----- Original Message ----- > On 07/19/2011 10:23 AM, Przemek Klosowski wrote: > > On 07/19/2011 11:07 AM, Eric Sandeen wrote: > >> On 07/19/2011 09:59 AM, Jaroslav Skarvada wrote: > > > >>> Sad that the daemon gone. It was able to dynamically switch speed > >>> (and save power) on systems that have CPUs with high transition > >>> latency (e.g. old P4, some Atoms, etc.). On such systems the > >> > >> Actually, no... > >> > >> http://codemonkey.org.uk/2009/01/18/forthcoming-p4clockmod/ > >> > >>> So the 1.00GHz ‘frequency’ is actually “run at 2GHz, but only do > >>> work 50% of the time”. > >>> > >>> On the surface, this sounds like a good idea. The other 50%, the > >>> CPU is idle, so you’re saving power, right? > >>> Not so much. In fact, you could be burning more power. The reason > >>> for this is that when the processor is sitting there doing > >>> nothing, it isn’t lower frequency, and more importantly, it very > >>> likely isn’t entering C states. So you’re burning the same amount > >>> of power, but now you’re only doing work for 50% of the time. As a > >>> result of this, your workload takes twice as long to complete. > >> > >> I've measured it, and Dave is right. You might get something saying > >> "1.0Ghz" but you're not saving anything at all. > > > > There are second-order effects---the processor probably doesn't use > > significantly less power but the graphic card and chipset do, for > > some > > overall system effects---people quoted numbers like 20% battery > > savings > > for 50% slowdown (if p4_clockmod really stopped the CPU 50% of the > > time, > > it'd double the battery life, so this really is a very inefficient > > and > > crude method). > > I would suggest getting a wattmeter and measuring it... probably the > simplest way to know for sure. > > I'm pretty sure I measured it directly with a kill-a-watt meter, but I > no longer have a P4, so can't retest. > > -Eric > -- Well, I can do it (just for curiosity). But this was not only about P4. Anything with transition latency over 10ms is rejected by ondemand governor (and it should be so) Jaroslav -- devel mailing list devel@xxxxxxxxxxxxxxxxxxxxxxx https://admin.fedoraproject.org/mailman/listinfo/devel
