[plug] Distro Day !!!

[indy at THE-TECH.MIT.EDU]

On Thu, Jul 03, 2003 at 09:22:08AM +0800, Ben Jensz wrote:
> >
> Just one comment on that.  It might be close, but it won't be the same 
> as having a 10% slower/faster CPU, as you'll be adjusting the Front-Side 
> Bus (FSB) as the multiplier is always locked.  Whereas a real 10% (or 
> whatever step) up CPU will have a higher multiplier, rather than FSB.  
> Maybe an idea to note that in the article.


Absolutely!

As has been mentioned, the idea that led us to this test is to
look into the optimization differences between Gentoo (full customization),
Mandrake (i686) and Debian (i386) optimizations to see how it affects
the user experience.

Obvious "hard data" tests are stuff like kernel compiles, and I
hope to rig up a test script for the Gimp that we can time.

More suggestions for tests needed! Show off your l33tness,
come up with good ways to compare the different distros' performance!

Anyone think of some good OOo tests?

Obviously, user experience is another part of this.
Hopefully we'll get enough people to try out the machines to
average their impressions and get a good feeling for the snappiness.


In that regard, Mandrake and Gentoo experts, out of KDE and Gnome,
is one "package" (and yes I know the word package is slightly inappropriate)
known to be better implemented on your distro? We don't want to
queer the tests by ignorance..


Indy.


-- 
Indranath Neogy
<indy at the-tech.mit.edu>