Igor 10 on macOS (in VM) experience

With IP10 coming out soon,

it may be useful to have place where to exchange how to run IP10 on macOS for those, who want to keep macOS as their main system. 

Here are some notes about my setup and experience. I hope others add more to  make this useful. If more information is added, may be WM can add summary their web pages for users to find, if they need to. 

I have been using IP10 beta in Parallels VM with Arm64 version of Windows 11 successfully for months now. I have tested the same setup using VMware Fusion, which is free for most use, and results were similar with bit more limitations and annoyances.  

Setup:

Parallels Desktop (I have business license) with 8GB RAM and 4 cpus assigned to it on MacBook Pro, with M1 Pro cpu, 32GM RAM, and 1TB SSD. This leaves enough RAM for macOS to run anything I tried (including local some LLMs!). Parallels can be run in "coherence" mode which hides the Windows desktop and Windows applications run in separate windows as if they were Mac applications. Network mapping can be used to map macOS home folder (and macOS drives) to Windows.  For Windows license, I bought regular (reseller discounted to $20) license and it works with ARM64 Windows.  The virtual machine data bundle on macOS SSD grew over years from ~20GB fresh to ~40GB. This contains now Windows OS, some support apps, IP8, IP9, and IP10, ~8GB of Python packages, and VScode. I can actually use VSCode to edit Igor procedure files and that works well enough aslo, including using some AI help (separate discussion on which LLM best understand Igor coding). 
By now I simply leave running the VM all the time, it is set to sleep in not used for more than 10 minutes.

VMware Fusion - free application, license for Windows is the same method (different license) as for Parallels. Fusion does not have "coherence" mode, so you get Windows desktop to run in. That is small annoyance. Somehow bigger annoyance is access to Mac home folder which is more challenging to setup. The speed of graphics seems to be somehow slower, looks like less optimization. Not really issues, just not as smooth. Overall, fine solution if needed.  

What works (Parallels mostly): 

IP10 runs very well, speed seems to be as fast as I would expect (comparable to IP9 on macOS). I did not run into any problems. I run Python code from within IP10 and it works perfectly fine with Python packages installed by using miniconda. Access to data and Iggor experiments on macOS home is fine, no problems observed. Printing works well through Windows printers which are configured automatically. 

What are problems:

1. Copy-paste. Clipboard between macOS and Windows is shared through some Parallels magic asnd seems to mostly work. But as far as I can say you cannot copy graphics (e.g., graph) from Windows IP10 and paste into macOS apps. Not sure where this comes from, if this is something one could configure, but bit annoying.
Solution : (more like workaround) is screen copy area using macOS screen copy tools and paste where needed, which is lower resolution but works mostly fine. Or, one can save image file and insert that where needed etc. 

2. Procedure file access. If you want to keep procedure files on macOS but edit in Windows IP10, do not do it... I tried linking procedure files location from macOS home, through Parallels sharing networking, to Windows Igor Pro User files location and it turned up to be really annoying. While just using them was fine, editing was really annoying as the system kept complaining about procedure files I was editing changing in their location and was asking me for resolution of version conflict. Like if two different editors were changing the files at once (which was not the case). I suspect this is the same issue as storing procedure files in network drives. Networks drives seems to have mind of their own and some caching/slowness in them which causes troubles. And these files are located on the same (very fast) ssd, just there is added networking layer to get to them. And that is enough to make editing them impractical. 
Solution : make local copy in VM and use some syncing method (Github. subversion, syncing script) to keep locations in sync. 

Anyone else has suggestions, comments, different experience? I'd like to optimize this setup since it looks like I will need to use it for some time now (until WM makes macOS IP11???).