Galcial slow performance of M2 Max chip for engineering software like Etabs

Does the Etabs software have a WIndows/ARM version?

If not, you will be seeing

• The overhead that Windows for ARM incurs to run Intel machine code on an ARM machine, and
• The overhead that Parallels incurs to run one ARM-based OS (Windows for ARM) within another (macOS)

I would suggest checking with your engineering software application vendors to see if they have applications that would not require this overhead.

macOS applications optimized for Apple Silicon would be best – but even Windows applications built for ARM, or macOS applications built for Intel that had to run under Rosetta 2, might be better than ones that are hitting you with a double dose of overhead.

Have you looked at the fine print of running Windows within Parallels on an Apple Silicon Mac? There are some limitations & restrictions....some which may apply to the GPU (I think some of that may have improved with a recent update to Parallels or Windows ARM, but there still may be issues).

Here is a Parallels article noting some of the limitations...the most severe for your needs is the limitation of DirectX 11.1 and OpenGL 3.1 as these will have a direct impact on GPU performance. Most likely current versions of any Windows app will be using DirectX 12+ (if the app must utilize a lesser version of DirectX even on bare metal performance is usually severely impacted):

https://kb.parallels.com/129497?_gl=1*1nxesq3*_gcl_au*MTM1MzEwMjM2Ni4xNjkzNjI3ODc4*_ga*NDU5MDkwMTU0LjE2OTM2Mjc4Nzg.*_ga_CEVYC7924W*MTY5Mzc5OTA4NS4yLjEuMTY5Mzc5OTM1OC4yNS4wLhttps://www.parallels.com/blogs/windows-11-arm-apple-m-series/

Here is an equivalent Microsoft article as well:

https://support.microsoft.com/en-us/windows/options-for-using-windows-11-with-mac-computers-with-apple-m1-and-m2-chips-cd15fd62-9b34-4b78-b0bc-121baa3c568c

General info:

https://www.parallels.com/blogs/windows-11-arm-apple-m-series/

Even on an Intel Mac, using a VM will impact performance for certain types of apps requiring lots of system resources especially involving access to the hardware like a GPU. The OS within the VM is never going to have anywhere near the same performance as Windows running on bare metal with direct access to the GPU hardware. Every call to access the GPU is being routed & translated through various layers so by the time the instruction reaches the GPU it is not going to be in an optimized state, but may be using multiple other instructions to perform what a single instruction would have achieved on bare metal.

If your high performance & resource intensive app is Windows only, then it seems like you need a laptop which can run Windows natively on bare metal. Or you need to find a native macOS app which can perform a similar function....you may need to make compromises with the latter option since some types of apps may not have any direct equivalent for macOS.

Hi! I get why you're doing what you're doing. I have the exact MAC setup with the cores and RAM. I've been using Parallels for about 7 years on many Macs. Please know that if you allocate over 50% of the cores and memory, your performance will degrade. Make sure you are running the Pro version of Parallels, and set your VM to no more than 48G of memory and 6 cores (half of your power) and give it a shot.

Jane

There have been sporadic reports of a few M-series Macs that do not wake properly from sleep. Instead, they blunder along at -- funny thing about that -- about 40 percent of the full potential processor speed.

Restart your Mac, and before it has a chance to sleep again, re-run your experiment.

If applicable, some fixes were released in MacOS 13.5.

Why would you even try running Windows apps in a virtual environment on a Mac, let alone a silicon Mac which only runs the ARM version of Windows in a professional engineering environment? Get a Windows computer to run Windows apps and use your Mac computer for work-related apps that natively run on a Mac silicon.