WikiBot: New page (2026-10-30 wiki refresh): Project Goals

2026-10-30T17:04:55Z

New page (2026-10-30 wiki refresh): Project Goals

New page

{{About message}}

= Project Goals =

<blockquote>'''Scope.''' What Open-IOV is for, what it is not for, and the dated status of each goal. Condensed from [https://git.vgpu.io/open-iov.git/tree/docs/01-orientation/01-purpose-and-scope.md <code>../docs/01-orientation/01-purpose-and-scope.md</code>] (mission, scope and success criteria) and [https://git.vgpu.io/open-iov.git/tree/docs/README.md <code>../docs/README.md</code>] (the project front door); the docs tree wins on disagreement. The live milestone ladder is on [[Project Status|<code>Project-Status.md</code>]], and the current execution detail on [[The Execution Frontier|<code>Execution-Frontier.md</code>]] and [[Graphics Execution|<code>Graphics-Execution.md</code>]].</blockquote>

Status: 2026-10-30, tree <code>master @ b4930168</code>. Tags: <code>[live]</code> working with the stock guest driver; <code>[wip]</code> in progress; <code>[planned]</code> designed, not built; <code>[target]</code> bring-up ahead; <code>[non-goal]</code> explicitly out of scope.

== 1. Mission ==

'''Make NVIDIA GPUs fully virtualizable, by anyone, for any hypervisor, without modifying NVIDIA software, firmware or hardware.'''

Concretely, Open-IOV builds an independent, maintained IP block that:

* presents a virtual NVIDIA GPU to a guest operating system;
* lets the guest run the '''stock NVIDIA driver''' (Linux and Windows);
* supports '''compute (CUDA)''' first, then '''3D graphics''' and '''video''';
* runs on ordinary NVIDIA GPUs already deployed in the field;
* isolates guests from each other and from the host;
* and requires no license key, no proprietary host service, no patched guest driver, and no specific hypervisor.

The guiding sentence is: '''''Emulate the interface. Pass through the work. Never patch anyone.'''''

== 2. The gap it fills ==

{| class="wikitable"
|-
! Approach
! Limitation
|-
| Discrete device assignment (VFIO passthrough)
| One GPU per VM; no sharing; fragile on consumer/laptop systems.
|-
| NVIDIA proprietary vGPU
| Requires a license and a vendor host stack; driver-version locked.
|-
| Software GPU emulation (QEMU virtual GPUs, <code>virgl</code>)
| No real CUDA; no real 3D performance; not usable for AI.
|-
| Cloud-only inference
| Requires an external service; unacceptable for secure edge, offline, or private data.
|}

Open-IOV occupies the gap: '''real GPU acceleration, multiple VMs, stock drivers, no license.'''

