aboutsummaryrefslogtreecommitdiff
path: root/gcc/tree-vectorizer.h
AgeCommit message (Collapse)AuthorFilesLines
2013-05-16tree-vectorizer.h (NUM_PATTERNS): Increment.Jakub Jelinek1-1/+1
* tree-vectorizer.h (NUM_PATTERNS): Increment. * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add vect_recog_rotate_pattern. (vect_recog_rotate_pattern): New function. * gcc.target/i386/rotate-3.c: New test. * gcc.target/i386/rotate-3a.c: New test. * gcc.target/i386/rotate-4.c: New test. * gcc.target/i386/rotate-4a.c: New test. * gcc.target/i386/rotate-5.c: New test. * gcc.target/i386/rotate-5a.c: New test. From-SVN: r198978
2013-04-23This patch extracts approved portions of the hash_table patches to the...Lawrence Crowl1-1/+26
This patch extracts approved portions of the hash_table patches to the cxx-conversion branch for files not under gcc/config. Update various hash tables from htab_t to hash_table. Modify types and calls to match. * tree-ssa-coalesce.c'coalesce_list_d.list from htab_t to hash_table. Fold coalesce_pair_map_hash and coalesce_pair_map_eq into new struct coalesce_pair_hasher. Removed struct coalesce_pair_iterator, as did not meet the hash_table iterator interface and it provided no significant code reduction. This leads to a change in the implementation of FOR_EACH_PARTITION_PAIR. * statistics.c'statistics_hashes Fold hash_statistics_eq into new struct stats_counter_hasher. * hash-table.h'hash_table Add documentation. Add nested class iterator and methods to hash_table. Add FOR_EACH_HASH_TABLE_ELEMENT implemented with those iterators. Change uses of FOR_EACH_HTAB_ELEMENT to FOR_EACH_HASH_TABLE_ELEMENT. * tree-ssa-sccvn.c'vn_tables_s.nary Fold vn_nary_op_hash, vn_nary_op_eq into new struct vn_nary_op_hasher. Add typedef vn_nary_op_table_type. Add typedef vn_nary_op_iterator_type. * tree-ssa-sccvn.c'vn_tables_s.phis Fold vn_phi_hash, free_phi into new struct vn_phi_hasher. Add typedef vn_phi_table_type. Add typedef vn_phi_iterator_type. * tree-ssa-sccvn.c'vn_tables_s.references Fold vn_reference_hash, vn_reference_op_eq, free_reference into new struct vn_reference_hasher. Add typedef vn_reference_table_type. Add typedef vn_reference_iterator_type. * tree-ssa-sccvn.c'constant_value_ids Fold vn_constant_hash, vn_constant_eq into new struct vn_constant_hasher. * tree-into-ssa.c'var_infos Fold var_info_hash, var_info_eq into new struct var_info_hasher. * tree-vectorizer.h'_loop_vec_info::peeling_htab * tree-vectorizer.h New struct peel_info_hasher. * tree-vect-loop.c Update dependent calls and types to match. * tree-vect-data-refs.c Fold vect_peeling_hash and vect_peeling_hash_eq into struct peel_info_hasher. * tree-ssa-reassoc.c'undistribute_ops_list::ctable Fold oecount_hash and oecount_eq into new struct oecount_hasher. * tree-ssa-loop-im.c'memory_accesses.refs Fold memref_hash and memref_eq into new struct mem_ref_hasher. Tested on x86_64. Index: gcc/ChangeLog 2013-04-23 Lawrence Crowl <crowl@google.com> * Makefile.in: Update as needed below. * hash-table.h (class hash_table): Correct many methods with parameter types compare_type to the correct value_type. (Correct code was unlikely to notice the change.) (hash_table::elements_with_deleted) New. (class hashtable::iterator): New. (hashtable::begin()): New. (hashtable::end()): New. (FOR_EACH_HASH_TABLE_ELEMENT): New. * statistics.c (statistics_hashes): Change type to hash_table. Update dependent calls and types. * tree-into-ssa.c (var_infos): Change type to hash_table. Update dependent calls and types. * tree-ssa-coalesce.c (struct coalesce_list_d.list): Change type to hash_table. Update dependent calls and types. * tree-ssa-loop-im.c (struct mem_ref.refs): Change type to hash_table. Update dependent calls and types. * tree-ssa-reassoc.c (undistribute_ops_list::ctable): Change type to hash_table. Update dependent calls and types. * tree-ssa-sccvn.c (vn_tables_s::nary): Change type to hash_table. Update dependent calls and types. (vn_tables_s::phis): Likewise. (vn_tables_s::references): Likewise. * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types. (vn_reference_eq): Update parameter and return types. * tree-ssa-structalias.c (pointer_equiv_class_table): Change type to hash_table. Update dependent calls and types. (location_equiv_class_table): Likewise. * tree-vect-data-refs.c: Consequential changes for making peeling a hash_table. * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update. (destroy_loop_vec_info): Dependent hash_table update. * tree-vectorizer.h (peeling_htab): Change type to hash_table. Update dependent calls and types. From-SVN: r198213
2013-04-19re PR tree-optimization/56270 (loop over array of struct float causes ↵Richard Biener1-6/+5
compiler error: segmentation fault) 2013-04-19 Richard Biener <rguenther@suse.de> * tree-vectorizer.h (struct _slp_instance): Move load_permutation member ... (struct _slp_tree): ... here. Make it a vector of unsigned ints. (SLP_INSTANCE_LOAD_PERMUTATION): Remove. (SLP_TREE_LOAD_PERMUTATION): Add. (vect_transform_slp_perm_load): Adjust prototype. * tree-vect-slp.c (vect_free_slp_tree): Adjust. (vect_free_slp_instance): Likewise. (vect_create_new_slp_node): Likewise. (vect_supported_slp_permutation_p): Remove. (vect_slp_rearrange_stmts): Adjust. (vect_supported_load_permutation_p): Likewise. Inline vect_supported_slp_permutation_p here. (vect_analyze_slp_instance): Compute load permutations per slp node instead of per instance. (vect_get_slp_defs): Adjust. (vect_transform_slp_perm_load): Likewise. (vect_schedule_slp_instance): Remove redundant code. (vect_schedule_slp): Remove hack for PR56270, add it ... * tree-vect-stmts.c (vectorizable_load): ... here, do not CSE loads for SLP. Adjust. From-SVN: r198095
2013-04-15re PR tree-optimization/56933 (Vectorizer missing read-write dependency for ↵Richard Biener1-6/+0
interleaved accesses) 2013-04-15 Richard Biener <rguenther@suse.de> PR tree-optimization/56933 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep member. (GROUP_READ_WRITE_DEPENDENCE): Remove. (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise. * tree-vect-data-refs.c (vect_analyze_group_access): Move dependence check ... vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence): ... here. * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize GROUP_READ_WRITE_DEPENDENCE. * gcc.dg/vect/pr56933.c: New testcase. From-SVN: r197972
2013-04-10tree-vectorizer.h (struct _slp_oprnd_info): Remove first_const_oprnd field, ↵Richard Biener1-2/+1
