performance: boot camp vs. virtual drive

Is there a performance difference between running Windows (XP, Vista, or 7) off a Boot Camp partition vs. doing so from a regular Parallels virtual disk? If so, which is faster?

Thanks!

 

Performance off the Parallels regular disk is considered slightly faster.
Also, there is a number of limitations for a Parallels virtual machine that uses the Boot Camp Windows partition either as a bootable volume or as a data disk:

* it cannot be suspended or paused
* such a virtual machine cannot have snapshots and the Safe Mode feature cannot be enabled for it
* compression or compacting cannot be performed
(you can see this in overview).

Using virtual machine based on a Boot camp partition, on the other hand, has the advantage that you can boot to Windows natively whenever you would need it.

 

I have heard there are performance speed issues

I would like to know how well running XP under Parallels works compared to boot camp when you look at a gaming?

I have read that parallels reduces some of the operational capacity of the mac by running as a virtual machine?

Has this been resolved for version 5?

 

Could you please clarify one more time what would you like to compare?

a) XP running as a virtual machine in Parallels vs. XP running on a Boot camp (no Parallels involved)

or

b) XP running as a virtual machine in Parallels vs. XP running as a virtual machine (in Parallels) based on a Boot camp partition.

 

Comparison

To clarify:

I am looking for performance comparisons of XP in parallels vs XP in Boot camp (no Parallels involved), as it is my understanding that XP in parallels being a virtual machine reduces some of the performance capabilities of XP since it is running with the Mac OSX, and both OS share the overall operating performance of the machine, versus boot camp which allows XP to access the full operating performance of the Mac.

On a side note, what would be the benefit of running XP in Parallels based upon a boot camp partition?

Thanks.

 

Thanks. In that case I can confirm, that running Windows XP natively, (on a Boot Camp partition) would indeed have more performance since Mac OS is not involved and thus all CPU, memory, etc. goes to Windows only.

Parallels in this scenario provides the advantage of running Windows while Mac OS is running at the same time.
Virtual machine based on a Boot camp partition is more like a hybrid between both. You can use Windows from Mac side mainly, but use the Boot Camp if you need, for example, to run the game that would not run in virtualized environment (i.e. having some specific hardware requirements).

 

outside of the obvious restrictions of no snapshots, etc...

this question here:

b) XP running as a virtual machine in Parallels vs. XP running as a virtual machine (in Parallels) based on a Boot camp partition.


I'd like more information - ie. is running XP as a VM based on boot camp a better selection ?

 

You only want to use a Boot Camp partition if you want to boot your Mac into Windows and use the same installation of Windows as you do in Parallels.

Some other advantages of using a Boot Camp partition:
1) You can move the drive to a real PC.
2) You have access to the disk media and not just the files. This wouldn't be an advantage if Parallels Mounter could mount a virtual disk as media (like Disk Utility) instead of as a file system.
3) Parallels Mounter can't mount Linux disks. It could if it could mount a virtual disk as media.

 

Performance numbers

I am very interested in knowing the performance statistics between the difference of running XP running as a virtual machine in Parallels 5 vs. XP running as a virtual machine (in Parallels 5) based on a Boot camp partition. Can you also discuss whether upgrading to Windows 7 and/or Parallels 6 has made any improvements. My main reason for asking this question is that I run a very memory intensive 3D modeling package sold by Autodesk named Revit Architecture. It is a memory and processor hog. The recommended system requirements are Windows 7 64-bit with 8 GB RAM but the required system requirements are Windows OS with min. of 3 GB. Currently I am running XP with 4 GB on my MBP. The files we deal with can be anywhere in size from 25 MB to 125 MB and we do a lot of rendering. Running Windows in Bootcamp is suppose to give the best performance but I would prefer to run using Parallels. It was suggested by Apple to create the VM partition of Bootcamp so that I only needed to have 1 install of windows and Revit to maintain, which I like very much. Revit has a lot of 3rd party apps and patches that I would like to be able to install once. I have been searching for an answer on specific numbers but can't find. I understand some of the advantages but don't think that those are critical to running Revit. I want speed and performance.

Are there tools I can use to measure the results other then a stop watch?

Also, one last question about the video card, does parallels allow you to assign a different video card instead of the Parallels Video Adapter? I am getting some odd display issues with materials.

thanks,
ken

 

Ken,

I have been using Revit with Parallels for a couple of years now. This year more so than in the first year. My hardware comprises a 2010 27" iMac quad core i5, 8GB RAM, ATI Radeon HD 5750, 1GB VRAM. I am running Revit Architecture 2011 on Windows 7 32-bit but I will move soon to Windows 7 64-bit as the 32-bit version is limited in how much memory it can use (i.e. <4GB). The VM is configured to use 2 cores and 4GB of RAM (no point assigning more with 32-bit Windows as it won't be used). The VM is set to provide maximum Windows performance, the idea being that if I'm running the VM that means that I'm doing mostly Revit work. When dealing with a complex Revit model or when rendering I tend to assign more cores to the VM.

Generally I am very satisfied with this setup as it means that I never have to leave my Mac. I am able to run Photoshop and InDesign natively while running Revit and there is no noticeable loss of responsiveness on the Mac. Revit does not run as smoothly in the VM as it does natively on a Windows PC. I believe that this is mostly due to the way that the gfx hardware is exposed to the host OS by the VM.

I would not recommend creating a boot camp partition. I don't believe that the performance gains justify it and you will lose a lot of the advantages of having a VM. For example, I have create my VM with just Windows and Revit on it. As I add more patches, plugins etc, I tend to make a copy of the whole VM first and then add new features to the copy. This way if anything goes wrong I simply delete the VM file and go back to the previously working VM (this manual procedure is slightly different from Parallel's snapshotting function but I prefer it as it gives me more flexibility). I also have other VMs, all configured differently with different software—when it comes to Windows, less is definitely better.

To measure your performance just time a complex render (tell Revit not to close the render window). Do it once on a Windows PC and then again on your Mac. Assuming that both are comparable machines this simple test will give you an idea of the performance hit imposed by the VM. I don't believe that it's much, especially for complex renders.

Also, Windows 7 has a built-in benchmarking tool. Right click the "Computer" icon on the desktop and select "properties".
In the dialog, under the "System" section click the "this system has not been rated" link (or something like that). Windows will then display new dialog giving you the option to rate your system. It tests CPU, memory, graphics, gaming graphics and disk performance. The scale is from 1.0 to 7.9 (7.9? What the hell? You gotta love Microsoft). Anyway, my system scores are:

CPU: 7.3
Memory: 5.5
Graphics: 6.0
Gaming Graphics: 6.0
Disk: 5.9

As you can see disk and memory are the lowest scores as is to be expected when running a VM.

cheers,
-tomek

 

I'm interested to know if anything has changed since this thread was originally posted back in 2006. Is it still faster to run a VM off a virtual Parallels disk than a real (Boot Camp) partition? If so, how significant is the difference? And, does the OS in question (XP, Windows 7, Linux) affect the answer?

 

Same here... any changes?