0s L1, Exit Latency L0s <64ns, L1 <1us\n                LnkSta: Speed 2.5GT/s, Width x1\n                LnkCap2: Supported Link Speeds: 2.5GT/s, Crosslink- Retimer- 2Retimers- DRS-\n\n========== BAR / VRAM APERTURE ==========\n\n### COMMAND:\nlspci -vv -s 83:00.0 | grep -E 'Region 2|Resizable BAR'\n### OUTPUT:\n        Region 2: Memory at 1e800000000 (64-bit, prefetchable) [size=32G]\n        Capabilities: [420 v1] Physical Resizable BAR\n        Capabilities: [220 v1] Virtual Resizable BAR\n                Region 2: Memory at 000001f000000000 (64-bit, prefetchable)\n\n### COMMAND:\nlspci -vv -s c4:00.0 | grep -E 'Region 2|Resizable BAR'\n### OUTPUT:\n        Region 2: Memory at 16000000000 (64-bit, prefetchable) [size=32G]\n        Capabilities: [420 v1] Physical Resizable BAR\n        Capabilities: [220 v1] Virtual Resizable BAR\n                Region 2: Memory at 0000016800000000 (64-bit, prefetchable)\n\n========== VRAM DETECTED BY XE DRIVER ==========\n\n### COMMAND:\ndmesg | grep -E 'VRAM\\[|VISIBLE VRAM'\n### OUTPUT:\n[   10.962210] xe 0000:c4:00.0: [drm] VISIBLE VRAM: 0x0000016000000000, 0x0000000800000000\n[   10.962245] xe 0000:c4:00.0: [drm] VRAM[0]: Actual physical size 0x0000000540000000, usable size exclude stolen 0x000000053a600000, CPU accessible size 0x000000053a600000\n[   10.962250] xe 0000:c4:00.0: [drm] VRAM[0]: DPA range: [0x0000000000000000-540000000], io range: [0x0000016000000000-1653a600000]\n[   10.962254] xe 0000:c4:00.0: [drm] VRAM[0]: Actual physical size 0x0000000540000000, usable size exclude stolen 0x000000053a600000, CPU accessible size 0x000000053a600000\n[   10.962257] xe 0000:c4:00.0: [drm] VRAM[0]: DPA range: [0x0000000000000000-540000000], io range: [0x0000016000000000-1653a600000]\n[   11.680582] xe 0000:83:00.0: [drm] VISIBLE VRAM: 0x000001e800000000, 0x0000000800000000\n[   11.680700] xe 0000:83:00.0: [drm] VRAM[0]: Actual physical size 0x0000000540000000, usable size exclude stolen 0x000000053a600000, CPU accessible size 0x000000053a600000\n[   11.680704] xe 0000:83:00.0: [drm] VRAM[0]: DPA range: [0x0000000000000000-540000000], io range: [0x000001e800000000-1ed3a600000]\n[   11.680709] xe 0000:83:00.0: [drm] VRAM[0]: Actual physical size 0x0000000540000000, usable size exclude stolen 0x000000053a600000, CPU accessible size 0x000000053a600000\n[   11.680712] xe 0000:83:00.0: [drm] VRAM[0]: DPA range: [0x0000000000000000-540000000], io range: [0x000001e800000000-1ed3a600000]\n\n========== XE DRIVER STATE ==========\n\n### COMMAND:\nmodinfo xe | grep -E 'filename|vermagic|srcversion'\n### OUTPUT:\nfilename:       /lib/modules/6.17.13-2-pve/kernel/drivers/gpu/drm/xe/xe.ko\nsrcversion:     9DF73FAD15D81C9378B16E4\nvermagic:       6.17.13-2-pve SMP preempt mod_unload modversions\n\n### COMMAND:\ndmesg | grep -E 'xe .