rename first_def_type to first_op_type. 2013-04-10 Richard Biener <rguenther@suse.de> * tree-vectorizer.h (struct _slp_oprnd_info): Remove first_const_oprnd field, rename first_def_type to first_op_type. * tree-vect-slp.c (vect_create_oprnd_info): Adjust. (vect_get_and_check_slp_defs): Always use the type of the operand. Allow mixed vect_external_def, vect_constant_def types. (vect_get_constant_vectors): Handle mixed vect_external_def, vect_constant_def types. * gcc.dg/vect/slp-39.c: New testcase. From-SVN: r197669
2013-04-09tree-vectorizer.h (slp_void_p): Remove.Richard Biener1-5/+4
2013-04-09 Richard Biener <rguenther@suse.de> * tree-vectorizer.h (slp_void_p): Remove. (slp_tree): Typedef before _slp_tree declaration. (struct _slp_tree): Use a vector of slp_tree as children. (vect_get_place_in_interleaving_chain): Remove. * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain): Move ... * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here and make static. (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts, vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts, vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations, vect_schedule_slp_instance, vect_remove_slp_scalar_calls): Use slp_node instead of slp_void_p and adjust. From-SVN: r197629
2013-03-21tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.Richard Biener1-2/+2
2013-03-21 Richard Biener <rguenther@suse.de> * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove. (vect_insert_into_interleaving_chain): Likewise. (vect_drs_dependent_in_basic_block): Inline ... (vect_slp_analyze_data_ref_dependence): ... here. New function, split out from ... (vect_analyze_data_ref_dependence): ... here. Simplify. (vect_check_interleaving): Simplify. (vect_analyze_data_ref_dependences): Likewise. Split out ... (vect_slp_analyze_data_ref_dependences): ... this new function. (dr_group_sort_cmp): New function. (vect_analyze_data_ref_accesses): Compute data-reference groups here instead of in vect_analyze_data_ref_dependence. Use a more efficient algorithm. * tree-vect-slp.c (vect_slp_analyze_bb_1): Use vect_slp_analyze_data_ref_dependences. Call vect_analyze_data_ref_accesses earlier. * tree-vect-loop.c (vect_analyze_loop_2): Likewise. * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust. (vect_slp_analyze_data_ref_dependences): New prototype. * gcc.dg/vect/vect-outer-3a-big-array.c: Adjust. * gcc.dg/vect/vect-outer-3a.c: Likewise. From-SVN: r196872
2013-02-28re PR middle-end/56461 (GCC is leaking lots of memory)Jakub Jelinek1-1/+1
PR middle-end/56461 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument type to vec<vec<tree> > *. * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs to be vec<tree> instead of vec<tree> *, set vec_defs to vNULL and call vec_defs.create (number_of_vects), adjust other uses of vec_defs. * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call, vectorizable_condition): Adjust vect_get_slp_defs callers. From-SVN: r196360
2013-02-05re PR tree-optimization/53342 (rnflow.f90 is ~5% slower after revision 187340)Richard Biener1-1/+0
2013-02-05 Richard Biener <rguenther@suse.de> PR tree-optimization/53342 PR tree-optimization/53185 * tree-vectorizer.h (vect_check_strided_load): Remove. * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not disallow peeling for vectorized strided loads. (vect_check_strided_load): Make static and simplify. (vect_analyze_data_refs): Adjust. * tree-vect-stmts.c (vectorizable_load): Handle peeled loops correctly when vectorizing strided loads. * gcc.dg/vect/pr53185-2.c: New testcase. From-SVN: r195759
2013-01-10Update copyright years in gcc/Richard Sandiford1-2/+1
From-SVN: r195098
2012-11-17This patch rewrites the old VEC macro-based interface into a new one based ↵Diego Novillo1-59/+45
on the template class 'vec'. This patch rewrites the old VEC macro-based interface into a new one based on the template class 'vec'. The user-visible changes are described in http://gcc.gnu.org/wiki/cxx-conversion/cxx-vec. I have tested the patch pretty extensively: - Regular bootstraps on x86_64, ppc, ia64, sparc and hppa. - Bootstraps with --enable-checking=release - Bootstraps with --enable-checking=gc,gcac - Basic builds on all targets (using contrib/config-list.mk). We no longer access the vectors via VEC_* macros. The pattern is "VEC_operation (T, A, V, args)" becomes "V.operation (args)". The only thing I could not do is create proper ctors and dtors for the vec class. Since these vectors are stored in unions, we have to keep them as PODs (C++03 does not allow non-PODs in unions). This means that creation and destruction must be explicit. There is a new method vec<type, allocation, layout>::create() and another vec<type, allocation, layout>::destroy() to allocate the internal vector. For vectors that must be pointers, there is a family of free functions that implement the operations that need to tolerate NULL vectors. These functions all start with the prefix 'vec_safe_'. See the wiki page for details. The gengtype change removes the special handling for VEC() that used to exist in gengtype. Additionally, it allows gengtype to recognize templates of more than one argument and introduces the concept of an undefined type (useful for template arguments that may or may not be types). When a TYPE_UNDEFINED is reached, gengtype will ignore it if it happens inside a type marked with GTY((user)). Otherwise, it will emit an error. Finally, gengtype rejects root types marked GTY((user)) that are not first class pointers. 2012-11-16 Diego Novillo <dnovillo@google.com> VEC API overhaul (http://gcc.gnu.org/wiki/cxx-conversion/cxx-vec) * vec.c (register_overhead): Convert it into member function of vec_prefix. (release_overhead): Likewise. (calculate_allocation): Likewise. (vec_heap_free): Remove. (vec_gc_o_reserve_1): Remove. (vec_heap_o_reserve_1): Remove. (vec_stack_o_reserve_1): Remove. (vec_stack_o_reserve_exact): Remove. (register_stack_vec): New. (stack_vec_register_index): New. (unregister_stack_vec): New. (vec_assert_fail): Remove. * vec.h: Conditionally include ggc.h. Document conditional hackery. Update top-level documentation. (ALONE_VEC_CHECK_INFO): Remove. (VEC_CHECK_INFO): Remove. (ALONE_VEC_CHECK_DECL): Remove. (VEC_CHECK_DECL): Remove. (ALONE_VEC_CHECK_PASS): Remove. (VEC_CHECK_PASS): Remove. (VEC_ASSERT): Remove. (vec_prefix): Add friends va_gc, va_gc_atomic, va_heap and va_stack. Mark fields alloc_ and num_ as protected. (struct vec_t): Remove. Remove all function members. (struct vl_embed): Declare. (struct