I wouldn't doubt that there are distros of Linux that could improve performance quite a bit, as well as the little tricks (xinit). I was talking with a friend last night about the benchmark, and he suggested a couple things I could try out. I figure I'll probably do my next comparison in about a month, after a little more learned. I don't know if I'll do multi-distros as I don't have *that* much time on my hands. I don't really know how I would go about benchmarking Linux due to the modular design of it, and how scalable it is. One has to wonder - should the benchmark be performed on a standard workstation install, or on a highly customized build? Didn't MS cut Windows2K down to about 32KB of code for the XBox? This would seem to prove that Win has the ability to be optimized, and probably runs really fast with so little code. Would a completely custom Linux install be "fair" to benchmark against WinXP? I guess it is MS's fault for not making an OS that can be cut down, but how many people would actually use such a customized install of Linux? If we (the Linux community) want to attract new users (to the advantages of Linux), should we be demonstrating a supar-leet install, or something more for the "average" user? I guess these are all judgement calls that are left up to the benchmarker, which makes my life just a little more difficult.
Last edited by AdamDunn; 04-14-03 at 09:54 PM.