Many factors affect how much increase in efficiency you can realize by
multiple core or multiple processors.  In a design with good
infrastructure (such as a good server), you may get up to .85 times
faster processing with two processors versus one, especially in a 64-bit
design with plenty of memory and processor cache.  This depends on the
tasks being performed and how well the system and the OS can distribute
the tasks over multiple processors, bus speeds and characteristics,
memory (capacity and transfer speed) and even disk characteristics. 

So, in an efficient computer with two processors and an OS and
application(s) capable of taking advantage of multiple processors, two
CPUs could get up to 1.7 (1.85, tops on a very good server) faster
performance.  But Tom's guess of 1.5 times is probably a more typical
limit for two processors in a personal computer.  

In some cases, though, you may not see any increased performance.  If
you drive a powerful car through a parking lot with speed bumps, you'll
probably go at about the same speed as any other car, for example.
Also, in stop and go rush hour, that fast car won't help you go any
faster.  An OS or an application that can't take advantage of multiple
processors (or processor cores) will run on one processor, so the second
processor (or core) will not add anything to the performance of that
application.

Thank you,
 
Mark Snyder
-----Original Message-----
Typically when you are increasing processor speed a number of things are
happening which increase your speed more than just a mere multiple of
the processor clock speed.

when you go from say a 1.5 to a 3.0 you are not merely just doubling
your speed, but you will notice far more bang.  Why is that?  I am not
an engineer, but I know that motherboards, Front Speed Buss, Memory
efficiency and a number of other factors help the speed throughput so it
is not just a multiple but much faster.

And yes there is some down side to all of it so that it is never a
straight graph increase, but one that has all sorts of dips and valleys.

Dual Cores lose some of their efficiency to a number of engineering
things and such that I cannot possibly explain.  How much do the newer
Quad Cores lose?

Stewart

(Who will probably upgrade to a quad core when octocores come out.)



At 09:27 PM 9/19/2007, you wrote:
> >Increases of processor speed increase processing not mathematically 
> >but algebraically.  (I may have the terms wrong but it is not 2+2+2, 
> >it is 2x2x2)
>
>Who told you that? Next time they tell you anything put your fingers in

>your ears and sing la-la-la until they go away.
>
>Due to overhead of various sorts increases have diminishing returns. So

>a dual processor is only 150% as effective as a single processor of the

>same speed. Etc., etc.

Rev. Stewart A. Marshall
mailto:[EMAIL PROTECTED]
Prince of Peace
Ozark, AL  SL 82


************************************************************************
* ==> QUICK LIST-COMMAND REFERENCE - Put the following commands in  <==
* ==> the body of an email & send 'em to: [EMAIL PROTECTED] <==
* Join the list: SUBSCRIBE COMPUTERGUYS-L Your Name
* Too much mail? Try Daily Digests command: SET COMPUTERGUYS-L DIGEST
* Tired of the List? Unsubscribe command: SIGNOFF COMPUTERGUYS-L
* New address? From OLD address send: CHANGE COMPUTERGUYS-L
YourNewAddress
* Need more help? Send mail to: [EMAIL PROTECTED]
************************************************************************
* List archive at www.mail-archive.com/[email protected]/
* RSS at
www.mail-archive.com/[email protected]/maillist.xml
* Messages bearing the header "X-No-Archive: yes" will not be archived
************************************************************************


************************************************************************
* ==> QUICK LIST-COMMAND REFERENCE - Put the following commands in  <==
* ==> the body of an email & send 'em to: [EMAIL PROTECTED] <==
* Join the list: SUBSCRIBE COMPUTERGUYS-L Your Name
* Too much mail? Try Daily Digests command: SET COMPUTERGUYS-L DIGEST
* Tired of the List? Unsubscribe command: SIGNOFF COMPUTERGUYS-L
* New address? From OLD address send: CHANGE COMPUTERGUYS-L YourNewAddress
* Need more help? Send mail to: [EMAIL PROTECTED]
************************************************************************
* List archive at www.mail-archive.com/[email protected]/
* RSS at www.mail-archive.com/[email protected]/maillist.xml
* Messages bearing the header "X-No-Archive: yes" will not be archived
************************************************************************

Reply via email to