On Feb 9, 2012, at 10:43 AM, Jaromir Capik wrote: I very sorry Jim. > > After some tests here I believe that we have no chance > to implement such feature. > I built my request on a wrong presumption, that it's > the total sum, that changes with the dynamic frequency scaling. > But that's not true. In fact it's the user counter that changes, > because the particular values represent time units depleted by > the processes and in such case we have absolutely no chance > to make it frequency independent. > Even if the issue is present, we can't solve it. > That's frustrating. > > Sorry for stealing Your time ... Cheer up, Jaromir! You still gave us the Nehalem approach. That's a worthwhile improvement, isn't it?. Regards, Jim