vl_ptr): Declare. (free): Remove. (reserve_exact): Remove. (reserve): Remove. (safe_splice): Remove. (safe_push): Remove. (safe_grow): Remove. (safe_grow_cleared): Remove. (safe_insert): Remove. (DEF_VEC_I): Remove. (DEF_VEC_ALLOC_I): Remove. (DEF_VEC_P): Remove. (DEF_VEC_ALLOC_P): Remove. (DEF_VEC_O): Remove. (DEF_VEC_ALLOC_O): Remove. (DEF_VEC_ALLOC_P_STACK): Remove. (DEF_VEC_ALLOC_O_STACK): Remove. (DEF_VEC_ALLOC_I_STACK): Remove. (DEF_VEC_A): Remove. (DEF_VEC_ALLOC_A): Remove. (vec_stack_p_reserve_exact_1): Remove. (vec_stack_o_reserve): Remove. (vec_stack_o_reserve_exact): Remove. (VEC_length): Remove. (VEC_empty): Remove. (VEC_address): Remove. (vec_address): Remove. (VEC_last): Remove. (VEC_index): Remove. (VEC_iterate): Remove. (VEC_embedded_size): Remove. (VEC_embedded_init): Remove. (VEC_free): Remove. (VEC_copy): Remove. (VEC_space): Remove. (VEC_reserve): Remove. (VEC_reserve_exact): Remove. (VEC_splice): Remove. (VEC_safe_splice): Remove. (VEC_quick_push): Remove. (VEC_safe_push): Remove. (VEC_pop): Remove. (VEC_truncate): Remove. (VEC_safe_grow): Remove. (VEC_replace): Remove. (VEC_quick_insert): Remove. (VEC_safe_insert): Remove. (VEC_ordered_remove): Remove. (VEC_unordered_remove): Remove. (VEC_block_remove): Remove. (VEC_lower_bound): Remove. (VEC_alloc): Remove. (VEC_qsort): Remove. (va_heap): Declare. (va_heap::default_layout): New typedef to vl_ptr. (va_heap::reserve): New. (va_heap::release): New. (va_gc): Declare. (va_gc::default_layout): New typedef to vl_embed. (va_gc::reserve): New. (va_gc::release): New. (va_gc_atomic): Declare. Inherit from va_gc. (va_stack): Declare. (va_stack::default_layout): New typedef to vl_ptr. (va_stack::alloc): New. (va_stack::reserve): New. (va_stack::release): New. (register_stack_vec): Declare. (stack_vec_register_index): Declare. (unregister_stack_vec): Declare. (vec<T, A = va_heap, L = typename A::default_layout>): Declare empty vec template. (vec<T, A, vl_embed>): Partial specialization for embedded layout. (vec<T, A, vl_embed>::allocated): New. (vec<T, A, vl_embed>::length): New. (vec<T, A, vl_embed>::is_empty): New. (vec<T, A, vl_embed>::address): New. (vec<T, A, vl_embed>::operator[]): New. (vec<T, A, vl_embed>::last New. (vec<T, A, vl_embed>::space): New. (vec<T, A, vl_embed>::iterate): New. (vec<T, A, vl_embed>::iterate): New. (vec<T, A, vl_embed>::copy): New. (vec<T, A, vl_embed>::splice): New. (vec<T, A, vl_embed>::quick_push New. (vec<T, A, vl_embed>::pop New. (vec<T, A, vl_embed>::truncate): New. (vec<T, A, vl_embed>::quick_insert): New. (vec<T, A, vl_embed>::ordered_remove): New. (vec<T, A, vl_embed>::unordered_remove): New. (vec<T, A, vl_embed>::block_remove): New. (vec<T, A, vl_embed>::qsort): New. (vec<T, A, vl_embed>::lower_bound): New. (vec<T, A, vl_embed>::embedded_size): New. (vec<T, A, vl_embed>::embedded_init): New. (vec<T, A, vl_embed>::quick_grow): New. (vec<T, A, vl_embed>::quick_grow_cleared): New. (vec_safe_space): New. (vec_safe_length): New. (vec_safe_address): New. (vec_safe_is_empty): New. (vec_safe_reserve): New. (vec_safe_reserve_exact): New. (vec_alloc): New. (vec_free): New. (vec_safe_grow): New. (vec_safe_grow_cleared): New. (vec_safe_iterate): New. (vec_safe_push): New. (vec_safe_insert): New. (vec_safe_truncate): New. (vec_safe_copy): New. (vec_safe_splice): New. (vec<T, A, vl_ptr>): New partial specialization for the space efficient layout. (vec<T, A, vl_ptr>::exists): New. (vec<T, A, vl_ptr>::is_empty): New. (vec<T, A, vl_ptr>::length): New. (vec<T, A, vl_ptr>::address): New. (vec<T, A, vl_ptr>::operator[]): New. (vec<T, A, vl_ptr>::operator!=): New. (vec<T, A, vl_ptr>::operator==): New. (vec<T, A, vl_ptr>::last): New. (vec<T, A, vl_ptr>::space): New. (vec<T, A, vl_ptr>::iterate): New. (vec<T, A, vl_ptr>::copy): New. (vec<T, A, vl_ptr>::reserve): New. (vec<T, A, vl_ptr>::reserve_exact): New. (vec<T, A, vl_ptr>::splice): New. (vec<T, A, vl_ptr>::safe_splice): New. (vec<T, A, vl_ptr>::quick_push): New. (vec<T, A, vl_ptr>::safe_push): New. (vec<T, A, vl_ptr>::pop): New. (vec<T, A, vl_ptr>::truncate): New. (vec<T, A, vl_ptr>::safe_grow): New. (vec<T, A, vl_ptr>::safe_grow_cleared): New. (vec<T, A, vl_ptr>::quick_grow): New. (vec<T, A, vl_ptr>::quick_grow_cleared): New. (vec<T, A, vl_ptr>::quick_insert): New. (vec<T, A, vl_ptr>::safe_insert): New. (vec<T, A, vl_ptr>::ordered_remove): New. (vec<T, A, vl_ptr>::unordered_remove): New. (vec<T, A, vl_ptr>::block_remove): New. (vec<T, A, vl_ptr>::qsort): New. (vec<T, A, vl_ptr>::lower_bound): New. (vec_stack_alloc): Define. (FOR_EACH_VEC_SAFE_ELT): Define. * vecir.h: Remove. Update all users. * vecprim.h: Remove. Update all users. Move uchar to coretypes.h. * Makefile.in (VEC_H): Add $(GGC_H). Remove vecir.h and vecprim.h dependencies everywhere. 2012-11-16 Diego Novillo <dnovillo@google.com> * gengtype-lex.l (VEC): Remove. Add characters in the set [\!\>\.-]. * gengtype-parse.c (token_names): Remove "VEC". (require_template_declaration): Remove handling of VEC_TOKEN. (type): Likewise. Call create_user_defined_type when parsing GTY((user)). * gengtype-state.c (type_lineloc): handle TYPE_UNDEFINED. (write_state_undefined_type): New. (write_state_type): Call write_state_undefined_type for TYPE_UNDEFINED. (read_state_type): Call read_state_undefined_type for TYPE_UNDEFINED. * gengtype.c (dbgprint_count_type_at): Handle TYPE_UNDEFINED. (create_user_defined_type): Make extern. (type_for_name): Factor out of resolve_typedef. (create_undefined_type): New (resolve_typedef): Call it when we cannot find a previous typedef and the type is not a template. (find_structure): Accept TYPE_UNDEFINED. (set_gc_used_type): Add argument ALLOWED_UNDEFINED_TYPES, default to false. Emit an error for TYPE_UNDEFINED unless LEVEL is GC_UNUSED or ALLOWED_UNDEFINED_TYPES is set. Set ALLOWED_UNDEFINED_TYPES to true for TYPE_USER_STRUCT. (filter_type_name): Accept templates with more than one argument. (output_mangled_typename): Handle TYPE_UNDEFINED (walk_type): Likewise. (write_types_process_field): Likewise. (write_func_for_structure): If CHAIN_NEXT is set, ORIG_S should not be a user-defined type. (write_types_local_user_process_field): Handle TYPE_ARRAY, TYPE_NONE and TYPE_UNDEFINED. (write_types_local_process_field): Likewise. (contains_scalar_p): Return 0 for TYPE_USER_STRUCT. (write_root): Reject user-defined types that are not pointers. Handle TYPE_NONE, TYPE_UNDEFINED, TYPE_UNION, TYPE_LANG_STRUCT and TYPE_PARAM_STRUCT. (output_typename): Handle TYPE_NONE, TYPE_UNDEFINED, and TYPE_ARRAY. (dump_typekind): Handle TYPE_UNDEFINED. * gengtype.h (enum typekind): Add TYPE_UNDEFINED. (create_user_defined_type): Declare. (enum gty_token): Remove VEC_TOKEN. 