*drm'\n### OUTPUT:\n[   10.960229] xe 0000:c4:00.0: [drm] Running in SR-IOV PF mode\n[   10.960461] xe 0000:c4:00.0: [drm] Found battlemage (device ID e211) discrete display version 14.01 stepping B0\n[   10.962210] xe 0000:c4:00.0: [drm] VISIBLE VRAM: 0x0000016000000000, 0x0000000800000000\n[   10.962245] xe 0000:c4:00.0: [drm] VRAM[0]: Actual physical size 0x0000000540000000, usable size exclude stolen 0x000000053a600000, CPU accessible size 0x000000053a600000\n[   10.962250] xe 0000:c4:00.0: [drm] VRAM[0]: DPA range: [0x0000000000000000-540000000], io range: [0x0000016000000000-1653a600000]\n[   10.962254] xe 0000:c4:00.0: [drm] VRAM[0]: Actual physical size 0x0000000540000000, usable size exclude stolen 0x000000053a600000, CPU accessible size 0x000000053a600000\n[   10.962257] xe 0000:c4:00.0: [drm] VRAM[0]: DPA range: [0x0000000000000000-540000000], io range: [0x0000016000000000-1653a600000]\n[   11.118986] xe 0000:c4:00.0: [drm] GT0: Using GuC firmware from xe/bmg_guc_70.bin version 70.55.3\n[   11.132091] xe 0000:c4:00.0: [drm] Finished loading DMC firmware i915/bmg_dmc.bin (v2.6)\n[   11.538272] xe 0000:c4:00.0: [drm] GT0: ccs2 fused off\n[   11.538277] xe 0000:c4:00.0: [drm] GT0: ccs3 fused off\n[   11.566086] xe 0000:c4:00.0: [drm] GT1: Using GuC firmware from xe/bmg_guc_70.bin version 70.55.3\n[   11.574804] xe 0000:c4:00.0: [drm] GT1: Using HuC firmware from xe/bmg_huc.bin version 8.2.10\n[   11.587116] xe 0000:c4:00.0: [drm] GT1: vcs1 fused off\n[   11.587120] xe 0000:c4:00.0: [drm] GT1: vcs3 fused off\n[   11.587123] xe 0000:c4:00.0: [drm] GT1: vcs4 fused off\n[   11.587126] xe 0000:c4:00.0: [drm] GT1: vcs5 fused off\n[   11.587128] xe 0000:c4:00.0: [drm] GT1: vcs6 fused off\n[   11.587131] xe 0000:c4:00.0: [drm] GT1: vcs7 fused off\n[   11.587134] xe 0000:c4:00.0: [drm] GT1: vecs2 fused off\n[   11.587137] xe 0000:c4:00.0: [drm] GT1: vecs3 fused off\n[   11.623751] xe 0000:c4:00.0: [drm] Registered 4 planes with drm panic\n[   11.663881] xe 0000:c4:00.0: [drm] Cannot find any crtc or sizes\n[   11.676439] xe 0000:c4:00.0: [drm] Using mailbox commands for power limits\n[   11.676905] xe 0000:c4:00.0: [drm] PL2 is supported on channel 0\n[   11.678699] xe 0000:83:00.0: [drm] Running in SR-IOV PF mode\n[   11.678906] xe 0000:83:00.0: [drm] Found battlemage (device ID e211) discrete display version 14.01 stepping B0\n[   11.680582] xe 0000:83:00.0: [drm] VISIBLE VRAM: 0x000001e800000000, 0x0000000800000000\n[   11.680700] xe 0000:83:00.0: [drm] VRAM[0]: Actual physical size 0x0000000540000000, usable size exclude stolen 0x000000053a600000, CPU accessible size 0x000000053a600000\n[   11.680704] xe 0000:83:00.0: [drm] VRAM[0]: DPA range: [0x0000000000000000-540000000], io range: [0x000001e800000000-1ed3a600000]\n[   11.680709] xe 0000:83:00.0: [drm] VRAM[0]: Actual physical size 0x0000000540000000, usable size exclude stolen 0x000000053a600000, CPU accessible size 0x000000053a600000\n[   11.680712] xe 0000:83:00.0: [drm] VRAM[0]: DPA range: [0x0000000000000000-540000000], io range: [0x000001e800000000-1ed3a600000]\n[   11.743874] xe 0000:c4:00.0: [drm] Cannot find any crtc or sizes\n[   11.783868] xe 0000:c4:00.0: [drm] Cannot find any crtc or sizes\n[   11.849326] xe 0000:83:00.0: [drm] Finished loading DMC firmware i915/bmg_dmc.bin (v2.6)\n[   11.851220] xe 0000:83:00.0: [drm] GT0: Using GuC firmware from xe/bmg_guc_70.bin version 70.55.3\n[   12.267897] xe 0000:83:00.0: [drm] GT0: ccs2 fused off\n[   12.267903] xe 0000:83:00.0: [drm] GT0: ccs3 fused off\n[   12.295326] xe 0000:83:00.0: [drm] GT1: Using GuC firmware from xe/bmg_guc_70.bin