Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Parallels or VMWare Fusion
- Thread starter VladM
- Start date
- Sort by reaction score
Parallels is mainly eye-candy and doesn't offer the stability of Fusion. Fusion is rock solid product whereas Parallels scared me a few times because it turned up errors. Use Fusion, it's cheaper to buy (think of upgrade fees on Parallels) and upgrade in the future. Parallels keeps charging and quality keeps slipping.
I was using VMWare 2.0 to run Access 2007 in WinXP to batch load thousands of records for some research I'm doing, and when Parallels came out with 4.0 I decided to give it a try. I've used every version of both these programs since they came out because I'm always looking for something better. 
Anyway, Parallels is loading these records about 3x faster than VMWare on my MBP. I've configured both virtual machines with the same amount of memory and they're both using only one virtual processor so I don't know why it would be so much faster but it's consistent. Since that's the only thing I use Windows for on my Mac at the moment, I'm currently using Parallels exclusively. The new interface in 4.0 is very nice too, and I find it handles Expose better than VMWare. With a number of windows open, they both show some relics when you activate Expose, but Parallels much less so than Fusion. Of course, YRMV
Anyway, Parallels is loading these records about 3x faster than VMWare on my MBP. I've configured both virtual machines with the same amount of memory and they're both using only one virtual processor so I don't know why it would be so much faster but it's consistent. Since that's the only thing I use Windows for on my Mac at the moment, I'm currently using Parallels exclusively. The new interface in 4.0 is very nice too, and I find it handles Expose better than VMWare. With a number of windows open, they both show some relics when you activate Expose, but Parallels much less so than Fusion. Of course, YRMVMy choice is Parallels
I tried three virtualization packages - the Sun VirtualBox, VMWare Fusion, and Parallels. After hours experimentation and testing, I have decided on Parallels. Here are my findings.
VirtualBox is just not for prime time, lots of graphics corruption right out of the box. For instance, when you switch from full screen Windows to the OS X desktop, some of the Windows desktop would be left on the screen, obscuring the OS X screen. The interaction between the Windows desktop and the OS X desktops is not as seamless as that provided by Fusion and Parallels, and numerous other little things that detract from the overall user experience. So even though VirtualBox is free, it's not worth what I need to put up with.
VMWare Fusion and Parallels are actually very similar in how they work, and I don't see any difference in terms of performance. But after extensive usability testing I have found that Parallels gets more of the little things right. And it's the little things, the attention to details, that differentiates the excellent from the merely good.
For example, when you maximize a Windows application in Windows full screen mode under Fusion, the application would maximize to cover the Taskbar. Under Parallels, however, the application correctly maximizes to cover only the client area of the desktop not cover by the Taskbar. I know VMWare must know how a correct Windows application should behave, and I attribute this behavioral error to either that they were lazy or that they hoped people don't notice. Neither speaks well for the code that runs underneath the surface.
Another example, when you task switch using command-tab in OS X, you can tell which applications are Windows native applications because Parallels would superimpose the two vertical bars Parallels logo over the Windows application icons. Parallels gets credit for extra effort. Fusion on the other hand not only does not show you which applications are Windows apps, but cmd-tab is simply not implemented correctly in virtual Windows and thus task switch is almost impossible when you're not in unity mode (Windows and OS X apps side by side). Very bad.
Yet another example, you can access the entire collection of Mac applications right from the Windows desktop because Parallels sees fit to create shortcuts to all the Mac application right on the Windows Start menu and you can of course duplicate those short cuts anywhere in the Windows environment.
Example number four, you can control the boot sequence of the virtual machine using the Parallels application menu, while with Fusion you need to get in to the virtual machine's BIOS to do that - and the BIOS screen flashes by so fast you need to manually edit the configuration file to slow down the welcome screen so you can actually see which keys to hit to enter Setup or choose the boot device. This is another example of lazy application development.
And not to belabor the point, example number five, VMWare only emulates multiple processors, NOT MULTIPLE CORES, so that Core 2 Dual you got there is useless if you run XP Home which only supports a single processor.
Now I don't know about you, but speaking as a developer, I think thoroughness and thoughtfulness on the part of the user interface is a good indication of the amount of effort that has gone under the surface to make the application work right.
As far as Parallels' performance is concerned, this is what I found. Now I can't speak for gamers, as I don't play games on my laptop. But Parallels is able to run industrial strength servers and development environments on my 2.0GHz, 2GB dual core MacBook faster than my 1.7GHz, 2GB single core Dell. So it definitely does what I want it to. And if you only run Office or Quicken, then you have nothing to worry about.