== 3. Goals and their status ==

{| class="wikitable"
|-
! Goal
! What success looks like
! Status (2026-10-30)
|-
| Stock guest driver initialisation
| The unmodified driver probes, initialises and enumerates as on bare metal (<code>nvidia-smi</code>)
| <code>[live]</code> — TU104 reference platform
|-
| Full CUDA driver-API surface
| Context creation, allocation, copies, free, destroy complete with correct results
| <code>[live]</code> — the driver-API smoke prints <code>ALL OK</code>; the 64-byte payload round-trips
|-
| CUDA kernel execution on real hardware
| Verified guest kernels execute on the physical GPU, with engine-authored completions waking the guest
| <code>[live]</code> — vector-add, matrix-multiply and transformer clients green on fresh boots; the staged-transfer class, the every-boot acceptance campaign and per-guest isolation are in progress ([[The Execution Frontier|<code>Execution-Frontier.md</code>]])
|-
| OpenGL (3D graphics)
| Stock guest GL applications render real pixels through the emulated vGPU
| <code>[wip]</code> — Xorg/GLX run, the engine executes GL method streams (verbatim 3D replay + synthesized launches); rendered pixels are the frontier ([[OpenGL|<code>OpenGL.md</code>]])
|-
| Vulkan
| Stock guest Vulkan applications render real pixels
| <code>[wip]</code> — the pipeline runs end to end (device → command buffer → fence → readback); frames are black — program delivery and program-VA mapping are the frontier ([[Vulkan|<code>Vulkan.md</code>]])
|-
| Windows guests (DirectX, CUDA)
| A stock WDDM driver and stock D3D/CUDA applications run in Windows guests
| <code>[target]</code> — not started; same emulation target ([[DirectX|<code>DirectX.md</code>]])
|-
| Video (NVENC/NVDEC)
| Encode/decode through the emulated device
| <code>[planned]</code>
|-
| Multi-tenant isolation
| Multiple VMs on one GPU, isolated by the IOMMU and the device model, with explicit scheduling
| <code>[wip]</code> — two mediated devices (different emulated VRAM sizes) and two guests coexist; per-guest execution isolation is in progress ([[GMMU Multi-Tenant Protection|<code>GMMU-Multi-Tenant-Protection.md</code>]])
|-
| Hypervisor portability
| The same IP block attaches to KVM, Xen, Argo/OpenXT, Qubes OS Air, ...
| <code>[planned]</code> — the emulation core is hypervisor-independent by construction; KVM/QEMU with VFIO mdev is the reference transport ([[Hardware and Software Matrix|<code>Hardware-Software-Matrix.md</code>]])
|-
| Version and hardware portability
| New driver, firmware or GPU generations enter by configuration plus measurement
| <code>[wip]</code> — 570.144/TU104 is the anchor; per-version evidence workflows exist; newer versions are gated work
|-
| Live migration
| A VM migrates with its GPU state
| <code>[planned]</code> — the protocol analysis is kept for the record ([[Migration Protocol|<code>Migration-Protocol.md</code>]])
|-
| Performance
| Within an agreed factor of native for the supported cases, with published methodology
| <code>[planned]</code> — no published numbers yet; the first scheduled measurement is the 60 fps graphics target ([[Graphics Execution|<code>Graphics-Execution.md</code>]])
|-
| Security posture
| Isolation survives adversarial review; the device fails closed
| <code>[wip]</code> — fail-closed design and audit practice; per-tenant execution isolation is in progress ([https://git.vgpu.io/open-iov.git/tree/docs/05-operations/01-performance-correctness-security.md <code>../docs/05-operations/01-performance-correctness-security.md</code>])
|}

== 4. How success is judged ==

The project evaluates itself on five axes, each with a methodology document in [https://git.vgpu.io/open-iov.git/tree/docs/05-operations/01-performance-correctness-security.md <code>../docs/05-operations/01-performance-correctness-security.md</code>]:

# '''Initialization''' — the driver reaches bare-metal parity through boot, probe and initialisation.
# '''Runtime''' — the guest runs real work: CUDA kernels with correct results, OpenGL/Vulkan rendering, encode/decode; repeated, sustained, resilient.
# '''Interoperability''' — no changes anywhere: guest driver, host driver, firmware, hypervisor.
# '''Stability''' — crash-free under stress: start/stop cycles, VM kill, driver reload, suspend/resume, migration.
# '''Performance and security''' — within an agreed factor of native, with isolation that survives adversarial review.

== 5. Non-goals ==

* '''Replacing bare-metal inference serving.''' Continuous batching on bare-metal hosts remains the right answer for centralised datacenter inference; Open-IOV is for compute that must live ''inside'' guests.
* '''Forking or patching''' the NVIDIA driver (guest or host), the GPU firmware, or hypervisors to make the device work.
* '''Anti-cheat / DRM circumvention.''' The project interoperates with interfaces and implements '''its own''' device; it does not counterfeit vendor licensing.
* '''Emulating GPU instruction execution in software.''' Software emulation of CUDA kernels is orders of magnitude too slow; work must run on real silicon.

== 6. Related pages ==

* [[Project Status|<code>Project-Status.md</code>]] — the dated milestone ladder.
* [[The Execution Frontier|<code>Execution-Frontier.md</code>]] — the compute execution frontier.
* [[Graphics Execution|<code>Graphics-Execution.md</code>]] — the 3D execution frontier.
* [[HAAEEP and Collaboration|<code>HAAEEP-and-Collaboration.md</code>]] — the agent-edge use case and collaboration models.
* [[Hardware and Software Matrix|<code>Hardware-Software-Matrix.md</code>]] — the support ladder.
* [https://git.vgpu.io/open-iov.git/tree/docs/01-orientation/01-purpose-and-scope.md <code>../docs/01-orientation/01-purpose-and-scope.md</code>] — the canonical purpose-and-scope document.

Project Goals - Revision history

WikiBot: New page (2026-10-30 wiki refresh): Project Goals