version 70.55.3\n[   12.298426] xe 0000:83:00.0: [drm] GT1: Using HuC firmware from xe/bmg_huc.bin version 8.2.10\n[   12.310724] xe 0000:83:00.0: [drm] GT1: vcs1 fused off\n[   12.310729] xe 0000:83:00.0: [drm] GT1: vcs3 fused off\n[   12.310732] xe 0000:83:00.0: [drm] GT1: vcs4 fused off\n[   12.310735] xe 0000:83:00.0: [drm] GT1: vcs5 fused off\n[   12.310737] xe 0000:83:00.0: [drm] GT1: vcs6 fused off\n[   12.310740] xe 0000:83:00.0: [drm] GT1: vcs7 fused off\n[   12.310742] xe 0000:83:00.0: [drm] GT1: vecs2 fused off\n[   12.310745] xe 0000:83:00.0: [drm] GT1: vecs3 fused off\n[   12.347087] xe 0000:83:00.0: [drm] Registered 4 planes with drm panic\n[   12.426868] xe 0000:83:00.0: [drm] Cannot find any crtc or sizes\n[   12.429864] xe 0000:83:00.0: [drm] Using mailbox commands for power limits\n[   12.430327] xe 0000:83:00.0: [drm] PL2 is supported on channel 0\n[   12.466870] xe 0000:83:00.0: [drm] Cannot find any crtc or sizes\n\n========== FIRMWARE ==========\n\n### COMMAND:\ndmesg | grep -Ei 'guc|huc|gsc'\n### OUTPUT:\n[   11.118986] xe 0000:c4:00.0: [drm] GT0: Using GuC firmware from xe/bmg_guc_70.bin version 70.55.3\n[   11.566086] xe 0000:c4:00.0: [drm] GT1: Using GuC firmware from xe/bmg_guc_70.bin version 70.55.3\n[   11.574804] xe 0000:c4:00.0: [drm] GT1: Using HuC firmware from xe/bmg_huc.bin version 8.2.10\n[   11.851220] xe 0000:83:00.0: [drm] GT0: Using GuC firmware from xe/bmg_guc_70.bin version 70.55.3\n[   12.295326] xe 0000:83:00.0: [drm] GT1: Using GuC firmware from xe/bmg_guc_70.bin version 70.55.3\n[   12.298426] xe 0000:83:00.0: [drm] GT1: Using HuC firmware from xe/bmg_huc.bin version 8.2.10\n[   12.486910] 0x000000001000-0x00000054e000 : \"xe.nvm.50176.GSC\"\n[   12.498589] 0x000000001000-0x00000054e000 : \"xe.nvm.33536.GSC\"\n\n========== DRM DEVICES ==========\n\n### COMMAND:\nls -l /dev/dri\n### OUTPUT:\ntotal 0\ndrwxr-xr-x 2 root root        140 Mar 16 11:26 by-path\ncrw-rw---- 1 root video  226,   0 Mar 16 11:26 card0\ncrw-rw---- 1 root video  226,   1 Mar 16 11:26 card1\ncrw-rw---- 1 root video  226,   2 Mar 16 11:26 card2\ncrw-rw---- 1 root render 226, 128 Mar 16 11:26 renderD128\ncrw-rw---- 1 root render 226, 129 Mar 16 11:26 renderD129\n\n========== IOMMU GROUPS ==========\n\n### COMMAND:\nfind /sys/kernel/iommu_groups -type l | grep 83:00.0\n### OUTPUT:\n/sys/kernel/iommu_groups/37/devices/0000:83:00.0\n\n### COMMAND:\nfind /sys/kernel/iommu_groups -type l | grep c4:00.0\n### OUTPUT:\n/sys/kernel/iommu_groups/16/devices/0000:c4:00.0\n\n========== COMPLETE ==========\nDiagnostic output saved to gpu_diag_20260316_121003.txt\nroot@proxmox:~#\n```\n\n[anlin](https://forum.level1techs.com/u/anlin)\n2\nThe gen1 x1 link status is a known quirk with Arc cards. The PCIe bridge upstream of the G21 device should show the actual link status (16 GT/s x8 would be expected in your case).