[nSLUG] 2.6 kernel performance

Donald Teed donald.teed at gmail.com
Tue Jun 21 16:51:09 ADT 2005

On 6/21/05, Ben Armstrong <synrg at sanctuary.nslug.ns.ca> wrote:
> The trouble is that many people don't really consider the real goal of
> optimization.  They just have an obsession with their systems scoring
> well on benchmarks.

I don't have that obsession.  However, my usual
experience with hdparm is that if there is
a 40% change there, you can notice the difference
in end user experience.

> I don't believe for a minute Stephen was saying
> "accept mediocre performance".  What I heard was "don't waste effort
> trying to tweak things that don't give real gains".

Well, what he literally said was "don't waste you time",
and "there isn't much that you should bother doing".

In my opinion, getting 100% of the 2.4 kernel performance
is worthwhile.  As well, I want to understand what
is happening.  I'm not about to just shrug it off and
think "oh well, the computer wants to run slower today".

I am the master and commander of this rig and I
aim for control.  Isn't that natural, or do the
rest of you let chaos reign?

> But tweaking for the sake of watching
> benchmark numbers go up (or down) is meaningless.

I agree with that.  I've seen so much written by Maximum PC
types about pushing for performance increases, and
they do the Time Magazine trick of showing graphs
that don't start at zero to exaggerate the changes.

I read years ago about a study where a group of non-savvy
PC users had their rigs upgraded while they were out at
lunch, and no one told them.  The subject of the
experiment was to determine when an average person 
notices, all on their own, that the system is faster.
The study concluded that the average user does not report
the speed up until there is a 2x performance factor
increase.  I've often used that as my formula on
planning upgrades.


More information about the nSLUG mailing list