Hi guys, basically as the title says I want to make external SSD drive with “Windows to Go” for the stuff that I really need Windows for unfortunately (proprietary CAD software) but there is no software for making this on Linux that I can find
Edit: typo
It should.
Hell I have hardware dongles I have to deal with that work fine in a VM.
Are you sure you can’t go the easier route with a VM? Running off USB will be slow and unpleasant, especially with windows.
What’s the VM issue?
Have you ever tried running Solidworks or Fusion 360 or Ansys or other Autodesk software in a vm? Painful is all I can say
I have a client running both Solidworks and Fusion 360 in VMWare Workstation with GPU pass-thru enabled, it’s pretty straightforward to setup and the end-users were pleased with it’s performance. If you have USB license keys those usually work as well, just setup the USB device pass-thru.
Yes.
On my virt host, which I pass the GPU to. I remote into it from my laptop.
In my case it’s been hardware acceleration. Sure I can make the adobe suite “run” but it’s not really usable at a professional level without hardware accel.
Single GPU or dual GPU laptop?
Have you looked at libvf.io?
virtio-gpu (if you look to the more recent work) has made big strides, but I also wouldn’t consider it “ready”.
Single gpu on my desktop. AMD 7900xt. I haven’t looked at either of those. I’ll give them a browse!
Take a look, and consider one more option (I have this setup in one of my workstations).
A second GPU. Older is fine, I have an older quadro 2000 in one machine thats great for cad, when I load up my VM for AutoCAD architecture, thats the GPU its using.
I have a few others with quadro 620s in them, the Intel iGPU is accessed shared by being an lxc, the quadros I pass through to the VMs. Some need very different configs, so I only power up the VM when I need it, so they can all use the same passed through GPU.
Ohh I see. You just dedicate another gpu to the VM? That sounds cool. I do have an extra.
Yup! Doesnt matter if its running in a reduced capacity by being in a x8 instead of an x16 either (which honestly won’t matter as much to performance as you might think anyway), just about any hardware acceleration makes a world of difference with those applications.