https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70471
Richard Biener <rguenth at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Keywords| |missed-optimization, ra
Status|UNCONFIRMED |NEW
Last reconfirmed| |2016-03-31
Ever confirmed|0 |1
--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> ---
Please provide the compiler command plus output when -v appended so we can see
what -march=native expands to for you.
As there are a load of spills in this code the extra move is likely a remanent
of a late removed spill.
It looks like you are talking about the function
_ZN11ML_BSSN_FD4L34ML_BSSN_FD4_EvolutionInterior_BodyEPK4_cGHiiPKdS4_S4_PKiS6_iPrKPd._omp_fn.0
created from ML_BSSN_FD4::ML_BSSN_FD4_EvolutionInterior_Body.
Confirmed there are a load of spills in the code, couldn't find the specific
sequence with using -march=core-avx2 (but I'm also not running on darwin
which will make assembler look differently).
Looks like GCC 6 spills a lot less here.
