Archived from groups: comp.sys.ibm.pc.hardware.chips (
More info?)
George Macdonald wrote:
> On Thu, 21 Apr 2005 14:56:48 GMT, Rob Stow <rob.stow.nospam@shaw.ca> wrote:
>
>
>>George Macdonald wrote:
>>
>>>On Thu, 21 Apr 2005 05:44:50 GMT, Rob Stow <rob.stow.nospam@shaw.ca> wrote:
>>>
>>>
>>>
>>>>Felger Carbon wrote:
>>>
>>>
>>>>>This message is not tongue in cheek: I looked in Pricewatch for the
>>>>>Pent M, and finally found it as the last item in a list of P4s. 2GHz,
>>>>>2M L2, $449!
>>>>>
>>>>>Which raises this question: in that A64-PentM comparison, which
>>>>>provided the better performance/price ratio? Like the man said, not
>>>>>tongue in cheek.
>>>>>
>>>>>
>>>>
>>>>For performance/(purchase price), the Athlon 64 wins by a mile.
>>>>
>>>>However, a 2 GHz P-M uses substantially less power (ie., has a
>>>>lower operating cost) and with a decent fan/sink the fan never
>>>>needs to spin fast enough to be audible.
>>>>
>>>>I only know one person using a P-M in a desktop (and it is the 2
>>>>GHz one), but I am quite impressed by it. I matches or beats a
>>>>2.2 GHz Opteron 248 in just about everything, while using about
>>>>30 W less power and running more quietly.
>>>
>>>
>>>How do you mean 30W less power?... at full tilt, idling or average, i.e.
>>>SpeedStep is just that much better than Cool 'n' Quiet?
>>>
>>
>>At full tilt.
>
>
> By full tilt I do not mean just working... rather clock speeds pegged at
> the max beacuse of intensive calcs and memory demands. This is actually
> quite rare unless you're running something really heavy duty, including FP.
>
>
>>I also need to make a correction: the P-M system I checked out
>>was indeed a 2 GHz as I originally stated - but overclocked to
>>2.27 GHz.
>
>
> So they are close to on par with the P-M overclocked significantly... and
> the A64 *not* - interesting. One wonders what the hell Intel is thinking
> about... or do they know something they'd rather we don't?
>
>
>>There are also some power numbers here:
>>http://techreport.com/reviews/2005q1/dfi-855gme-mgf/index.x?pg=17
>>that put the difference at about 40W between a 2 GHz Athlon64 and
>>a 2 GHz P-M.
>>
>>I know it also uses less at idling - I just never bothered to
>>personally measure it: idle for me means its time to turn it off
>>and go to bed.
>
>
> Idling != idle. With both SpeedStep and Cool 'n' Quiet, your system is
> mostly idling for just general computing tasks. Unless I run some of my
> math intensive stuff, my Athlon 64 3500+ sits at 1000Mhz and 1.0V the
> majority of the time - you really have to pound on it to get it to ramp up.
>
For me - and for most review sites - idle is when the computer is
powered on and the OS has been loaded but it is not really doing
anything. It is just sitting there waiting for someone to put it
to work.
What you are describing is a "normal" workload - and some of the
review sites provide separate power consumption numbers for that.
>
>>Again, there are numbers for you in the
>>TechReport article.
>
>
> Note the caveat of the lack of "clock throttling" - not how most people are
> going to run either system, if they're "informed". All those tables of
> results with such huge "caveat(S)" on them and all because that's the mbrd
> they "had on hand"?... seems like such a err, waste?
>
I got the impression that throttling *does* work the motherboards
the TechReport used - when single core chips are used. There are
issues in getting the throttling to work with the dual-core chips
- which hopefully will be fixed soon.