On Mon, 3 Jan 2005, Satish Balay wrote: > On Mon, 3 Jan 2005, Satish Balay wrote: > > On Mon, 3 Jan 2005, Dave Jones wrote: > > > > > > 'cpuspeed' needs to support '--ignore-intermediate-frequencies' option > > > > to speed up the transtiton between max & min speeds. > > > > > > It can't. Some implementations of speed scaling can't handle > > > huge leaps, and need to be 'stepped'. Some of the drivers do this > > > internally anyway, so even if you removed it from cpuspeed, the > > > multiple transitions would still be occuring. > > > > I guess internal multiple transitions is fine. But looks like cpuspeed > > is not taking advantage of it. > > > > Each transition is a single step (600 -> 800 -> 1000 etc..) - and each > > step occurs only ater the requisite '-i' interaval. > > (to add some perspective) - I've used cpudyn on FC1 - and I liked > it. (However it doesn't work with fedora 2.6 kernels - hence using > cpuspeed) > > http://mnm.uib.es/gallir/cpudyn/faq.html > The cpudyn FAQ has the following text: > > 7. Whay does it have just two states: powersave and performance? Why > don't you allow to specify other frequencies. > > Basically because I dont need it. I just need these two states, go to > the maximum when it's needed it. Go to the minimum if it's not > needed. cpudynd is very good at reacting interactively to the CPU > load. Another datapoint from 'powernowd' (Trevor alluded to). From the README: --- Mode 3, LEAPS : Immediately jump to the highest frequency if usage above 80%. Immediately jump to the lowest frequency if usage below 20%. --- Satish
