Hi folks. After using Parallels for a few years and checking out the interface of Fusion, I was very intrigued to possibly switch as I like the interface far better than that of Parallels. I decided to download both Parallels 11 and Fusion 8 to benchmark them and see where they stand. Results are really not what I expected given that VMWare is the industry leader in virtualization. I could immediately notice Fusion 8 was slower than Parallels 11. Here is what I did the testing on:
MacBook pro 2015 - i5/8GB/256GB
Both VMs were configured with the following:
60GB Static disk
512MB Graphics Mem
Windows High Performance Power Profile
Network Disable on Mac host
No other processes running on Mac host
The results were what I expected from use, Fusion 8 performed slower in every aspect, some aspects were very far behind Parallels actually. Disk access which is obviously very important was not good I'm wondering if anyone has any suggestions for settings I'm missing or reasons for why the performance was less than that of Parallels or reasons why anyone would go to Fusion over Parallels. I've attached the results using the latest version of Passmark. Thoughts/Info appreciated.
I haven't had a chance to test against Parallels, but I can say that I'm extremely disappointed with Fusion 8...it actually runs Windows 10 slower than Fusion 7 did and slows down even more after my MBP (2014) wakes up from sleep. I should have switched to Parallels when it came out instead of waiting and upgrading Fusion.
Graphics performance is very poor in those benchmarks, and it doesn't look like DirectX 10 is enabled. Do you have the option under Settings > Display for the VM? It should certainly be available for this year's MB Pro line, but some users seem to be having problems with the detection and can only enable DirectX 9.0EX:
I did some similar tests running a boot camp partition on both Parallels 11 and Fusion 8. I did not run benchmark tools, but real life apps. Here are some examples:
Starting Photoshop: 4.2s on parallels, 5.35 on Fusion.
Starting Visual Studio: 6.97 on parallels, 8.93 on Fusion.
Starting Delphi: 22.57 on parallels, 27.93 on Fusion.
Copy a folder with lot of small files: 9.65 on Parallels, 11.42 on Fusion
Copy one big file: 2.48 on parallels, 4.92 on Fusion
Shutdown Windows: 9.37 on parallels, 15.87 on Fusion.
The difference is clearly noticeable when working on both. The only advantage of Fusion is the possibility to move VM to Workstation or VMWare Player, and maybe a better power management. The CPU load is a bit lower when running Fusion. It could be important for people on the road. If someone has advices to improve Fusion perf, it would be great to test.
The Real Advantage of using Fusion instead of Parallels is the fact that you do not have to buy into
a "ball and chain" subscription in order to get a useful virtualization environment.
The latest release of Parallels has significantly crippled the standard version and forcing anyone
that wants to do anything useful with virtualization into a must pay forever to keep using product!!!
Another advantage with Fusion is that you do not have to purchase a license for a second machine,
which is great for us who have the need for desktop and laptop usage.
I had been a very satisfied Parallels customer for years and have returned to Fusion due to this recent
marketing scheme at Parallels. I will accept some slower performance (which I have not yet seen in real
world usage yet of Fusion) for a product that won't turn into a pumpkin at midnight!
It's a good argument. But if you buy every Fusion upgrade to take advantage of the latest material support, you are also renewing each year. I agree, the difference is that you have the choice to renew or not.
Does it make a difference running Fusion 8 on El Capitan? VMWare seems to advertize lot of speed optimizations for new OSX version.
I had also been a satisfied Parallels customer, but after I discovered that I had to "subscribe" to the Pro version just to get all the features I have had for years, I decided to check out Fusion again. I ended up buying it and am happy with it. I really like that I only need one copy and that covers my desktop and my laptop. I really don't notice much of a speed difference between the two (although I never tried Parallels 11). I also tried converting a VM, as well as loading one fresh. I was really surprised at how well the converted one ran.
I'm curious how many users paid $49 for the Parallels Pro version thinking it was an upgrade only to be surprised next year when they receive a bill or a hit to their credit card when their subscription is up for renewal. Will be interesting to read their forums next summertime around this time. Bet there will be a lot of angry people who find out that if they don't renew, they will basically have their VMs disabled. The whole subscription concept is not that obvious if you aren't reading the page carefully and think this is just another year and another upgrade.
Sure. But since 2008, it only happened once. I will continue my testing. But I have the feeling that Fusion 8 has reduced the gap, but is still a bit behind for performances and boot camp partition support.
I am still poking in now and then in the Parallels forums, and people in several threads
that paid for the Pro version thinking it was a good deal had no clue that they were buying into
a ball and chain.
I also, am not really seeing much in slowdown compared to Parallels 10. Then again, I was not extensively
working in Windows 10 so I will see as time goes on. Most of my work in the Windows world is more CPU
and file intensive rather than graphics intensive.
I've just finished a small trial of Parallels to see if I would benefit from moving back there (I have an old Parallels license that I could upgrade).
I did notice that my VM, after converting, was snappier in Parallels but there wasn't a big difference. I also think I must have got used to Fusion's way of working, as there were a handful of small features I missed when using Parallels.
As you say, I suppose it depends on what you use your VM for. Historically Parallels has always been better for running Windows games, but I think it's pretty close for a lot of other applications.
I actually noticed a greater speed-up reinstalling my Windows 10 VM from scratch, after a previous upgrade from Windows 8.1.
Interestingly the VMware Fusion 8 seems to easily outperform Parallels Desktop 11 in quite a lot of cases, especially those GPU bound.
The worse performance most of you mention here seems to be caused by the often significantly worse results in the virtual machine management and file transfer tests.
Actually the real conclusion of the OP's vs TekRevue's benchmarks seems to be that you're only going to see the benefits of Fusion 8's graphics improvements if you have a dedicated GPU. 2014-2015 Macbooks without one (i.e. the majority) perform significantly worse than Parallels, and don't seem to support the advertised DX10 or OpenGL 3.3 goodness at all. This perhaps needs a fuller explanation from VMWare - or at least a more detailed requirements list.
I'm running virtualization using the Boot Camp partition. In that case, I'm not getting the same results at all. Parallels is much quicker and use less CPU. Did someone try to convert a Boot Camp partition (instead of running directly from the partition)? Did you see improvements (in perf not talking about suspend or mirroring folders) over running directly from Boot Camp partition?