2012-11-16 Diego Novillo <dnovillo@google.com> Adjust for new vec API (http://gcc.gnu.org/wiki/cxx-conversion/cxx-vec) * coretypes.h (uchar): Define. * alias.c: Use new vec API in vec.h. * asan.c: Likewise. * attribs.c: Likewise. * basic-block.h: Likewise. * bb-reorder.c: Likewise. * builtins.c: Likewise. * calls.c: Likewise. * cfg.c: Likewise. * cfganal.c: Likewise. * cfgcleanup.c: Likewise. * cfgexpand.c: Likewise. * cfghooks.c: Likewise. * cfghooks.h: Likewise. * cfgloop.c: Likewise. * cfgloop.h: Likewise. * cfgloopanal.c: Likewise. * cfgloopmanip.c: Likewise. * cfgrtl.c: Likewise. * cgraph.c: Likewise. * cgraph.h: Likewise. * cgraphclones.c: Likewise. * cgraphunit.c: Likewise. * combine.c: Likewise. * compare-elim.c: Likewise. * coverage.c: Likewise. * cprop.c: Likewise. * data-streamer.h: Likewise. * dbxout.c: Likewise. * dce.c: Likewise. * df-core.c: Likewise. * df-problems.c: Likewise. * df-scan.c: Likewise. * dominance.c: Likewise. * domwalk.c: Likewise. * domwalk.h: Likewise. * dse.c: Likewise. * dwarf2cfi.c: Likewise. * dwarf2out.c: Likewise. * dwarf2out.h: Likewise. * emit-rtl.c: Likewise. * except.c: Likewise. * except.h: Likewise. * expr.c: Likewise. * expr.h: Likewise. * final.c: Likewise. * fold-const.c: Likewise. * function.c: Likewise. * function.h: Likewise. * fwprop.c: Likewise. * gcc.c: Likewise. * gcse.c: Likewise. * genattr.c: Likewise. * genattrtab.c: Likewise. * genautomata.c: Likewise. * genextract.c: Likewise. * genopinit.c: Likewise * ggc-common.c: Likewise. * ggc.h: Likewise. * gimple-low.c: Likewise. * gimple-ssa-strength-reduction.c: Likewise. * gimple-streamer-in.c: Likewise. * gimple.c: Likewise. * gimple.h: Likewise. * gimplify.c: Likewise. * graph.c: Likewise. * graphds.c: Likewise. * graphds.h: Likewise. * graphite-blocking.c: Likewise. * graphite-clast-to-gimple.c: Likewise. * graphite-dependences.c: Likewise. * graphite-interchange.c: Likewise. * graphite-optimize-isl.c: Likewise. * graphite-poly.c: Likewise. * graphite-poly.h: Likewise. * graphite-scop-detection.c: Likewise. * graphite-scop-detection.h: Likewise. * graphite-sese-to-poly.c: Likewise. * graphite.c: Likewise. * godump.c: Likewise. * haifa-sched.c: Likewise. * hw-doloop.c: Likewise. * hw-doloop.h: Likewise. * ifcvt.c: Likewise. * insn-addr.h: Likewise. * ipa-cp.c: Likewise. * ipa-inline-analysis.c: Likewise. * ipa-inline-transform.c: Likewise. * ipa-inline.c: Likewise. * ipa-inline.h: Likewise. * ipa-prop.c: Likewise. * ipa-prop.h: Likewise. * ipa-pure-const.c: Likewise. * ipa-ref-inline.h: Likewise. * ipa-ref.c: Likewise. * ipa-ref.h: Likewise. * ipa-reference.c: Likewise. * ipa-split.c: Likewise. * ipa-utils.c: Likewise. * ipa-utils.h: Likewise. * ipa.c: Likewise. * ira-build.c: Likewise. * ira-color.c: Likewise. * ira-emit.c: Likewise. * ira-int.h: Likewise. * ira.c: Likewise. * loop-invariant.c: Likewise. * loop-unroll.c: Likewise. * lower-subreg.c: Likewise. * lra-lives.c: Likewise. * lra.c: Likewise. * lto-cgraph.c: Likewise. * lto-section-out.c: Likewise. * lto-streamer-in.c: Likewise. * lto-streamer-out.c: Likewise. * lto-streamer.h: Likewise. * lto-symtab.c: Likewise. * mcf.c: Likewise. * modulo-sched.c: Likewise. * omp-low.c: Likewise. * opts-common.c: Likewise. * opts-global.c: Likewise. * opts.c: Likewise. * opts.h: Likewise. * passes.c: Likewise. * predict.c: Likewise. * print-tree.c: Likewise. * profile.c: Likewise. * profile.h: Likewise. * read-rtl.c: Likewise. * ree.c: Likewise. * reg-stack.c: Likewise. * regrename.c: Likewise. * regrename.h: Likewise. * reload.c: Likewise. * reload.h: Likewise. * reload1.c: Likewise. * rtl.h: Likewise. * sched-deps.c: Likewise. * sched-int.h: Likewise. * sdbout.c: Likewise. * sel-sched-dump.c: Likewise. * sel-sched-ir.c: Likewise. * sel-sched-ir.h: Likewise. * sel-sched.c: Likewise. * sese.c: Likewise. * sese.h: Likewise. * statistics.h: Likewise. * stmt.c: Likewise. * stor-layout.c: Likewise. * store-motion.c: Likewise. * tlink.c: Likewise. * toplev.c: Likewise. * trans-mem.c: Likewise. * tree-browser.c: Likewise. * tree-call-cdce.c: Likewise. * tree-cfg.c: Likewise. * tree-cfgcleanup.c: Likewise. * tree-chrec.c: Likewise. * tree-chrec.h: Likewise. * tree-complex.c: Likewise. * tree-data-ref.c: Likewise. * tree-data-ref.h: Likewise. * tree-dfa.c: Likewise. * tree-diagnostic.c: Likewise. * tree-dump.c: Likewise. * tree-eh.c: Likewise. * tree-emutls.c: Likewise. * tree-flow.h: Likewise. * tree-if-conv.c: Likewise. * tree-inline.c: Likewise. * tree-inline.h: Likewise. * tree-into-ssa.c: Likewise. * tree-iterator.c: Likewise. * tree-loop-distribution.c: Likewise. * tree-mudflap.c: Likewise. * tree-optimize.c: Likewise. * tree-outof-ssa.c: Likewise. * tree-parloops.c: Likewise. * tree-phinodes.c: Likewise. * tree-predcom.c: Likewise. * tree-pretty-print.c: Likewise. * tree-scalar-evolution.c: Likewise. * tree-sra.c: Likewise. * tree-ssa-address.c: Likewise. * tree-ssa-alias.c: Likewise. * tree-ssa-ccp.c: Likewise. * tree-ssa-coalesce.c: Likewise. * tree-ssa-dce.c: Likewise. * tree-ssa-dom.c: Likewise. * tree-ssa-forwprop.c: Likewise. * tree-ssa-live.c: Likewise. * tree-ssa-live.h: Likewise. * tree-ssa-loop-im.c: Likewise. * tree-ssa-loop-ivcanon.c: Likewise. * tree-ssa-loop-ivopts.c: Likewise. * tree-ssa-loop-manip.c: Likewise. * tree-ssa-loop-niter.c: Likewise. * tree-ssa-loop-prefetch.c: Likewise. * tree-ssa-math-opts.c: Likewise. * tree-ssa-operands.c: Likewise. * tree-ssa-phiopt.c: Likewise. * tree-ssa-phiprop.c: Likewise. * tree-ssa-pre.c: Likewise. * tree-ssa-propagate.c: Likewise. * tree-ssa-reassoc.c: Likewise. * tree-ssa-sccvn.c: Likewise. * tree-ssa-sccvn.h: Likewise. * tree-ssa-strlen.c: Likewise. * tree-ssa-structalias.c: Likewise. * tree-ssa-tail-merge.c: Likewise. * tree-ssa-threadedge.c: Likewise. * tree-ssa-threadupdate.c: Likewise. * tree-ssa-uncprop.c: Likewise. * tree-ssa-uninit.c: Likewise. * tree-ssa.c: Likewise. * tree-ssanames.c: Likewise. * tree-stdarg.c: Likewise. * tree-streamer-in.c: Likewise. * tree-streamer-out.c: Likewise. * tree-streamer.c: Likewise. * tree-streamer.h: Likewise. * tree-switch-conversion.c: Likewise. * tree-vect-data-refs.c: Likewise. * tree-vect-generic.c: Likewise. * tree-vect-loop-manip.c: Likewise. * tree-vect-loop.c: Likewise. * tree-vect-patterns.c: Likewise. * tree-vect-slp.c: Likewise. * tree-vect-stmts.c: Likewise. * tree-vectorizer.c: Likewise. * tree-vectorizer.h: Likewise. * tree-vrp.c: Likewise. * tree.c: Likewise. * tree.h: Likewise. * value-prof.c: Likewise. * value-prof.h: Likewise. * var-tracking.c: Likewise. * varasm.c: Likewise. * varpool.c: Likewise. * vmsdbgout.c: Likewise. * config/bfin/bfin.c: Likewise. * config/c6x/c6x.c: Likewise. * config/darwin.c: Likewise. * config/i386/i386.c: Likewise. * config/ia64/ia64.c: Likewise. * config/mep/mep.c: Likewise. * config/mips/mips.c: Likewise. * config/pa/pa.c: Likewise. * config/rs6000/rs6000-c.c: Likewise. * config/rs6000/rs6000.c: Likewise. * config/rx/rx.c: Likewise. * config/spu/spu-c.c: Likewise. * config/vms/vms.c: Likewise. * config/vxworks.c: Likewise. * config/epiphany/resolve-sw-modes.c: Likewise. From-SVN: r193595
