https://gcc.gnu.org/bugzilla/show_bug.cgi?id=79506
--- Comment #3 from Jan Hubicka ---
Actually it is visible from the dump
Scaling time by probability:0.00
means that we expect quite few values to be "almost invariant". It may come
from busted BB profile of course.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=79506
--- Comment #2 from Jan Hubicka ---
badness being a very small negative number is actually normal for large
functions like this one (perhaps I should print it better though). I can check
from where the estimated speedup comes - perhaps we work
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=79506
--- Comment #1 from Richard Biener ---
Ok, so it's IVOPTs becoming relatively slower fastest (with --param
inline-min-speedup=8 it surpasses IRA as the most expensive pass). 34% of IVO
is
spent in simple_iv (and siblings), the other 34% are