Re: Revamp loop profile scaling to profile_probability

2017-07-03 Thread Jan Hubicka
> On Sat, Jul 1, 2017 at 7:14 PM, Jan Hubicka wrote: > > Hi, > > this patch makes loop profile scaling to use profile_probability. This > > is mostly trivial change except for vect_do_peeling which seems to scale > > profile down and then back up. This is a bad idea, because things may > > simp

Re: Revamp loop profile scaling to profile_probability

2017-07-03 Thread Richard Biener
On Sat, Jul 1, 2017 at 7:14 PM, Jan Hubicka wrote: > Hi, > this patch makes loop profile scaling to use profile_probability. This > is mostly trivial change except for vect_do_peeling which seems to scale > profile down and then back up. This is a bad idea, because things may simply > drop to 0.

Revamp loop profile scaling to profile_probability

2017-07-01 Thread Jan Hubicka
Hi, this patch makes loop profile scaling to use profile_probability. This is mostly trivial change except for vect_do_peeling which seems to scale profile down and then back up. This is a bad idea, because things may simply drop to 0. So I kept that one to use integer scaling (because probabili