I tried three virtualization packages - the Sun VirtualBox, VMWare Fusion, and Parallels. After hours experimentation and testing, I have decided on Parallels. Here are my findings.
VirtualBox is just not for prime time, lots of graphics corruption right out of the box. For instance, when you switch from full screen Windows to the OS X desktop, some of the Windows desktop would be left on the screen, obscuring the OS X screen. The interaction between the Windows desktop and the OS X desktops is not as seamless as that provided by Fusion and Parallels, and numerous other little things that detract from the overall user experience. So even though VirtualBox is free, it's not worth what I need to put up with.
VMWare Fusion and Parallels are actually very similar in how they work, and I don't see any difference in terms of performance. But after extensive usability testing I have found that Parallels gets more of the little things right. And it's the little things, the attention to details, that differentiates the excellent from the merely good.
For example, when you maximize a Windows application in Windows full screen mode under Fusion, the application would maximize to cover the Taskbar. Under Parallels, however, the application correctly maximizes to cover only the client area of the desktop not cover by the Taskbar. I know VMWare must know how a correct Windows application should behave, and I attribute this behavioral error to either that they were lazy or that they hoped people don't notice. Neither speaks well for the code that runs underneath the surface.
Another example, when you task switch using command-tab in OS X, you can tell which applications are Windows native applications because Parallels would superimpose the two vertical bars Parallels logo over the Windows application icons. Parallels gets credit for extra effort. Fusion on the other hand not only does not show you which applications are Windows apps, but cmd-tab is simply not implemented correctly in virtual Windows and thus task switch is almost impossible when you're not in unity mode (Windows and OS X apps side by side). Very bad.
Yet another example, you can access the entire collection of Mac applications right from the Windows desktop because Parallels sees fit to create shortcuts to all the Mac application right on the Windows Start menu and you can of course duplicate those short cuts anywhere in the Windows environment.
Example number four, you can control the boot sequence of the virtual machine using the Parallels application menu, while with Fusion you need to get in to the virtual machine's BIOS to do that - and the BIOS screen flashes by so fast you need to manually edit the configuration file to slow down the welcome screen so you can actually see which keys to hit to enter Setup or choose the boot device. This is another example of lazy application development.
And not to belabor the point, example number five, VMWare only emulates multiple processors, NOT MULTIPLE CORES, so that Core 2 Dual you got there is useless if you run XP Home which only supports a single processor.
Now I don't know about you, but speaking as a developer, I think thoroughness and thoughtfulness on the part of the user interface is a good indication of the amount of effort that has gone under the surface to make the application work right.
As far as Parallels' performance is concerned, this is what I found. Now I can't speak for gamers, as I don't play games on my laptop. But Parallels is able to run industrial strength servers and development environments on my 2.0GHz, 2GB dual core MacBook faster than my 1.7GHz, 2GB single core Dell. So it definitely does what I want it to. And if you only run Office or Quicken, then you have nothing to worry about.
I regularly move VMs between Fusion and Workstation with no issues (aside from needing to make sure that the Tools are up to date).
Interesting post.
VMware does not do this on my set up.
When running in full screen mode, the Windows task bar shows per your preferences, either always showing or when you mouse over it.
Most interesting. I am wondering why there is a difference?
Happens sometimes but not all the time. Have you tried switch back and forth between full screen and unity? I have already uninstalled Fusion so I am not going to try to find out the exact circumstance when it happens.
Thanks for the response.
I usually run VMware in single window and full screen mode. If I have an external monitor, I just throw Windows XP over on the external and run it full screen.
I tried switching back and forth between all three today, but everything worked okay. I am running Tiger (10.4.11) and VMware Fusion (1.1.3).
Anyhow, enjoyed your post and the details that you provided. I use both Parallels and VMware. I flip flop on which one I like best these days. Currently if I had to make a decision, I would probably choose VMware, but either is okay for my needs.
Definitely agreed with you on VirtualBox. Sure it's free but it leaves a lot to be desired. I figure if you are going to spend the $ on Windows XP, you might as well have a good application to run your virtual machine.
...great info about these programs - was in the store yesterday trying to figure them out...
Guess it seems like VM Fusion would be a little easier to setup for a "beginner"...
Guess it seems like VM Fusion would be a little easier to setup for a "beginner"...
I have used all 3 main players being discussed here, Virtualbox, parallels and fusion.
Virtualbox is great for the price .... ie free .. covers the basics and a few nice bits too ... great if you just need the occasional bit of software to fire up on windows or wanna play with linux etc
Parallels i started with v3 and it was ok ... upgraded to early v4 on release and promptly got a refund of the upgrade price and went to fusion2
I would say however, that if parallels have sorted out their early v4 issues then it might be a nice product but the initial release totally felt beta and many many customers were very unhappy
Fusion2 is decent, i use vmware elsewhere (esx and workstation at work) so i've always been a fan of vmware's stuff... fusion performs well and i don't often reboot the vm's cos mainly i suspend them.
