On Fri, Dec 05, 2008 at 01:09:36PM +0100, Wojciech Puchar wrote: > >a Windows system it's the virtual memory management, with the same > >amount of RAM Windows swaps a *lot* more. > > it may be not VM subsystem but memory usage of windoze software. or both. > > again - it's too different to be benchmarked
There's no reason one cannot generate benchmarks comparing the two. You just have to choose your benchmark tasks carefully. Of course, microbenchmarks are usually suspect no matter what systems you're testing -- whether it's FreeBSD vs. MS Windows, OpenBSD vs. Linux 2.6.x, or Ruby 1.9 vs. Python 3.0, there are always ways to arrange your benchmark tests to favor whatever you want to favor. That doesn't change the fact that FreeBSD vs. MS Windows benchmark tests can be every bit as (un)useful as any other benchmark tests. They're not "just too different". -- Chad Perrin [ content licensed PDL: http://pdl.apotheon.org ] Quoth C. Hoare: "Two ways of constructing software: (1) make it so simple that there are obviously no bugs, (2) make it so complicated that there are no obvious bugs. Making it simple is far more difficult."
Description: PGP signature