Re: cpufreq* help.

Path	csiph.com!x330-a1.tempe.blueboxinc.net!usenet.pasdenom.info!news.albasani.net!news.swapon.de!fu-berlin.de!uni-berlin.de!individual.net!not-for-mail
From	Robert Riches <spamtrap42@jacob21819.net>
Newsgroups	comp.os.linux.misc
Subject	Re: cpufreq* help.
Date	23 Apr 2011 04:47:09 GMT
Organization	none-at-all
Lines	34
Message-ID	<slrnir4med.i7f.spamtrap42@one.localnet> (permalink)
References	<iossat$mt6$1@speranza.aioe.org> <iot1uh$srf$1@dont-email.me> <slrnir4khi.i7f.spamtrap42@one.localnet> <iotl46$n3a$1@dont-email.me>
Reply-To	spamtrap42@verizon.net
Mime-Version	1.0
Content-Type	text/plain; charset=us-ascii
Content-Transfer-Encoding	7bit
X-Trace	individual.net V7Lr9EVNpn91JZQ46wVnFgV5kfJNFqPQL4JQh7kcKnDn8grZO2
Cancel-Lock	sha1:x846DO/NbfzrSNQ6fQcsJ+O6RYk=
User-Agent	slrn/0.9.9p1 (Linux)
Xref	x330-a1.tempe.blueboxinc.net comp.os.linux.misc:887

Show key headers only | View raw

On 2011-04-23, Todd <Todd@invalid.com> wrote:
> On 04/22/2011 09:14 PM, Robert Riches wrote:
>> The relationship you describe holds true for CMOS circuits.
>> While nearly all current digital ICs are CMOS, there are other
>> logic families (NMOS, PMOS, TTL, ECL, etc.) that do not
>> necessarily run cooler with slower clock frequencies.
>
> Hi Robert,
>
> Definitely CMOS.  The current drain on others can be
> asymmetrical, but the transition is still the soar
> point for current consumption.  The faster you
> run them, the more work (work done or energy
> converted per unit of time) they perform,
> the hotter they get.  The big push on smaller traces
> is because the traces act as capacitors that have
> to be discharged/recharged every time a circuit
> switches state.  The faster you can charge/discharge
> the trace capacitance, the faster you can go and the
> less power you have to use.  And the less power, the
> less heat.

Look for the word "constant" in the "Characteristics" section of
this page:

        http://en.wikipedia.org/wiki/Emitter-coupled_logic

NMOS and PMOS are also pretty close to constant in current drain,
largely independent of clock frequency.

-- 
Robert Riches
spamtrap42@jacob21819.net
(Yes, that is one of my email addresses.)

Back to comp.os.linux.misc | Previous | Next — Previous in thread | Next in thread | Find similar

Thread

cpufreq* help. "s. keeling" <keeling@nucleus.com> - 2011-04-22 21:35 +0000
  Re: cpufreq* help. Todd <Todd@invalid.com> - 2011-04-22 16:11 -0700
    Re: cpufreq* help. Robert Riches <spamtrap42@jacob21819.net> - 2011-04-23 04:14 +0000
      Re: cpufreq* help. Todd <Todd@invalid.com> - 2011-04-22 21:38 -0700
        Re: cpufreq* help. Robert Riches <spamtrap42@jacob21819.net> - 2011-04-23 04:47 +0000
          Re: cpufreq* help. Todd <Todd@invalid.com> - 2011-04-22 22:01 -0700
            Re: cpufreq* help. Todd <Todd@invalid.com> - 2011-04-22 22:04 -0700
              Re: cpufreq* help. Michael Black <et472@ncf.ca> - 2011-04-23 11:37 -0400
                Re: cpufreq* help. Bill Marcum <bill@lat.localnet> - 2011-04-23 13:22 -0400
  Re: cpufreq* help. Baron <baron@linuxmaniac.net> - 2011-04-23 20:10 +0100

csiph-web