Virtualbox is great for the price .... ie free .. covers the basics and a few nice bits too ... great if you just need the occasional bit of software to fire up on windows or wanna play with linux etc
Parallels i started with v3 and it was ok ... upgraded to early v4 on release and promptly got a refund of the upgrade price and went to fusion2
I would say however, that if parallels have sorted out their early v4 issues then it might be a nice product but the initial release totally felt beta and many many customers were very unhappy
Fusion2 is decent, i use vmware elsewhere (esx and workstation at work) so i've always been a fan of vmware's stuff... fusion performs well and i don't often reboot the vm's cos mainly i suspend them.
Well I reinstalled Fusion and got some more info.
The reason I reinstalled Fusion was that Parallels does not work with Solaris 10 release Nov2008. So I gave Fusion a try, as I have had luck running Solaris using VMware's Windows products. Fusion does indeed works with Solaris. Since I had it installed, I imported the latest Parallels image I had and did a couple more comparisons.
I compared the startup time of an application that I use often. The startup time includes both application load time and workspace preparation. When started up cold, the time on Parallels was 1:50. The time on Fusion was 1:40. When started up warm, meaning if I close the app and restart it right away, the time on Parallels was 15 seconds, while on Fusion it was 30 seconds. Parallels is thus about 10% faster, which is just about what MacTech magazine found.
I also compared the memory food print of both VMs. The size of the VMs are 2048KB. My Macbook has 4GB of RAM. With Parallels, with XP running in the background but no application loaded, the wired memory foot print is about 1GB. With Fusion, the foot print is about 2.3GB. Now this includes OS X. So what it shows is that Fusion allocates all 2GB of the memory of the entire VM all at once, such that it sits in the memory space like a brick, without it being able to page out to disk. This is bad. Parallels takes up about 700KB bring idle in the background, and thus is more practical to be kept open and ready even when you are not using a Windows app at the moment.
I also confirmed the Windows maximizing to cover the task bar behavior. This is really really bad because when I work in Windows full screen I tend to run all apps full screen, and having no access to the Start button when I got three four full screen apps running is extremely annoying. And I don't like the task bar to auto-hide either.
I also found that Parallels has the same bug. However, it happens very rarely, only when you switch from window mode to full screen mode and the window is bigger than your display, which just doesn't happen often.
And again that cmd-tab behavior in Fusion where in unity mode it does not allow you to cycle through the Windows apps. What kind of unity is that??
And not to mention that whenever you attach a USB device in Fusion you need to go through the menu options to connect it to your VM instead of having a dialog box popping up to ask you if you want to connect it to OS X or the VM the way Parallels does. Of course you can predefine what device gets attached to where, but sometime you really do need to make an on-the-fly decision every time and there is no way around it.
I use VMware on Windows to host Solaris and Linus. I wish they have a better product for the Mac so I can standardize on one VM platform.
The reason I reinstalled Fusion was that Parallels does not work with Solaris 10 release Nov2008. So I gave Fusion a try, as I have had luck running Solaris using VMware's Windows products. Fusion does indeed works with Solaris. Since I had it installed, I imported the latest Parallels image I had and did a couple more comparisons.
I compared the startup time of an application that I use often. The startup time includes both application load time and workspace preparation. When started up cold, the time on Parallels was 1:50. The time on Fusion was 1:40. When started up warm, meaning if I close the app and restart it right away, the time on Parallels was 15 seconds, while on Fusion it was 30 seconds. Parallels is thus about 10% faster, which is just about what MacTech magazine found.
I also compared the memory food print of both VMs. The size of the VMs are 2048KB. My Macbook has 4GB of RAM. With Parallels, with XP running in the background but no application loaded, the wired memory foot print is about 1GB. With Fusion, the foot print is about 2.3GB. Now this includes OS X. So what it shows is that Fusion allocates all 2GB of the memory of the entire VM all at once, such that it sits in the memory space like a brick, without it being able to page out to disk. This is bad. Parallels takes up about 700KB bring idle in the background, and thus is more practical to be kept open and ready even when you are not using a Windows app at the moment.
I also confirmed the Windows maximizing to cover the task bar behavior. This is really really bad because when I work in Windows full screen I tend to run all apps full screen, and having no access to the Start button when I got three four full screen apps running is extremely annoying. And I don't like the task bar to auto-hide either.
