http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59058
--- Comment #10 from Richard Biener <rguenth at gcc dot gnu.org> --- Author: rguenth Date: Thu Nov 21 14:09:15 2013 New Revision: 205217 URL: http://gcc.gnu.org/viewcvs?rev=205217&root=gcc&view=rev Log: 2013-11-21 Richard Biener <rguent...@suse.de> PR tree-optimization/59058 * tree-scalar-evolution.h (number_of_exit_cond_executions): Remove. * tree-scalar-evolution.c (number_of_exit_cond_executions): Likewise. * tree-vectorizer.h (LOOP_PEELING_FOR_ALIGNMENT): Rename to ... (LOOP_VINFO_PEELING_FOR_ALIGNMENT): ... this. (NITERS_KNOWN_P): Fold into ... (LOOP_VINFO_NITERS_KNOWN_P): ... this. (LOOP_VINFO_PEELING_FOR_NITER): Add. * tree-vect-loop-manip.c (vect_gen_niters_for_prolog_loop): Use LOOP_VINFO_PEELING_FOR_ALIGNMENT. (vect_do_peeling_for_alignment): Re-use precomputed niter instead of re-emitting it. * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Use LOOP_VINFO_PEELING_FOR_ALIGNMENT. * tree-vect-loop.c (vect_get_loop_niters): Use number_of_latch_executions. (new_loop_vec_info): Initialize LOOP_VINFO_PEELING_FOR_NITER. (vect_analyze_loop_form): Simplify. (vect_analyze_loop_operations): Move epilogue peeling code ... (vect_analyze_loop_2): ... here and adjust it to compute LOOP_VINFO_PEELING_FOR_NITER. (vect_estimate_min_profitable_iters): Use LOOP_VINFO_PEELING_FOR_ALIGNMENT. (vect_build_loop_niters): Emit on the preheader. (vect_generate_tmps_on_preheader): Likewise. (vect_transform_loop): Use LOOP_VINFO_PEELING_FOR_NITER instead of recomputing it. Adjust. Modified: trunk/gcc/ChangeLog trunk/gcc/tree-scalar-evolution.c trunk/gcc/tree-scalar-evolution.h trunk/gcc/tree-vect-data-refs.c trunk/gcc/tree-vect-loop-manip.c trunk/gcc/tree-vect-loop.c trunk/gcc/tree-vectorizer.h
