On Wed, Jun 6, 2012 at 3:12 PM, Jan Hubicka <hubi...@ucw.cz> wrote: >> The patch looks like 4_7 based which can be different from 4_6 based >> one. Do you have the latter one for review? > > Hi, > the code determining where program spend most of time in more sane way than > current hot/cold decision logic is long time on my TODO. My plan was to count > number of gimple statments executed at link time in ipa-profile pass and then > scale the logic based on it. But siply sorting gcov counts ignoring the fact > htat we have no clue how many instruction given counter controls don't seem > like bad logic. (especially for non-LTO compilations). > Any plans fo rmainline version?
Yes, I would like to put this in mainline as well. I'll send that patch out after I have added this to the google branches. Thanks, Teresa > Honza -- Teresa Johnson | Software Engineer | tejohn...@google.com | 408-460-2413
