GNU Fortran 95 (GCC 4.1.0 20050715 (experimental)), when run with the options
-Wall -O3 -march=pentium4 -S, optimises one of the two calls to the sin
intrinsic away in the following function:
function optsin (x)
implicit none
double precision :: optsin
double precision, intent(in) :: x
optsin = sin(x) / (1 + sin(x))
end function optsin
In the "equivalent" function
function optmatmul (a, b, ni, nj)
implicit none
integer, intent(in) :: ni, nj
double precision :: optmatmul(ni, nj)
double precision, intent(in) :: a(ni, nj), b(ni, nj)
optmatmul = matmul(a, b) / (1 + matmul(a, b))
end function optmatmul
the two calls to matmul both remain present in the assembler output, although
they are known to return the same result.
--
Summary: Double occurrence of matmul intrinsic not optimised
Product: gcc
Version: 4.1.0
Status: UNCONFIRMED
Severity: normal
Priority: P2
Component: fortran
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: schnetter at aei dot mpg dot de
CC: gcc-bugs at gcc dot gnu dot org
GCC build triplet: i686-pc-linux-gnu
GCC host triplet: i686-pc-linux-gnu
GCC target triplet: i686-pc-linux-gnu
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=22572
