http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48636
--- Comment #28 from Dominique d'Humieres <dominiq at lps dot ens.fr> 2012-10-20 11:22:16 UTC --- If I understand correctly the patch, the default value for max-inline-min-speedup is 20. This could be over-agressive: for fatigue.f90 the threshold is between 94 (fast) and 95 (slow). I see a similar threshold for test_fpu.f90 from 100 (fast) to ~97 (slow), unfortunately above 94. Also the choice of true or false for big_speedup_p is based on a test 't-t1>a*t' which is equivalent to 't1<=b*t' with b=1-a. Any reason for the choice? Last point, AFAICT the behavior of the different param tuning the inlining is often non monotonic (I am trying to investigate that in more detail).
