My Experience - Bootcamp vs Parallels

[Toronto1970]

I have been testing out both Parallels and Bootcamp on my new MacBook (Black Core2Duo 2GHz with 2Gb RAM). I work for a company that develops Windows financial modeling software involving many floating point operations, and a large volume of results written to disk.

With this application, the main drawback of Parallels is that it can only recognize one core. Under Bootcamp, Windows can see both cores, and the software can distribute the calculations and almost halve the run times.

I also found that the runtimes improved greatly under Parallels when I switched from a variable sized disk to a fixed size.

The runtimes for a sample run were (times in seconds):

147 - Bootcamp running with 2 cores
275 - Bootcamp running with 1 core only
407 - Parallels with fixed size disk
514 - Parallels with variable size disk

The Bootcamp times are (as expected) similar to runtimes seen on native Windows machines with similar hardware.

Parallels (with fixed size disk) is running about 50% slower than Bootcamp under 1 core. I've been trying to find other Parallels settings which would improve runtime more, but I haven't been able to find any.

 

[Westside guy]

Thank you for posting that.

I find that Parallels works well for the non-intensive stuff I need Windows for (basically debugging Web issues involving Internet Explorer or Front Page for work, and of course some simple games for home ). With those sorts of apps, I don't notice the performance hit at all.

But it's important for people to note that there IS a significant performance trade-off that will impact demanding apps - like yours. People shouldn't have unrealistic expectations regarding virtualization; especially not at this early stage in the game.