2012-10-05tree-vectorizer.h (vect_estimate_min_profitable_iters): Remove.Jan Hubicka1-1/+0
* tree-vectorizer.h (vect_estimate_min_profitable_iters): Remove. * tree-vect-loop.c (vect_estimate_min_profitable_iters): Declare here. (vect_analyze_loop_operations): Use loop count estimate to rule out unprofitable vectorization. (vect_estimate_min_profitable_iters): Return ret_min_profitable_estimate. From-SVN: r192116
2012-10-01invoke.texi: Add documentation for the new option.Sharad Singhai1-5/+2
2012-09-30 Sharad Singhai <singhai@google.com> * doc/invoke.texi: Add documentation for the new option. * tree-dump.c: Move general dump file related functionality into dumpfile.c. Remove unneeded headers. * tree-dump.h: Move function declarations into dumpfile.h. * dumpfile.h: Include "line-map.h". Add defines for MSG flags. (struct dump_file_info): Move here from tree-dump.c. Rename flags to pflags, state to pstate, stream to pstream, filename to pfilename. All callers updated. Add alt_flags, alt_state, alt_filenmae, alt_stream. * tree-vectorizer.c: Include "dumpfile.h". Remove vect_dump. (vect_set_dump_settings): Remove. (vect_print_dump_info): Ditto. * tree-vectorizer.h: Remove declaration of vect_dump and vect_print_dump_info. * tree-vect-loop.c: Include "dumpfile.h". Use new dump style. * tree-vect-data-refs.c: Ditto. * tree-vect-stmts.c: Ditto. * tree-vect-slp.c: Ditto. * tree-vect-patterns.c: Ditto. * tree-vect-loop-manip.c: Ditto. * testsuite/gcc.target/i386/vect-double-1.c: Fix test. * opts.c (vect_set_verbosity_level): Remove. (common_handle_option): Handle -fopt-info flag. Deprecate -ftree-vectorizer-verbose. * tree-parloops.c (gather_scalar_reductions): Remove reference to vect_dump. * flag-types.h: Remove vect_verbosity_levels. * common.opt: Add -fopt-info. Deprecate -ftree-vectorizer-verbose. * opts-global.c (dump_remap_tree_vectorizer_verbose): New function. (handle_common_deferred_options): Handle -fopt-info and -ftree-vectorizer-verbose. * Makefile.in: Add dumpfile.o. (tree-dump.o): Update dependencies. (tree-vect-loop.o): Ditto. (tree-vect-loop-manip.o): Ditto. (tree-vect-slp.o): Ditto. (tree-vect-stmts.o): Ditto. (tree-vectorizer.o): Ditto. (opts.o): Ditto. * passes.c (finish_optimization_passes): Instead of using dump_begin/dump_end, use dump_start/dump_finish. Do not use dump_file. (pass_init_dump_file): Ditto. * c-decl.c (c_write_global_declarations): Use a different method to determine if the dump has ben initialized. * decl2.c (cp_write_global_declarations): Use a different method to determine if the dump has ben initialized. From-SVN: r191883
2012-09-10Remove unnecessary VEC function overloads.Diego Novillo1-1/+1
Several VEC member functions that accept an element 'T' used to have two overloads: one taking 'T', the second taking 'T *'. This used to be needed because of the interface dichotomy between vectors of objects and vectors of pointers. In the past, vectors of pointers would use pass-by-value semantics, but vectors of objects would use pass-by-reference semantics. This is no longer necessary, but the distinction had remained. The main side-effect of this change is some code reduction in code that manipulates vectors of objects. For instance, - struct iterator_use *iuse; - - iuse = VEC_safe_push (iterator_use, heap, iterator_uses, NULL); - iuse->iterator = iterator; - iuse->ptr = ptr; + struct iterator_use iuse = {iterator, ptr}; + VEC_safe_push (iterator_use, heap, iterator_uses, iuse); Compile time performance was not affected. Tested on x86_64 and ppc64. Also built all-gcc on all targets using VEC routines: arm, bfin, c6x, epiphany, ia64, mips, sh, spu, and vms. 2012-09-10 Diego Novillo <dnovillo@google.com> * vec.h (vec_t::quick_push): Remove overload that accepts 'T *'. Update all users. (vec_t::safe_push): Likewise. (vec_t::quick_insert): Likewise. (vec_t::lower_bound): Likewise. (vec_t::safe_insert): Likewise. (vec_t::replace): Change second argument to 'T &'. From-SVN: r191165
2012-07-31re PR tree-optimization/53773 (Vectorizer generates non-canonical multiplies)Bill Schmidt1-0/+7
gcc: 2012-07-31 Bill Schmidt <wschmidt@linux.ibm.com> PR tree-optimization/53773 * tree-vectorizer.h (struct _loop_vec_info): Add operands_swapped. (LOOP_VINFO_OPERANDS_SWAPPED): New macro. * tree-vect-loop.c (new_loop_vec_info): Initialize LOOP_VINFO_OPERANDS_SWAPPED field. (destroy_loop_vec_info): Restore canonical form. (vect_is_slp_reduction): Set LOOP_VINFO_OPERANDS_SWAPPED field. (vect_is_simple_reduction_1): Likewise. gcc/testsuite: 2012-07-31 Bill Schmidt <wschmidt@linux.ibm.com> PR tree-optimization/53773 * testsuite/gcc.dg/vect/pr53773.c: New test. From-SVN: r190007
2012-07-25tm.texi: Regenerate.Bill Schmidt1-49/+32
2012-07-24 Bill Schmidt <wschmidt@linux.ibm.com> * doc/tm.texi: Regenerate. * targhooks.c (default_init_cost): Add prologue and epilogue costs. (default_add_stmt_cost): Likewise; also handle NULL stmt_info. (default_finish_cost): Add prologue and epilogue costs. * targhooks.h (default_add_stmt_cost): Change parameter list. (default_finish_cost): Likewise. * target.def (init_cost): Change documentation string. (add_stmt_cost): Change documentation string and parameter list. (finish_cost): Likewise. * target.h (vect_cost_model_location): New enum. * tree-vectorizer.h (struct _slp_tree): Remove cost substruct. (struct _slp_instance): Remove cost substruct; rename stmt_cost_vec to body_cost_vec. (SLP_INSTANCE_OUTSIDE_OF_LOOP_COST): Remove. (SLP_INSTANCE_STMT_COST_VEC): Rename to SLP_INSTANCE_BODY_COST_VEC. (SLP_TREE_OUTSIDE_OF_LOOP_COST): Remove. (struct _vect_peel_extended_info): Rename stmt_cost_vec to body_cost_vec. (struct _stmt_vec_info): Remove cost substruct. (STMT_VINFO_OUTSIDE_OF_LOOP_COST): Remove. (stmt_vinfo_set_outside_of_loop_cost): Remove. (builtin_vectorization_cost): New function. (vect_get_stmt_cost): Change to use builtin_vectorization_cost. (add_stmt_cost): Change parameter list. (finish_cost): Likewise. (vect_model_simple_cost): Likewise. (vect_model_store_cost): Likewise. (vect_model_load_cost): Likewise. (record_stmt_cost): Likewise. (vect_get_load_cost): Likewise. (vect_get_known_peeling_cost): Likewise. * tree-vect-loop.c (vect_get_known_peeling_cost): Change