Passthrough

One physical GPU split into virtual functions handed to separate Windows desktops

Licence-Free Windows VDI on Proxmox with an Intel Arc Pro B50 — and the Firmware Limit That Stopped It

Intel’s Arc Pro cards do SR-IOV natively, with no vGPU licence to buy — which makes a licence-free Windows VDI on Proxmox genuinely possible. The whole build on a B50, the Windows-first bootstrap problem, why the firmware caps it at two virtual functions, and how many each card in the B-series gives you.

Kernel boot flags sorted into a host column and a guest column

Two Boot Lines — Which Kernel Flags Belong on a Proxmox Host, and Which Belong in the Guest

Most Proxmox tuning lines you find online are one blob of kernel flags. Half of them belong on the hypervisor, half belong inside the guest, two of the popular ones do nothing at all, and a few change meaning depending on which side of the boundary they land.

A curve rising from 36 percent of bare metal at queue depth 1 to within a few percent by queue depth 32

PCIe Passthrough Performance on Proxmox VE — The IOMMU Tax and How to Minimise It

Why PCIe devices lose throughput when passed through to a VM via VFIO, and the practical tuning steps that claw most of it back.

Two CPU sockets with a DMA path crossing the inter-socket link between them

NUMA Alignment on Proxmox VE — Why It Matters and How to Get It Right

On multi-socket systems, a VM with its vCPUs on one NUMA node and its passed-through device on another loses 20–30% throughput before you’ve even looked at anything else.

A latency-percentile curve that stays flat to p90 then climbs steeply, beside one that stays flat

PCIe ASPM and Why You Should Disable It for Passthrough

Active State Power Management saves a few watts on idle PCIe links. Under VFIO passthrough, it adds latency jitter that’s hard to diagnose and easy to fix.

The same 4 KB payload drawn as 32 packets at MPS 128 and 8 packets at MPS 512

PCIe MaxPayloadSize — A Free Performance Win for Passthrough

QEMU’s virtual root complex defaults to 128-byte TLP payloads. Most devices support 256 or 512. One kernel parameter fixes it.

A flat 1996-era PCI bus beside a 2007-era PCIe root complex

Always Use Q35, Not i440fx — Why It Matters on Proxmox VE

The two QEMU virtual chipsets are not interchangeable. Q35 provides a proper PCIe topology that passthrough, modern Windows, and the wider KVM ecosystem all depend on.