[Bug fortran/100235] New: 10.3.0 Performance regressions for compile-time math intrinsics computation on arrays

[molah at ucar dot edu]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100235

            Bug ID: 100235
           Summary: 10.3.0 Performance regressions for compile-time math
                    intrinsics computation on arrays
           Product: gcc
           Version: 10.3.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
          Assignee: unassigned at gcc dot gnu.org
          Reporter: molah at ucar dot edu
  Target Milestone: ---

I am observing a significant compile time increase for gfortran-10.3.0, when
computing math intrinsic operations over arrays constructed at compile-time.

======= Test case =========

#define N 8000
PROGRAM TEST

REAL :: V(0:N)
V  = ACOS( (/ (I, I=0,N) /)/100000.0 )

END PROGRAM

==========================

Performance:
gfortran-9.3.0: 0.05s
gfortran-10.2.0: 0.05s
gfortran-10.3.0: 27.0s !!

The runtime is increasing exponentially, so that N=1600 takes a few minutes for
10.3.0.

The optimization flags appear to have no effect on the runtime.