parameter list; call record_stmt_cost for prologue and epilogue costs. (vect_estimate_min_profitable_iters): Call add_stmt_cost for prologue and epilogue costs; remove computation of vec_outside_cost; return vec_prologue_cost and vec_epilogue_cost from finish_cost. (vect_model_reduction_cost): Revise call to add_stmt_cost for body costs; call add_stmt_cost for prologue and epilogue costs. (vect_model_induction_cost): Revise call to add_stmt_cost for body costs; call add_stmt_cost for prologue costs. * tree-vect-data-refs.c (vect_get_data_access_cost): Change parameter list for function and arguments for calls to vect_get_load_cost and vect_get_store_cost. (vect_peeling_hash_get_lowest_cost): Change argument list for calls to vect_get_data_access_cost and vect_get_known_peeling_cost; use temporary vectors prologue_cost_vec and epilogue_cost_vec for the latter call and discard their results; rename stmt_cost_vec to body_cost_vec; correct possible storage leak for body_cost_vec. (vect_peeling_hash_choose_best_peeling): Rename stmt_cost_vec to body_cost_vec. (vect_enhance_data_refs_alignment): Rename stmt_cost_vec to body_cost_vec; add extra dummy parameter on calls to vect_get_data_access_cost; tolerate null si->stmt; add vect_body to argument list on call to add_stmt_cost. * tree-vect-stmts.c (record_stmt_cost): Change parameter list; rename stmt_cost_vec to body_cost_vec; tolerate null stmt_info; call builtin_vectorization_cost; add "where" parameter on call to add_stmt_cost. (vect_model_simple_cost): Change parameter list; call record_stmt_cost for prologue costs; remove call to stmt_vinfo_set_outside_of_loop_cost; rename stmt_cost_vec to body_cost_vec. (vect_model_promotion_demotion_cost): Add vect_body argument to call to add_stmt_cost; call add_stmt_cost for prologue costs; remove call to stmt_vinfo_set_outside_of_loop_cost. (vect_model_store_cost): Change parameter list; call record_stmt_cost for prologue costs; add vect_body argument to call to record_stmt_cost; rename stmt_cost_vec to body_cost_vec; remove call to stmt_vinfo_set_outside_of_loop_cost. (vect_get_store_cost): Rename stmt_cost_vec to body_cost_vec; add vect_body argument to calls to record_stmt_cost. (vect_model_load_cost): Change parameter list; rename stmt_cost_vec to body_cost_vec; add vect_body argument to calls to record_stmt_cost; remove call to stmt_vinfo_set_outside_of_loop_cost. (vect_get_load_cost): Change parameter list; rename stmt_cost_vec to body_cost_vec; add vect_body argument to calls to record_stmt_cost; call record_stmt_cost for prologue costs. (vectorizable_store): Change argument list for call to vect_model_store_cost. (vectorizable_load): Change argument list for call to vect_model_load_cost. (new_stmt_vec_info): Remove assignment to STMT_VINFO_OUTSIDE_OF_LOOP_COST. * config/spu/spu.c (spu_init_cost): Add prologue and epilogue costs. (spu_add_stmt_cost): Likewise; also handle NULL stmt_info. (spu_finish_cost): Add prologue and epilogue costs. * config/i386/i386.c (i386_init_cost): Add prologue and epilogue costs. (i386_add_stmt_cost): Likewise; also handle NULL stmt_info. (i386_finish_cost): Add prologue and epilogue costs. * config/rs6000/rs6000.c (rs6000_init_cost): Add prologue and epilogue costs. (rs6000_add_stmt_cost): Likewise; also handle NULL stmt_info. (rs6000_finish_cost): Add prologue and epilogue costs. * tree-vect-slp.c (vect_free_slp_instance): Rename SLP_INSTANCE_STMT_COST_VEC to SLP_INSTANCE_BODY_COST_VEC. (vect_create_new_slp_node): Remove assignment to SLP_TREE_OUTSIDE_OF_LOOP_COST. (vect_get_and_check_slp_defs): Change parameter list; change argument lists to calls to vect_model_store_cost and vect_model_simple_cost. (vect_build_slp_tree): Change parameter list; change argument lists to calls to vect_model_load_cost, vect_get_and_check_slp_defs, and recursive self-calls; remove setting of outside_cost from SLP_TREE_OUTSIDE_OF_LOOP_COST; add vect_body argument to call to record_stmt_cost. (vect_analyze_slp_instance): Rename stmt_cost_vec to body_cost_vec; rename SLP_INSTANCE_STMT_COST_VEC to SLP_INSTANCE_BODY_COST_VEC; remove assignment to SLP_INSTANCE_OUTSIDE_OF_LOOP_COST; record SLP prologue costs. (vect_bb_vectorization_profitable_p): Rename stmt_cost_vec to body_cost_vec; handle null ci->stmt; add vect_body argument to call to add_stmt_cost; simplify calls to targetm.vectorize. builtin_vectorization_cost; return vec_prologue_cost and vec_epilogue_cost from finish_cost. (vect_update_slp_costs_according_to_vf): Rename stmt_cost_vec to body_cost_vec; add vect_body argument to call to add_stmt_cost. From-SVN: r189836
2012-07-17Clean up supportable_widening_operation.Richard Henderson1-3/+2
* tree-vect-stmts.c (supportable_widening_operation): Remove decl parameters. (vectorizable_conversion): Update supportable_widening_operation call. * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Likewise. (vect_recog_widen_shift_pattern): Likewise. * tree-vectorizer.h: Update decl. From-SVN: r189574
2012-07-13tm.texi: Regenerate.Bill Schmidt1-29/+92
2012-07-13 Bill Schmidt <wschmidt@linux.ibm.com> * doc/tm.texi: Regenerate. * doc/tm.texi.in (TARGET_VECTORIZE_INIT_COST): New hook. (TARGET_VECTORIZE_ADD_STMT_COST): Likewise. (TARGET_VECTORIZE_FINISH_COST): Likewise. (TARGET_VECTORIZE_DESTROY_COST_DATA): Likewise. * targhooks.c (default_init_cost): New function. (default_add_stmt_cost): Likewise. (default_finish_cost): Likewise. (default_destroy_cost_data): Likewise. * targhooks.h (default_init_cost): New decl. (default_add_stmt_cost): Likewise. (default_finish_cost): Likewise. (default_destroy_cost_data): Likewise. * target.def (init_cost): New DEFHOOK. (add_stmt_cost): Likewise. (finish_cost): Likewise. (destroy_cost_data): Likewise. * target.h (struct _stmt_vec_info): New extern decl. (stmt_vectype): Likewise. (stmt_in_inner_loop_p): Likewise. * tree-vectorizer.h (stmt_info_for_cost): New struct/typedef. (stmt_vector_for_cost): New VEC/typedef. (add_stmt_info_to_vec): New function. (struct _slp_tree): Remove cost.inside_of_loop field. (struct _slp_instance): Remove cost.inside_of_loop field; add stmt_cost_vec field. (SLP_INSTANCE_INSIDE_OF_LOOP_COST): Remove macro. (SLP_INSTANCE_STMT_COST_VEC): New accessor macro. (SLP_TREE_INSIDE_OF_LOOP_COST): Remove macro. (struct _vect_peel_extended_info): Add stmt_cost_vec field. (struct _loop_vec_info): Add target_cost_data field. (LOOP_VINFO_TARGET_COST_DATA): New accessor macro. (struct _bb_vec_info): Add target_cost_data field. (BB_VINFO_TARGET_COST_DATA): New accessor macro. (struct _stmt_vec_info): Remove cost.inside_of_loop