Sharing some Geekbench results of Parallels + W11ARM… and feedbacks needed (regarding REVIT)!!

[o’bluetouring]

Hello… sorry for the long post.

Just want to share some Geekbench results I ran yesterday on the new MBP 14 (10C/16C/16GB ram) inside the Parallels VM running W11 ARM insider preview edition. Hopefully these results will be useful for other members, and I’m also looking for some feedbacks on these results.

[Backstory]
I was planning to order a MBP with 32GB of ram to replace my current mid-2014 MBP 15.4 i7 2.5 with 16GB ram. The old laptop has been running strong and I even ran AutoCAD 2020 within Parallels/W10 during the pandemic to do some freelance architectural engineering projects, and other than some slow down when opening drawings, it actually ran pretty smooth during design. Only reason I thought about upgrading is it no longer can be upgraded to the latest OS, and I figure 32GB would be a nice bump if I plan to keep it for 7 years like I did on my current one.

But before I order the new MBP and for it to arrive in December, I want to make sure AutoCAD 2020 works with W11 ARM version first, so I decided to pick up a stock configuration MBP 14 from a local Apple Store this past weekend to test it out. The one I bought was the Pro version with 10C/16C with 16GB ram (and 1TB SSD).

I ran these Geekbench tests by assigning different number of cores and ram to the VM. As expected, the SC scores are similar and MC score gets better with more cores assigned to VM. I’ve also posted the results of my current MBP for comparison.

[Questions - REVIT Related]
I’m still trying to learn REVIT and I want to make sure this new MBP can handle it down the road. I’m no expert so I just opened up the sample project that came with REVIT 2020 and tried to render it, figured maybe that will give me an idea of how the MBP handles REVIT. I selected “Best” setting for rendering and it took about 1hr 46mins to finish the rendering. I got similar results with 4C/8GB, 6C/10GB and 8C/12GB (all within 30 seconds of each other). I’ve also attached the screenshot of the Activity Monitor when it was rendering, while having 10+ drawings opened on AutoCAD (also ran in VM).

(Q1) For those who uses REVIT, is this considered good or bad based on the rendering time? If this is not a good way to test how the machine handles REVIT, what other tests can do I?

(Q2) Is it weird that number of cores and amount of ram assigned to VM don’t affect the rendering time much?

[Other Questions]
(Q1) I don’t know how to read all the stats on Activity Monitor, but does the memory pressure and other stats look alright? Should I consider getting 32GB ram if I also want to run do some light Photoshop and video editing on the side? My wife and I might be sharing this laptop so she might install some accounting software in VM.

(Q2) I tried to merge (3) 40-min VR clips in QuickTime (MacOS side, not VM). Each clip is about 2GB but the end result is almost 20GB, and it took about 1hr to finish (when I first started the process, it estimated 2hr but it eventually finished in 1hr). Will 32GB ram speed up the process even more? I know the same merging process would take more than 3 hours to finish on my current 2014 MBP.

Again, sorry for the long post but hopefully these Geekbench results are useful.

Thanks for all feedbacks!!

 

[ahurst]

Thanks for sharing this! Could you also try running the Geekbench 5 tests again using the x64 version of Geekbench? I'd be very curious how Windows-on-ARM's translation layer on an M1 Pro stacks up against native virtualization on your older high-spec Intel MBP.

As to your questions, I don't know much about REVIT, but for Q2 the reason extra cores/RAM might have no effect is that it's a single-threaded CPU or GPU limited process, and thus it can't take advantage of the extra cores/RAM you're allocating the VM.

Also I'm no expert but that memory pressure looks fine: I'd only be concerned if it was hitting the 60% plus range on that graph or dipping into the yellow zone. I start hitting slowdowns when I get into high yellows and reds (80%+ memory pressure) on my older Intel iMac, and with a Fusion Drive it doesn't have anywhere near the virtual memory speeds a new M1 Pro system would have.

 

[o’bluetouring]

Thanks for sharing this! Could you also try running the Geekbench 5 tests again using the x64 version of Geekbench? I'd be very curious how Windows-on-ARM's translation layer on an M1 Pro stacks up against native virtualization on your older high-spec Intel MBP.

Thanks for the feedback! Yes I can help. Where do I get the x64 version of Geekbench 5? The Geekbench tests i ran was done on the Geekbench software I downloaded from their website within W11 VM environment.

 

[ahurst]

I honestly have no clue, maybe there's a preference pane or checkbox somewhere? If you can't find one, you can probably try Geekbench 4 which I don't believe has native ARM support.

 

[o’bluetouring]

I honestly have no clue, maybe there's a preference pane or checkbox somewhere? If you can't find one, you can probably try Geekbench 4 which I don't believe has native ARM support.

I don’t see any checkbox or preference for Geekbench 5 in W11 VM.

Here is the screenshot of Geekbench 5 result on the Mac OS side, which tested all 10C of the processor (instead of the 4C, 6C or 8C I tested within W11 VM).

 

[ahurst]

I don’t see any checkbox or preference for Geekbench 5 in W11 VM.

I think it's the "Architecture" dropdown menu in the bottom half: will the trial version let you select x86 or x64 instead of AArch64?

 

[o’bluetouring]

I think it's the "Architecture" dropdown menu in the bottom half: will the trial version let you select x86 or x64 instead of AArch64?

The trial version doesn’t show anything else under the pull-down menu sorry

 

[ultrakyo]

The trial version doesn’t show anything else under the pull-down menu sorry

Try this link…

https://cdn.geekbench.com/Geekbench-5.4.2-WindowsSetup.exe