Hi, This patch fixes the vectorizer testsuite failures at -m64 on darwin, reported in: http://gcc.gnu.org/ml/gcc/2007-08/msg00052.html The failures are because of this patch: http://gcc.gnu.org/ml/gcc-cvs/2007-07/msg00447.html http://gcc.gnu.org/ml/gcc-patches/2007-07/msg00082.html ...that, for 64bit darwin, tells the vectorizer that data cannot be assumed to be naturally aligned, which in turn means that we cannot use loop peeling to force data alignment on 64bit darwin. So testcases that expect peeling to happen need to be updated accordingly. Thanks to Jack Howarth for testing the patch on powerpc-apple-darwin9 (both -m32 and -m64). I tested it on powerp64-linux (both -m32 and -m64) and i386-linux (only -m32). Thanks to Ira for testing on x86_64. I will commit this patch to mainline later today dorit * gcc.dg/vect/vect-28.c: Fix test to not expect peeling on !vector_alignment_reachable targets. * gcc.dg/vect/vect-33.c: Likewise. * gcc.dg/vect/vect-42.c: Likewise. * gcc.dg/vect/vect-44.c: Likewise. * gcc.dg/vect/vect-50.c: Likewise. * gcc.dg/vect/vect-70.c: Likewise. * gcc.dg/vect/vect-71.c: Likewise. * gcc.dg/vect/vect-87.c: Likewise. * gcc.dg/vect/vect-88.c: Likewise. * gcc.dg/vect/vect-91.c: Likewise. * gcc.dg/vect/vect-93.c: Likewise. * gcc.dg/vect/vect-96.c: Likewise. * gcc.dg/vect/no-section-anchors-69.c: Likewise. * gcc.dg/vect/section-anchors-69.c: Likewise. * gcc.dg/vect/pr25413.c: Likewise. * gcc.dg/vect/costmodel/ppc/costmodel-33.c: Likewise. * gcc.dg/vect/costmodel/ppc/costmodel-76b.c: Likewise. * gfortran.dg/vect-2.f90: Likewise. * gfortran.dg/vect-3.f90: Likewise. * gfortran.dg/vect-4.f90: Likewise. * gfortran.dg/vect-5.f90: Likewise. * lib/target-supports.exp (check_effective_target_natural_alignment): Return false for darwin. (check_effective_target_vector_alignment_reachable_for_double): New. (See attached file: reachable_alignment_vecttestfixes.txt)