field. (STMT_VINFO_INSIDE_OF_LOOP_COST): Remove macro. (stmt_vinfo_set_inside_of_loop_cost): Remove function. (init_cost): New function. (add_stmt_cost): Likewise. (finish_cost): Likewise. (destroy_cost_data): Likewise. (vect_model_simple_cost): Change parameter list. (vect_model_store_cost): Likewise. (vect_model_load_cost): Likewise. (record_stmt_cost): New extern decl. (vect_get_load_cost): Change parameter list. (vect_get_store_cost): Likewise. * tree-vect-loop.c (new_loop_vec_info): Call init_cost. (destroy_loop_vec_info): Call destroy_cost_data. (vect_estimate_min_profitable_iters): Remove old calculation of inside costs; call finish_cost instead. (vect_model_reduction_cost): Call add_stmt_cost instead of old inside-costs calculation. (vect_model_induction_cost): Likewise. * tree-vect-data-refs.c (vect_get_data_access_cost): Change to return a stmt_vector_for_cost; modify calls to vect_get_load_cost and vect_get_store_cost to obtain the value to return. (vect_peeling_hash_get_lowest_cost): Obtain a stmt_cost_vec from vect_get_data_access_cost and store it in the minimum peeling structure. (vect_peeling_hash_choose_best_peeling): Change the parameter list to add a (stmt_vector_for_cost *) output parameter, and set its value. (vect_enhance_data_refs_alignment): Ignore the new return value from calls to vect_get_data_access_cost; obtain stmt_cost_vec from vect_peeling_hash_choose_best_peeling and pass its contents to the target cost model. * tree-vect-stmts.c (stmt_vectype): New function. (stmt_in_inner_loop_p): Likewise. (record_stmt_cost): Likewise. (vect_model_simple_cost): Add stmt_cost_vec parameter; call record_stmt_cost instead of old calculation; don't call stmt_vinfo_set_inside_of_loop_cost. (vect_model_promotion_demotion_cost): Call add_stmt_cost instead of old calculation; don't call stmt_vinfo_set_inside_of_loop_cost. (vect_model_store_cost): Add stmt_cost_vec parameter; call record_stmt_cost instead of old calculation; add stmt_cost_vec parameter to vect_get_store_cost call; don't call stmt_vinfo_set_inside_of_loop_cost. (vect_get_store_cost): Add stmt_cost_vec parameter; call record_stmt_cost instead of old calculation. (vect_model_load_cost): Add stmt_cost_vec parameter; call record_stmt_cost instead of old calculation; add stmt_cost_vec parameter to vect_get_load_cost call; don't call stmt_vinfo_set_inside_of_loop_cost. (vect_get_load_cost): Add stmt_cost_vec parameter; call record_stmt_cost instead of old calculation. (vectorizable_call): Add NULL parameter to vect_model_simple_cost call. (vectorizable_conversion): Likewise. (vectorizable_assignment): Likewise. (vectorizable_shift): Likewise. (vectorizable_operation): Likewise. (vectorizable_store): Add NULL parameter to vect_model_store_cost call. (vectorizable_load): Add NULL parameter to vect_model_load_cost call. (new_stmt_vec_info): Don't set STMT_VINFO_INSIDE_OF_LOOP_COST. * config/spu/spu.c (TARGET_VECTORIZE_INIT_COST): New macro def. (TARGET_VECTORIZE_ADD_STMT_COST): Likewise. (TARGET_VECTORIZE_FINISH_COST): Likewise. (TARGET_VECTORIZE_DESTROY_COST_DATA): Likewise. (spu_init_cost): New function. (spu_add_stmt_cost): Likewise. (spu_finish_cost): Likewise. (spu_destroy_cost_data): Likewise. * config/i386/i386.c (ix86_init_cost): New function. (ix86_add_stmt_cost): Likewise. (ix86_finish_cost): Likewise. (ix86_destroy_cost_data): Likewise. (TARGET_VECTORIZE_INIT_COST): New macro def. (TARGET_VECTORIZE_ADD_STMT_COST): Likewise. (TARGET_VECTORIZE_FINISH_COST): Likewise. (TARGET_VECTORIZE_DESTROY_COST_DATA): Likewise. * config/rs6000/rs6000.c (TARGET_VECTORIZE_INIT_COST): New macro def. (TARGET_VECTORIZE_ADD_STMT_COST): Likewise. (TARGET_VECTORIZE_FINISH_COST): Likewise. (TARGET_VECTORIZE_DESTROY_COST_DATA): Likewise. (rs6000_init_cost): New function. (rs6000_add_stmt_cost): Likewise. (rs6000_finish_cost): Likewise. (rs6000_destroy_cost_data): Likewise. * tree-vect-slp.c (vect_free_slp_instance): Free stmt_cost_vec. (vect_create_new_slp_node): Don't set SLP_TREE_INSIDE_OF_LOOP_COST. (vect_get_and_check_slp_defs): Add stmt_cost_vec parameter; add stmt_cost_vec parameter to vect_model_store_cost and vect_model_simple_cost calls. (vect_build_slp_tree): Remove inside_cost parameter; add stmt_cost_vec parameter; add stmt_cost_vec parameter to vect_get_and_check_slp_defs, vect_model_load_cost, and recursive vect_build_slp_tree calls; prevent calculating cost more than once for loads; remove inside_cost parameter from recursive vect_build_slp_tree calls; call record_stmt_cost instead of old calculation. (vect_analyze_slp_instance): Allocate stmt_cost_vec and save it with the instance; free it on premature exit; remove inside_cost parameter from vect_build_slp_tree call; add stmt_cost_vec parameter to vect_build_slp_tree call; don't set SLP_INSTANCE_INSIDE_OF_LOOP_COST. (new_bb_vec_info): Call init_cost. (destroy_bb_vec_info): Call destroy_cost_data. (vect_bb_vectorization_profitable_p): Call add_stmt_cost for each statement recorded with an SLP instance; call finish_cost instead of the old calculation. (vect_update_slp_costs_according_to_vf): Record statement costs from SLP instances, multiplying by the appropriate number of copies; don't update SLP_INSTANCE_INSIDE_OF_LOOP_COST. From-SVN: r189460
2012-06-14tree-vectorizer.h (vect_get_stmt_cost): Move from tree-vect-stmts.c.Bill Schmidt1-2/+14
2012-06-14 Bill Schmidt <wschmidt@linux.ibm.com> * tree-vectorizer.h (vect_get_stmt_cost): Move from tree-vect-stmts.c. (cost_for_stmt): Remove decl. (vect_get_single_scalar_iteration_cost): Correct typo in name. * tree-vect-loop.c (vect_get_cost): Remove. (vect_get_single_scalar_iteration_cost): Correct typo in name; use vect_get_stmt_cost rather than vect_get_cost. (vect_get_known_peeling_cost): Use vect_get_stmt_cost rather than vect_get_cost. (vect_estimate_min_profitable_iters): Correct typo in call to vect_get_single_scalar_iteration_cost; use vect_get_stmt_cost rather than vect_get_cost. (vect_model_reduction_cost): Use vect_get_stmt_cost rather than vect_get_cost. (vect_model_induction_cost): Likewise. * tree-vect-data-refs.c (vect_peeling_hash_get_lowest_cost): Correct typo in call to vect_get_single_scalar_iteration_cost. * tree-vect-stmts.c (vect_get_stmt_cost): Move to tree-vectorizer.h. (cost_for_stmt): Remove unnecessary function. * Makefile.in (TREE_VECTORIZER_H): Update dependencies. From-SVN: r188611
2012-05-09tree-vectorizer.h (vect_loop_versioning): Adjust prototype.Richard Guenther1-3/+3
