Processor Scaling

Ray Parrish crp at
Wed Jan 7 03:04:55 UTC 2009

Derek Broughton wrote:
> Ray Parrish wrote:
>> Derek Broughton wrote:
>>> Ray Parrish wrote:
>>>> Well, as I said above I'm not worried about the extra power the "cpu"
>>>> not the whole desktop takes... The cpu is just a small chip inside the
>>>> box, and runs on 5 volts. Reading comprehension my friend...
>>> Electricity comprehension my friend.  5V or 500V is (largely) irrelevant
>>> - it's the power, not the voltage.
>> Ok, I did some research, and this supports my view [From Wikipedia]
>>> Dynamic frequency scaling reduces the number of instructions a
>>> processor can issue in a given amount of time, thus reducing
>>> performance. Hence, it is generally used when the workload is not
>>> CPU-bound.
>>> Dynamic frequency scaling by itself is rarely worthwhile as a way to
>>> conserve switching power. Saving the most power requires dynamic
>>> voltage scaling too, because of the V^2 component and the fact that
>>> modern CPUs are strongly optimized for low power idle states. In most
>>> constant-voltage cases it is more efficient to run briefly at peak
>>> speed and stay in a deep idle state for longer (called "race to
>>> idle"), than it is to run at a reduced clock rate for a long time and
>>> only stay briefly in a light idle state. However, reducing voltage
>>> along with clock rate can change those tradeoffs.
> So what you've just said is that you didn't understand a word of what was 
> written...
> Did you have a clue what it meant by V^2?
> It also points out "modern CPUs are strongly optimized for low power idle 
> states" - and you've yet to provide any evidence that yours _is_ one of 
> those (it's not really even true - the _cutting edge_ cpus, are, but there 
> are still huge numbers of commodity cpus that don't even do frequency 
> scaling).
> Absolutely, it's right that voltage scaling is the best way to reduce power 
> consumption - but that doesn't bear any relationship to what either you or I 
> said above.
Well, I'm just throwing a guess out, but V^2 seems to be normal notation 
for voltage squared.

This is the closest information I could find on "low power idle states" 
and comes from a pdf I found on AMD's site. They aren't very forthcoming 
with the specifications for my AMD Sempron 3400+, and only list the 
1,8ghz AM2 socket version now, while mine is the 2ghz Socket 754 version.

> Power Management
> – Multiple low-power states
> – System Management Mode (SMM)
> – ACPI-compliant, including support for processor
> performance states
Here is a quote from Wikipedia about the AMD cpus.

> *Cool'n'Quiet* is a CPU 
> <> speed 
> throttling <> and power 
> saving technology introduced by AMD <> 
> with their Athlon 64 processor line. It works by reducing the 
> processor's clock rate <> and 
> voltage <> when the processor is 
> idle. The aim of this technology is to reduce overall power 
> consumption <> and lower 
> heat generation, allowing for slower (thus quieter 
> <>) cooling fan 
> <> operation. The objectives 
> of cooler and quieter result in the name Cool'n'Quiet. The technology 
> is similar to Intel's <> SpeedStep 
> <> and AMD's own PowerNow! 
> <>, which were developed with 
> the aim of increasing laptop battery life by reducing power consumption.
It says Athlon 64, but on another page it lists the Sempron 3400+ as 
having the 64 bit instruction set, and Cool'n'Quiet technology.

Anyway, as I've said in previous posts on this topic, I save power in 
quite a number of ways, but when it comes to my computer, I want it 
running at full speed at all times. The power I save by leaving the heat 
off far into the cold season, and wearing a coat indoors, eating 80% of 
my food cold, washing dishes in cold water, and turning the warmer off 
under the coffee pot as soon as it finishes brewing, as I don't mind 
cold coffee either, should outweigh the amount of power my 64 watt cpu 

Later, Ray Parrish

-- Web index of human reviewed links.
Trouble shooting and Fixing Windows My poetry in web pages

More information about the ubuntu-users mailing list