Re: [vox-tech] [OT] Windows Question for Relative

[Jeff Newmiller <jdnewmil@dcn.davMAPSis.ca.us>]

On Sun, 8 May 2005, Peter Jay Salzman wrote:

> On Sun 08 May 05,  8:05 PM, Bill Kendrick <nbs@sonic.net> said:
> > On Sun, May 08, 2005 at 05:47:14PM -0700, Karsten M. Self wrote:
> > > on Sun, May 08, 2005 at 12:58:34PM -0700, tech_dev(Alex Mandel)
> > > (tech_dev@wildintellect.com) wrote:
> > <snip>
> > > 
> > > > The other thing to think about although this is nit picky is to look
> > > > at the case cooling situation. Dell is very bad about this. You can
> > > > download a free program(for home use) called Everest to check the ram
> > > > performance and CPU temp.
> > > 
> > > Cooling won't affect performance.  It will affect component longevity.
> > 
> > I think the idea here may be that some CPUs can slow themselves down if
> > they get too hot, or something.  Maybe I'm thinking of laptops?
> > Maybe I should stop pretending to know anything about modern PC
> > capabilities. ;^)
>  
> It's called CPU throttling.  There are a few different approaches, but many
> of them work by essentially letting the CPU run at full speed while
> inserting NOPs to reduce heat generation.   Apparently, NOPs generate less
> heat.  :)

I don't see the humor here, Pete.  They do generate less heat than FP
instructions... but AFAIK no cpu throttling approaches use NOPs... rather,
the clock multiplier is usually tweaked down for fewer high-low or
low-high voltage transitions per second.

Letting the kernel stop the cpu (awaiting a timer or I/O interrupt to
reawaken) is pretty common for saving power, too.

---------------------------------------------------------------------------
Jeff Newmiller                        The     .....       .....  Go Live...
DCN:<jdnewmil@dcn.davis.ca.us>        Basics: ##.#.       ##.#.  Live Go...
                                      Live:   OO#.. Dead: OO#..  Playing
Research Engineer (Solar/Batteries            O.O#.       #.O#.  with
/Software/Embedded Controllers)               .OO#.       .OO#.  rocks...1k
---------------------------------------------------------------------------

_______________________________________________
vox-tech mailing list
vox-tech@lists.lugod.org
http://lists.lugod.org/mailman/listinfo/vox-tech