Earlier in the year I upgraded to a 2Tb ssd for my Surface Pro 9. I created three partitions with two having identical Win11 installs and the third being a Data drive for storage.
I’ve just read about Win11 Pro coming with VM capability and wondered whether creating a virtual machine running GP would be a better option than multiple install on two partitions.
So my question is - what are the advantages of GP running on a VM as compared to a normal instal on a partition? Would GP run any better and possibly faster? Or is there very little or no improvement at all?
Well, in principle, one major advantage of using a decent VM (VMWare for example), is that you can copy the entire machine to a new computer and just keep using it.
The issue really boils down to how well a VM will handle peripherals (USB, USBC, etc) and real-time audio.
Because your machine has the resources to do that - the host doesn’t have to be doing very much - and you get the other benefit of the ability to just move the VM.
I think if you’re on Linux, this is particularly beneficial (assuming it’s responsive enough) because you can have a very lean Linux server that does barely anything and then give the VM lots of resources.
Remember that it is Windows … and by default it does much. It first needs to be optimized to be a host. And then “guest” needs to be optimized for audio.
It is too much complicated and IMO without any benefits.
I wonder, however, did anybody try this? Or even had a gig on a virtual machine?
It would be fun to try, just for trying. My daily job involves providing the infrastructure for a few thousands of VMs and have them running well, so it’s really something that appeals to me
‘Complicated’ is like beauty: both in the eye of the beholder.
‘Benefits’ could be the possibility to take snapshots before doing an OS update.
But it will take resources anyway and licensing for plugins could be costly if someone likes to use them in multiple vms.
I must admit it is not a usual setup and I don’t think even if it really runs well, many users will want to use it.
At the moment I don’t have the time: there’s better fish to fry.
In his particular case, yes - but it doesn’t have to be. I have tested GP running on a Windows virtual machine inside my Intel Mac and it actually worked quite well. We are not in a position to official support this usage (because there are too many factors outside our control such as the VM being able to handle peripherals, response time and so on) but it can be done.
Many things can be done – e.g. I remember those Hackintosh projects – but that is not something I’d do for a critical use case such as live performance.
If you run mac vms on mac hardware, you don’t need a hackintosh. Although there’s this issue: VMware doesn’t sell hypervisors (at least not the bare metal version) for Mac anymore, because they won’t/can’t support apple-silicon.
I would not be comfortable with using either a VM or a Hackintosh in live performance. Critical real-time audio should be handled directly by supported hardware/operating system.
I don’t have any particular reason or inclination to test this right now but I bet that a decent VM host (VMWare Workstation) would handle real-time audio pretty well. I’ve used VMs for many many years (still am) on top of both Linux machines and Intel Macs and they’re pretty darn good.
This article was rather interesting describing how well games ran on a Windows Guest with a Linux host
Some/a lot of gamers are very picky about the frame rates and the response times of the display. I think their needs are very alike real-time processing. But then, they have the GPU to offload things to.
I understand that people shy away of using a VM for real-time processing, but today’s hypervisors make it possible to connect hardware devices without any interference of the hypervisor itself. This means there’s not necessarily additional latency in that respect. The VM itself, however, is competing with the other programs on the host, so that could be more problematic.
As I said before: I’d like to investigate this (just for the sake of knowing), but at the moment I lack the time.
Yes, but the biggest difference is that if a gamer gets a 30ms video glitch every 5 minutes, its not a big deal at all. Whereas a similar audio glitch(or any audio artifact really) for a musician is a huge red flag.
… and by default it does much. It first needs to be optimized to be a host. And then “guest” needs to be optimized for audio.