\n\nYeah, I figured out that I needed to load the cards into Intel Arc Pro Graphics on a Windows machine in order to update the firmware and disable the ECC reservation on the VRAM so that all of the capacity is usable. I have yet to figure out how to pass the whole GPU into a VM in Proxmox due to the weird bridge situation (I couldn’t change the driver for the root port of the bridge from ‘pcieport’ to ‘vfio-pci’ so that it would be selectable as a PCI device), so I just created an LXC and it seems to be working fine for the time being\n\nHmm. Is the passthrough situation weird? I didn’t have to change anything at all for my arc pro cards, it just worked. SR-IOV needs a tiny bit of fiddling but even that iirc is out of the box nearly.\n\nI was initially using an Ubuntu 25.04 VM and the xe driver wouldn’t bind to the devices that I passed through correctly. I just tried spinning up a new VM with a 25.10 image earlier today and everything worked fine in terms of llama.cpp using the GPU properly.\n\n[BillM](https://forum.level1techs.com/u/BillM)\n6\nI have the same with a B50 Pro, the link state shows 2.5GT/s at PCIe x1:\n\n```\n# lspci -s b5:00.0 -vv\nb5:00.0 VGA compatible controller: Intel Corporation Battlemage G21 [Intel Graphics] (prog-if 00 [VGA controller])\n        Subsystem: Intel Corporation Device 1114\n        Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr+ Stepping- SERR+ FastB2B- DisINTx-\n        Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-\n        Latency: 0, Cache Line Size: 32 bytes\n        Interrupt: pin ? routed to IRQ 185\n        NUMA node: 0\n        IOMMU group: 4\n        Region 0: Memory at fa000000 (64-bit, non-prefetchable) [size=16M]\n        Region 2: Memory at 38f800000000 (64-bit, prefetchable) [size=16G]\n        Expansion ROM at fb000000 [disabled] [size=2M]\n        Capabilities: [40] Vendor Specific Information: Len=0c <?>\n        Capabilities: [70] Express (v2) Endpoint, IntMsgNum 0\n                DevCap: MaxPayload 256 bytes, PhantFunc 0, Latency L0s unlimited, L1 unlimited\n                        ExtTag+ AttnBtn- AttnInd- PwrInd- RBE+ FLReset+ SlotPowerLimit 0W TEE-IO-\n                DevCtl: CorrErr- NonFatalErr- FatalErr+ UnsupReq-\n                        RlxdOrd+ ExtTag+ PhantFunc- AuxPwr- NoSnoop+ FLReset-\n                        MaxPayload 256 bytes, MaxReadReq 512 bytes\n                DevSta: CorrErr- NonFatalErr- FatalErr- UnsupReq- AuxPwr- TransPend-\n                LnkCap: Port #0, Speed 2.5GT/s, Width x1, ASPM L0s L1, Exit Latency L0s <64ns, L1 <1us\n                        ClockPM- Surprise- LLActRep- BwNot- ASPMOptComp+\n                LnkCtl: ASPM Disabled; RCB 64 bytes, LnkDisable- CommClk-\n                        ExtSynch- ClockPM- AutWidDis- BWInt- AutBWInt-\n                LnkSta: Speed 2.5GT/s, Width x1\n                        TrErr- Train- SlotClk- DLActive- BWMgmt- ABWMgmt-\n```\n\nWhich surprised me, as I thought it was PCIe x8.\n\nSo I look up the tree for the bridge:\n\n```\n# lspci -s b5:00.0 -tv\n-[0000:b2]---00.0-[b3-b6]----00.0-[b4-b6]----01.0-[b5]----00.0  Intel Corporation Battlemage G21 [Intel Graphics]\n```\n\nAnd, if I am reading it right (the tree and the bridge), it is really connected at 8GT/s with x8 width:\n\n```\nlspci -s b2:00.0 -vv\nb2:00.0 PCI bridge: Intel Corporation Sky Lake-E PCI Express Root Port A (rev 07) (prog-if 00 [Normal decode])\n        Subsystem: Super Micro Computer Inc Device 095e\n        Physical Slot: 9\n        Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr+ Stepping- SERR+ FastB2B- DisINTx+\n        Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-\n        Latency: 0, Cache Line Size: 64 bytes\n        Interrupt: pin A routed to IRQ 36\n        NUMA node: 0\n        IOMMU group: 0\n        Bus: primary=b2, secondary=b3, subordinate=b6, sec-latency=0\n        I/O behind bridge: [disabled] [16-bit]\n        Memory behind bridge: fa000000-fb3fffff [size=20M] [32-bit]\n        Prefetchable memory behind bridge: 38f800000000-38fc007fffff [size=16392M] [32-bit]\n        Secondary status: 66MHz- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- <SERR- <PERR-\n        BridgeCtl: Parity+ SERR+ NoISA- VGA+ VGA16+ MAbort- >Reset- FastB2B-\n                PriDiscTmr- SecDiscTmr- DiscTmrStat- DiscTmrSERREn-\n        Capabilities: [40] Subsystem: Super Micro Computer Inc Device 095e\n        Capabilities: [60] MSI: Enable+ Count=1/2 Maskable+ 64bit-\n                Address: fee00038  Data: 0000\n                Masking: 00000002  Pending: 00000000\n        Capabilities: [90] Express (v2) Root Port (Slot+), IntMsgNum 0\n                DevCap: MaxPayload 256 bytes, PhantFunc 0\n                        ExtTag+ RBE+ TEE-IO-\n                DevCtl: CorrErr- NonFatalErr- FatalErr+ UnsupReq-\n                        RlxdOrd- ExtTag+ PhantFunc- AuxPwr- NoSnoop-\n                        MaxPayload 256 bytes, MaxReadReq 128 bytes\n                DevSta: CorrErr- NonFatalErr- FatalErr- UnsupReq- AuxPwr- TransPend-\n                LnkCap: Port #9, Speed 8GT/s, Width x16, ASPM L1, Exit Latency L1 <16us\n                        ClockPM- Surprise+ LLActRep+ BwNot+ ASPMOptComp+\n                LnkCtl: ASPM Disabled; RCB 64 bytes, LnkDisable- CommClk+\n                        ExtSynch- ClockPM- AutWidDis- BWInt+ AutBWInt+\n                LnkSta: Speed 8GT/s, Width x8\n                        TrErr- Train- SlotClk+ DLActive+ BWMgmt- ABWMgmt-\n```\n\nSo I’m guessing that I am really getting x8 after all\n\nI’d test the RX 580 in another PC first. If the same thing happens, the card is probably the problem, not the motherboard.\n\nWould you be able to provide a link to more information about this? I’m seeing Gen 1.0 speeds on my Intel Arc A770 per lspci and I’m trying to figure out where the problem lies.", "url": "https://wpnews.pro/news/intel-arc-pro-b60-stuck-at-pcie-gen1-x1-and-21gb-vram-in-proxmox", "canonical_source": "https://forum.level1techs.com/t/intel-arc-pro-b60-stuck-at-pcie-gen1-x1-and-21gb-vram-in-proxmox/247563#post_8", "published_at": "2026-08-23 02:05:32+00:00", "updated_at": "2026-08-23 02:12:50.764161+00:00", "lang": "en", "topics": ["ai-infrastructure", "ai-tools"], "entities": ["Intel Arc Pro B60", "Proxmox", "AMD EPYC 7K62", "Supermicro H12SSL-i", "Newegg", "Intel", "xe driver"], "alternates": {"html": "https://wpnews.pro/news/intel-arc-pro-b60-stuck-at-pcie-gen1-x1-and-21gb-vram-in-proxmox", "markdown": "https://wpnews.pro/news/intel-arc-pro-b60-stuck-at-pcie-gen1-x1-and-21gb-vram-in-proxmox.md", "text": "https://wpnews.pro/news/intel-arc-pro-b60-stuck-at-pcie-gen1-x1-and-21gb-vram-in-proxmox.txt", "jsonld": "https://wpnews.pro/news/intel-arc-pro-b60-stuck-at-pcie-gen1-x1-and-21gb-vram-in-proxmox.jsonld"}}