All idea exchange posts in Fusion Tech Preview 22H2

Re: 3D Graphics Acceleration?

Technogeezer — Fri, 18 Nov 2022 15:43:58 GMT

You can't. They didn't include 3D support for Windows 11 ARM. 2D only.

A major disappointment.

Re: 3D Graphics Acceleration?

Geoffrey84 — Fri, 18 Nov 2022 15:36:02 GMT

I've just downloaded and installed the release version of VMware Fusion 13.0 Pro but don't seem to have any 3D acceleration in my Windows 11 x64 ARM virtual machines with macOS 13 Ventura as the host (MacBook Pro M1 Max).

I tried updating VMware Tools but the installer noted that no update was needed. Many items in the hardware device list are still missing, however, and any games I try to load crash instantly.

Can anyone please advise on how to enable 3D graphics acceleration in the final release version of VMware Fusion 13 in Windows x64 11 ARM guest machines?

Thanks for your help

Re: 3D Graphics Acceleration?

Geoffrey84 — Thu, 17 Nov 2022 12:59:58 GMT

You're not being dense, since 3D graphics acceleration is the most important feature all of us are anxiously awaiting. Rumor has it that a production release of VMware Fusion 13 for macOS is imminent, and I've been checking this forum daily hoping that magic moment has finally arrived. Without Boot Camp support and CrossOver performing miserably all we have besides UTM, which is also limited to 2D graphics, is Parallels Desktop, and I for one would love to move away from that option and switch to Fusion as soon as possible!

3D Graphics Acceleration?

bfranco991 — Thu, 17 Nov 2022 05:46:33 GMT

Forgive me for being naive (or dense or any other adjective) but are there plans for including 3D graphics acceleration in the next fusion iteration for Apple Silicon? My use case is in 3D CAD programs (mostly SolidWorks) and though it's bearable to run on my M1 Max MacBook, I would absolutely love to take advantage of the 32 core GPU for 3D Modeling.

Expand checks for proper permissions on directories at Fusion launch

Technogeezer — Wed, 28 Sep 2022 15:23:28 GMT

There have been a rash of incidents lately where Fusion failures are related to improper permissions on directories (in particular there have been way too many errors related to incorrect permissions on /private/tmp) that Fusion is looking to write into. This usually has resulted in the broken pipe 14 errors. The fix for this is relatively simple, but the ordinary user does not have the experience to review logs to determine if broken permissions are the cause of this error.

Could Fusion be enhanced to check at the time of the UI launch for proper permissions on directories that it's looking to write into (in particular, /tmp or /private/tmp)? Responding with an error that something's amiss at this point with an intelligent error message gives the user a chance to address it properly (before endless attempts and frustrations from trying to run a VM with the chance of the broken pipe error).

Re: Running intel binaries in linux vm's with rosetta

Technogeezer — Mon, 19 Sep 2022 01:58:37 GMT

There are reports that Apple is restricting the use of the Rosetta on Linux feature to virtual machines using the Virtualization Framework. It's also implemented using a virtio virtual device - which VMware doesn't support.

Lots has to change from both Apple and VMware for this to happen with Fusion.

Yes, looks like UTM 4.0.5 beta looks like it now has integrated the Virtualization framework, which allows it to get the Rosetta for Linux implementation. If you do try it, let us know how it works and behaves and more importantly what features for virtual devices are supported (for example, suspend/resume, resizing, snapshots).

From what I read, the Rosetta on Linux ARM is a JIT (just-in-time) implementation. It doesn't have the performance benefits of the AOT (ahead-of-time once-time-only) translation. Interesting to see how well Rosetta translated x86_64 apps run on Linux ARM.

Re: Running intel binaries in linux vm's with rosetta

Marshalleq — Sun, 18 Sep 2022 04:27:30 GMT

I have discovered that an option has been added into the next version of UTM to utilise this capability - of course I have no idea in what way that is implemented, only that UTM linux arm VM's will utilise this capability to run x86 binaries inside them via rosetta.

