I am using VMware Server 1.0.7 on Windows XP SP3 at the moment to test software in virtual machines.
I have also tried Microsoft Virtual PC (do not remeber the version, could be 2004 or 2007) and VMware was way faster at the time.
I have heard of Parallels and VirtualBox but I did not have the time to try them out. Anybody has some benchmarks how fast is each of them (or some other)?
I searched for benchmarks on the web, but found nothing useful.
I am looking primarily for free software, but if it is really better than free ones I would pay for it.
Also, if you are using (or know of) a good virtualization software but have no benchmarks for it, please let me know.
From my experience of Parallels and VMware (on the PC and more extensively on the Mac) the difference between any 2 competing versions of the software is usually quite small and often ‘reversed’ in the next releases.
I never found Parallels to be much faster (or slower) than VMware – it often would be a case of the state of the VM I was running, the host machine itself and the app(s) I was running within the VM. If VMWare brought out a new release which did something faster, you could be sure that Parallels would improve their performance in that area in the next release, too.
In the end I settled on VMWare Fusion and the key reason for this was just that it played nicely with VMware Workstation on the PC. I have trouble taking Parallels VMs from the Mac to the PC and back again, and this worked fine on VMware. Finally, though this is less of a concern, I was unhappy that sometimes it felt as if Parallels would release a version without proper regression testing – you’d get the up-to-date version and find that networking was suddenly unexplicably broken until they released another patch a few days later. I doubt this is still the case but VMware always felt a little more ‘in control’ and professional to me.
I’d go for a solution that you can get running in a stable fashion on your PC, that is compatible with your other requirements (such as your co-workers’ platforms and your overall budget). You can waste your lifetime trying to measure which one is faster at any given task!
One other thing – it’s worth checking the documentation that comes with the software, and any forums etc, before making judgements about performance. For instance, in my experience throwing huge amounts of ram at your VM (at the expense of free ram in the host system) does NOT automatically make it faster; better to split the ram up evenly, and certainly keep an eye on any recommended figure. In VMware, that recommended figure is a good guide.