I'm trying to find some VM software that can access more than 256mb of the video card. Parallels tops out at 256mb. It won't let you access more than that.

The machine I intend to run this on will be an 8core mac pro workstation desktop with 8gigs of RAM or more and a terabyte HD. This will be for doing 3d rendering, not to create images, but to create models that I can send to a 3d printer or CNC mill.

The problem is I already own the 3d rendering software and it is for Windows, and I am soooo done with windows, so i want to minimize my dependence on it.

If I can find a VM that can access as much as 1gig of a video card, I would probably have a second video card installed in the machine. I am willing to settle for 512 mb of usage.

Does such a VM exist? I haven't heard anything conclusive about VMware or WINE or anything else. VMware, Parallels, and WINE's customer support desk don't answer emails apparently.

Thanks for your help!

Dani AI

Generated

Building on ’s goal (heavy 3D work, lots of VRAM) and the suggestions from and to try different hypervisors, the practical reality is: if the renderer needs full, native access to the GPU and its framebuffer, a hosted VM on macOS is rarely the right tool. For Intel Macs, the simplest way to guarantee full GPU/driver/VRAM access is to run Windows natively (Boot Camp). This gives the guest direct access to the card and vendor drivers. See Apple’s Boot Camp documentation for the supported workflow and requirements. (support.apple.com)

If you want a VM-like convenience while keeping near-native GPU performance, the technical option is PCI/GPU passthrough (assigning a whole physical GPU to a VM). That can deliver almost bare‑metal performance, but it is advanced: it needs IOMMU/VT‑d support, a hypervisor that exposes passthrough (KVM/QEMU on Linux or ESXi in datacenter setups), and usually a separate GPU for the host or multiple GPUs for multiple guests. This is not a trivial plug‑and‑play solution on macOS desktop hypervisors; it’s an architecture-level feature of server/host hypervisors. (wiki.archlinux.org)

There are also commercial vGPU products (NVIDIA vGPU / GRID) that partition a physical GPU into multiple virtual GPUs with fixed frame‑buffer sizes for guests — but these are enterprise solutions that require supported hardware, special hypervisor integration and licensing, and still impose tradeoffs versus full passthrough. For many desktop CAD/CAM apps, compatibility and driver fidelity matter more than a “virtual” video buffer. (docs.nvidia.com)

WINE/CrossOver is a different approach (no Windows install), and it has improved 3D support over the years, but compatibility for complex professional 3D/CAD applications is variable and should be verified case-by-case via the Wine AppDB or vendor guidance. If avoiding Windows is the priority, consider (a) running the Windows renderer on a dedicated Windows machine (or a Boot Camp partition) and using file sync/remote-render workflows, or (b) investing the time in a passthrough setup on a supported host — both are far more likely to meet a 512MB–1GB GPU‑memory need than a hosted macOS VM. (winehq.org)

Recommended Answers

All 2 Replies

http://www.virtualbox.org/
It's another VM that runs on Macs, and can host windows.

I don't know about the specs though, that is something you would need to look into.

On the plus side, it's open source, so you might be able to arrange your own custom hack on the video side.

have you tried VM fusion?

Be a part of the DaniWeb community

We're a friendly, industry-focused community of developers, IT pros, digital marketers, and technology enthusiasts meeting, networking, learning, and sharing knowledge.