There are some intriguing possibilities.

I think I will go give UTM a spin and see what I can do.

Re: Running intel binaries in linux vm's with rosetta

palter — Sat, 17 Sep 2022 12:22:22 GMT

This mechanism is part of Apple's high-level Virtualization framework. VMware uses the low-level Hypervisor framework. It's not clear that there's any way to access Rosetta from the Hypervisor framework.

Of course, if anyone can figure out how to get this functionality, it's probably the VMware team. (Yes, I'd like to see it as well.)

Running intel binaries in linux vm's with rosetta

Marshalleq — Sat, 17 Sep 2022 03:57:11 GMT

https://developer.apple.com/documentation/virtualization/running_intel_binaries_in_linux_vms_with_rosetta

Could this be integrated?

Re: RaspberryPi Guest

cyvaquero — Fri, 26 Aug 2022 13:48:42 GMT

What would be the use case?

Raspbian is just Debian for ARM tuned for RasPi hardware. I'm not sure what the benefit would be, lacking the hardware it's just Debian with a couple management scripts thrown in.

Re: Bundled uninstaller application within the product

ColoradoMarmot — Wed, 10 Aug 2022 18:02:22 GMT

Great idea, or at least have the drag to trash trigger a full uninstall.

Re: Bundled uninstaller application within the product

Technogeezer — Wed, 10 Aug 2022 16:20:18 GMT

I should add that the uninstall process I have been recommending is the complete manual uninstall process in the KB article. Not simply dragging the app to the trash.

Bundled uninstaller application within the product

Technogeezer — Wed, 10 Aug 2022 16:10:03 GMT

Given the proliferation of files outside of the application bundle that Fusion relies on, simply dragging the Fusion app to the trash is not a sufficient uninstallation process.

Fusion needs an uninstaller application that has an option to fully remove these support files and vestiges of older Fusion releases that are now manual processes in a KB article.

As a side note, the Fusion installation and upgrade process needs to be more robust. I've answered too many questions here in the forum where the solution has been "uninstall and reinstall" because something is wonky after an installation or upgrade.

Re: eGPU tb4/pci path-trough on M1 if it's posible

rootstas — Sun, 07 Aug 2022 22:38:33 GMT

Linux port for the M1 can talk to eGPUs using TB3

Re: One Click Windows ARM deployment

rootstas — Sun, 07 Aug 2022 22:31:00 GMT

Why did you duplicate records?

https://communities.vmware.com/t5/Fusion-for-Apple-Silicon-Ideas/Wizard-VM-creator-with-OS-auto-downloader-as-one-of-steps/idi-p/2921613

Re: x86_64 virtualization on AARCH64-macOS - Status changed to: Declined

Mikero — Thu, 04 Aug 2022 18:33:57 GMT

Complete OS-level Emulation is not something we plan to pursue

One Click Windows ARM deployment

bmwnyrich — Thu, 04 Aug 2022 13:03:24 GMT

Include a one-click deployable Windows ARM image similar to Parallels.

Re: Bring support OpenGL 4+ & DirectX 11/12 for Windows On ARM Guest

Technogeezer — Tue, 02 Aug 2022 01:48:48 GMT

VMware does not typically comment on the content of unreleased product as a corporate policy. So it's not necessarily a "not willing", but rather "can't comment due to corporate policy".

Consider this area a "wish list", asking VMware to implement a feature, or make a change that would benefit the product. Unfortunately they don't tell you when or even if they will implement it.

Re: Bring support OpenGL 4+ & DirectX 11/12 for Windows On ARM Guest

yfx01 — Tue, 02 Aug 2022 00:24:50 GMT

I also want to know the answer to this question, but the official seems not willing to answer this question.

Re: eGPU tb4/pci path-trough on M1 if it's posible

ColoradoMarmot — Mon, 01 Aug 2022 03:35:15 GMT

M1's themselves don't support eGPU so neat if it could, but with the host OS not providing driver support, not sure how a virtualization product might accomplish that