[Pdx-pm] RE: [Pdx-pm] Software development and The Rules

Tyler F. Creelan creelan at engr.orst.edu
Wed Oct 16 20:11:50 CDT 2002


> 	Of course, there's still some point where you have to wonder.
> A modern pentium 4 processes what, 2.5 billion instructions per second?
> If it takes Windows two minutes to boot, how many of the billions of
> instructions were pure bloat?
>

Most of the boot-up sequence consists of loading pages into memory from
disk or initializing devices, so the cpu is largely idle. Hence, Window's
two minute (or 5+ minutes in some cases) boot time derives more from the
extra time needed to read bloated code off the disk than execute it. :)

Interestly enough, MS is trying to address the problem in new versions of
Windows with "informed prefetching" - storing the page fault sequence and
throwing it deterministically on every boot.

You'd think programmer time would be better spent making the
initialization sequence more modular, reducing the need for frequent
reboots like unix-based systems. Instead, they just try to make it reboot
faster. :)


Tyler

-----------------------------
Tyler F. Creelan
College of Engineering
Oregon State University
503-640-3101
-----------------------------


On Wed, 25 Sep 2002, Austin Schutz wrote:

> > For example:
> >
> > Q: What's the fastest, most space-efficient way to sort all
> > of Portland's 503 phone numbers?
> >
>
> for(@numbers) {
>   s/[^\d]//g;
> }
>
> @sorted = sort { $a <=> $b }(@numbers);
>
> 	Oh, you meant fast and small for the computer. Does it matter?
> 	Most of the time optimal solutions are ones that are easy to
> write, unless you have a specific need for performance.
>
> 	Of course, there's still some point where you have to wonder.
> A modern pentium 4 processes what, 2.5 billion instructions per second?
> If it takes Windows two minutes to boot, how many of the billions of
> instructions were pure bloat?
>
> 	Austin
>
> _______________________________________________
> Pdx-pm-list mailing list
> Pdx-pm-list at mail.pm.org
> http://mail.pm.org/mailman/listinfo/pdx-pm-list
>







More information about the Pdx-pm-list mailing list