I also found that Parallels has the same bug. However, it happens very rarely, only when you switch from window mode to full screen mode and the window is bigger than your display, which just doesn't happen often.
And again that cmd-tab behavior in Fusion where in unity mode it does not allow you to cycle through the Windows apps. What kind of unity is that??
And not to mention that whenever you attach a USB device in Fusion you need to go through the menu options to connect it to your VM instead of having a dialog box popping up to ask you if you want to connect it to OS X or the VM the way Parallels does. Of course you can predefine what device gets attached to where, but sometime you really do need to make an on-the-fly decision every time and there is no way around it.
I use VMware on Windows to host Solaris and Linus. I wish they have a better product for the Mac so I can standardize on one VM platform.
The last time I tried it was when Fusion was 1.0. Could be that workstation was not aware of the VM file format at the time.
Is there a function like adjust host's computer in full screen mode for vmware? Every time I try to play the game age of empires II in vmware it stalls, yet, when I enable this in parallels it does it correctly
I have Fusion 2.0 & Parallels 4.0. And use them both. I just recently picked up Parallels 4.0 after all the results of it being significantly faster, but Fusion's been my main VM software. I can't say that i've seen a tremendous difference, but it seems that Parallels' got a lil more pep in it. On the flip so, I'm so used to Fusion's keyboard shortcuts and what not that it's been a bit of an adjustment.
Honestly, the average user can't go wrong with either one. They both deliver and get the job done. Some people will say A is better than B, but opinions are like ******s....everyone's got one
Honestly, the average user can't go wrong with either one. They both deliver and get the job done. Some people will say A is better than B, but opinions are like ******s....everyone's got one
Running Solidworks in VM & Parallels
I have both Parallels 4 and Fusion 1.1.3 (never installed 2.04)
Switched to VM Fusion a year ago because it's so much faster
MacPro 2.66 dual core
6GB Ram
OSX 10.4.11
I'm hesitation to switch to VM Fusion 2.04 as I can not afford to have it not working ...
- I run Solidworks 2007 on it every day
This actually works pretty well (verses running it under Parallels where it never really seemed usable ..)
Reading performance data & blogs on to internet it seems parallels should be at least comparable when it comes to speed
So whats the trick to get parallels up to speed ? - does it run faster under X 10.5 ?
And what about VM Fusion 2.04? - any benefit for me to install the update ?
Any comments are appreciated
Georg
I have both Parallels 4 and Fusion 1.1.3 (never installed 2.04)
Switched to VM Fusion a year ago because it's so much faster
MacPro 2.66 dual core
6GB Ram
OSX 10.4.11
I'm hesitation to switch to VM Fusion 2.04 as I can not afford to have it not working ...
- I run Solidworks 2007 on it every day
This actually works pretty well (verses running it under Parallels where it never really seemed usable ..)
Reading performance data & blogs on to internet it seems parallels should be at least comparable when it comes to speed
So whats the trick to get parallels up to speed ? - does it run faster under X 10.5 ?
And what about VM Fusion 2.04? - any benefit for me to install the update ?
Any comments are appreciated
Georg
Personally, I like VMWare Fusion, as others have undoubtedly said, it's up to you, user preference, I think you can try them both out, see which one you like better and go with that!
I too was at version 1.1.3 for the longest time, I JUST now updated it 2 days ago to 2.0.4, and it works flawlessly, I really like it, the interface has got a lot better, plus some other features added. I'd recommend upgrading, I was wary at first, but I'm glad that I did!
Mac Pro
8-core 2.8ghz Xeon
10gb
3.32tb
10.5.7
I have both Parallels 4 and Fusion 1.1.3 (never installed 2.04)
Switched to VM Fusion a year ago because it's so much faster
MacPro 2.66 dual core
6GB Ram
OSX 10.4.11
I'm hesitation to switch to VM Fusion 2.04 as I can not afford to have it not working ...
- I run Solidworks 2007 on it every day
This actually works pretty well (verses running it under Parallels where it never really seemed usable ..)
Reading performance data & blogs on to internet it seems parallels should be at least comparable when it comes to speed
So whats the trick to get parallels up to speed ? - does it run faster under X 10.5 ?
And what about VM Fusion 2.04? - any benefit for me to install the update ?
Any comments are appreciated
Georg
I too was at version 1.1.3 for the longest time, I JUST now updated it 2 days ago to 2.0.4, and it works flawlessly, I really like it, the interface has got a lot better, plus some other features added. I'd recommend upgrading, I was wary at first, but I'm glad that I did!
Mac Pro
8-core 2.8ghz Xeon
10gb
3.32tb
10.5.7