Re: Challange preposed by Mindcraft

["Steve Poling" <sdp@xxxxxxxxx>]

methinks the best way to step up to the challenge is to use the benchmarks
to direct some kernel profiling (done already) and then hack the kernel to
fix anything that's found to be broken (done already). along the way,
collect on one or two web sites all the tuning info that people using *real
life* installations discover (done already).

then the best way to step up to the Microsoft/Mindcraft challenge is to
ignore the benchmark lash-up Microsoft/Mindcraft proposed, and devise a
benchmark that more accurately reflects *real life*. e.g. for the same
dollars as one might spend on a quad-xeon computer, put together a clustered
server-farm & benchmark *that* against NT & Solaris. yes, solaris, we ought
to be comparing Linux against the *best* OSes on the planet. By showing
Linux is almost as good as Solaris & FreeBSD, but better than NT, we rub
Micros~1's nose in the fact that NT is slower than a number of rivals.

Oh, and if the Linux-community benchmarks were to exploit known BSOD
conditions, that'd be a bit over the top. we really *don't* want Microsloth
to remove them. they'd make too much money selling service releases.


-----Original Message-----
From: Jeff Mayrand <jmayrand@ameritech.net>
To: klug@klug.armintl.com <klug@klug.armintl.com>
Date: Friday, May 28, 1999 6:06 PM
Subject: Challange preposed by Mindcraft


>I was curious if anyone has found out if the linux community has stepped up
>to the challenges that Mindcraft had posted on their site (IE New
>benchmarking with servers tweaked by the best engineers the linux community
>has...)  It would be a shame to balk and balk at the results and not show
>face at a second challange....
>
>
>
>
>Jeff Mayrand
>
>