https://gcc.gnu.org/bugzilla/show_bug.cgi?id=49627
dr.robert.kosik at gmail dot com changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |dr.robert.kosik at gmail dot com --- Comment #4 from dr.robert.kosik at gmail dot com --- This Bug is still open in 2016-01-27 with gfortran 4.8.4 double precision :: X integer(16) :: norm norm = NINT(X,KIND=16) internal compiler error: in build_round_expr, at fortran/trans-intrinsic.c:394 norm = NINT(X,KIND=16) ^ Please submit a full bug report. It works fine with other types of rounding, i.e., INT(X,16), FLOOR(X,16), CEILING(X,16) compile ok.