2012-05-09 Richard Guenther <rguenther@suse.de> * tree-vectorizer.h (vect_loop_versioning): Adjust prototype. (vect_do_peeling_for_loop_bound): Likewise. (vect_do_peeling_for_alignment): Likewise. * tree-vect-loop-manip.c (conservative_cost_threshold): Remove. (vect_do_peeling_for_loop_bound): Get check_profitability and threshold as parameters. (vect_do_peeling_for_alignment): Likewise. (vect_loop_versioning): Likewise. * tree-vect-loop.c (vect_transform_loop): Compute check_profitability and threshold here. Control where to put the check here. From-SVN: r187323
2012-04-25tree-vectorizer.h (vect_loop_versioning): Adjust prototype.Richard Guenther1-1/+1
2012-04-25 Richard Guenther <rguenther@suse.de> * tree-vectorizer.h (vect_loop_versioning): Adjust prototype. * tree-vect-loop.c (vect_transform_loop): Adjust. * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Record the maximum number of iterations for the epilogue loop. (vect_loop_versioning): Remove case re-using the peeled epilogue loop. From-SVN: r186813
2012-04-17re PR tree-optimization/18437 (vectorizer failed for matrix multiplication)Michael Matz1-0/+3
PR tree-optimization/18437 * tree-vectorizer.h (_stmt_vec_info.stride_load_p): New member. (STMT_VINFO_STRIDE_LOAD_P): New accessor. (vect_check_strided_load): Declare. * tree-vect-data-refs.c (vect_check_strided_load): New function. (vect_analyze_data_refs): Use it to accept strided loads. * tree-vect-stmts.c (vectorizable_load): Ditto and handle them. testsuite/ * gfortran.dg/vect/rnflow-trs2a2.f90: New test. From-SVN: r186530
2012-04-10tree-vectorizer.h (_loop_vec_info.strided_stores): Rename to grouped_stores.Michael Matz1-12/+12
* tree-vectorizer.h (_loop_vec_info.strided_stores): Rename to grouped_stores. (LOOP_VINFO_STRIDED_STORES): Rename to LOOP_VINFO_GROUPED_STORES. (struce _bb_vec_info.strided_stores): Rename to grouped_stores. (BB_VINFO_STRIDED_STORES): Rename to BB_VINFO_GROUPED_STORES. (STMT_VINFO_STRIDED_ACCESS): Rename to STMT_VINFO_GROUPED_ACCESS. (vect_strided_store_supported): Rename to vect_grouped_store_supported. (vect_strided_load_supported): Rename to vect_grouped_load_supported. (vect_transform_strided_load): Rename to vect_transform_grouped_load. (vect_record_strided_load_vectors): Rename to vect_record_grouped_load_vectors. * tree-vect-data-refs.c (vect_update_misalignment_for_peel): Rename use of above macros. (vect_verify_datarefs_alignment): Ditto. (vector_alignment_reachable_p): Ditto. (vect_peeling_hash_get_lowest_cost): Ditto. (vect_enhance_data_refs_alignment): Ditto. (vect_analyze_group_access): Ditto and rename stride to groupsize. (vect_analyze_data_ref_access): Rename "strided" to "grouped". (vect_strided_store_supported): Rename to vect_grouped_store_supported. (vect_strided_load_supported): Rename to vect_grouped_load_supported. (vect_transform_strided_load): Rename to vect_transform_grouped_load. (vect_record_strided_load_vectors): Rename to vect_record_grouped_load_vectors. * tree-vect-loop.c (new_loop_vec_info): Rename use of above macros. (destroy_loop_vec_info): Ditto. (vect_transform_loop): Ditto and rename strided_store to grouped_store. * tree-vect-slp.c (vect_build_slp_tree): Rename use of above macros. (vect_analyze_slp): Ditto. (new_bb_vec_info): Ditto. (destroy_bb_vec_info): Ditto. (vect_schedule_slp_instance): Ditto and rename strided_store to grouped_store. * tree-vect-stmts.c (vect_cost_strided_group_size): Rename to vect_cost_group_size. (vect_model_store_cost): Rename use of above macros and call to vect_cost_strided_group_size. (vect_model_load_cost): Ditto. (vectorizable_store): Ditto, rename strided_store to grouped_store and calls to renamed tree-vectorizer.h functions. (vectorizable_load): Ditto. (vect_transform_stmt): Rename use of above macros and strided_store to grouped_store. testsuite/ * gcc.dg/vect/vect-outer-1-big-array.c: Adjust. * gcc.dg/vect/vect-outer-1.c: Adjust. * gcc.dg/vect/vect-outer-1a-big-array.c: Adjust. * gcc.dg/vect/vect-outer-1a.c: Adjust. * gcc.dg/vect/vect-outer-1b-big-array.c: Adjust. * gcc.dg/vect/vect-outer-1b.c: Adjust. * gcc.dg/vect/vect-outer-2b.c: Adjust. * gcc.dg/vect/vect-outer-3b.c: Adjust. From-SVN: r186285
2012-03-23re PR tree-optimization/52678 (internal compiler error: in ↵Richard Guenther1-0/+8
vect_update_ivs_after_vectorizer, at tree-vect-loop-manip.c:1842) 2012-03-23 Richard Guenther <rguenther@suse.de> PR tree-optimization/52678 * tree-vectorizer.h (struct _stmt_vec_info): Add loop_phi_evolution_part member. (STMT_VINFO_LOOP_PHI_EVOLUTION_PART): New define. * tree-vect-loop.c (vect_analyze_scalar_cycles_1): Initialize STMT_VINFO_LOOP_PHI_EVOLUTION_PART. * tree-vect-loop-manip.c (vect_update_ivs_after_vectorizer): Use the cached evolution part and the PHI nodes value from the loop preheader edge instead of re-analyzing the evolution. * gfortran.dg/pr52678.f: New testcase. From-SVN: r185734
2012-03-15tree-vectorizer.h (vect_pattern_recog): Add new argument.Ira Rosen1-1/+1
ChangeLog: * tree-vectorizer.h (vect_pattern_recog): Add new argument. * tree-vect-loop.c (vect_analyze_loop_2): Update call to vect_pattern_recog. * tree-vect-patterns.c (widened_name_p): Pass basic block info to vect_is_simple_use. (vect_recog_dot_prod_pattern): Fail for basic blocks. (vect_recog_widen_sum_pattern): Likewise. (vect_handle_widen_op_by_const): Support basic blocks. (vect_operation_fits_smaller_type, vect_recog_over_widening_pattern): Likewise. (vect_recog_vector_vector_shift_pattern): Support basic blocks. Update call to vect_is_simple_use. (vect_recog_mixed_size_cond_pattern): Support basic blocks. Add printing. (check_bool_pattern): Add an argument, update call to vect_is_simple_use and the recursive calls. (vect_recog_bool_pattern): Update relevant function calls. Add printing. (vect_mark_pattern_stmts): Update calls to new_stmt_vec_info. (vect_pattern_recog_1): Check for reduction only in loops. (vect_pattern_recog): Add new argument. Support basic blocks. * tree-vect-stmts.c (vectorizable_conversion): Pass basic block info to vect_is_simple_use_1. * tree-vect-slp.c (vect_get_and_check_slp_defs): Support basic blocks. (vect_slp_analyze_bb_1): Call vect_pattern_recog. testsuite/ChangeLog: * gcc.dg/vect/bb-slp-pattern-1.c: New test. * gcc.dg/vect/bb-slp-pattern-2.c: New test. Co-Authored-By: Ulrich Weigand <ulrich.weigand@linaro.org> From-SVN: r185436
2012-02-05re PR tree-optimization/52091 (ICE: verify_gimple failed)Ira Rosen1-3/+5
PR tree-optimization/52091 * tree-vectorizer.h (vect_is_simple_use): Add an argument. (vect_is_simple_use_1): Likewise. * tree-vect-loop.c (vectorizable_reduction): Update calls to vect_is_simple_use_1 and vect_is_simple_use. (vectorizable_live_operation): Likewise. * tree-vect-patterns.c (widened_name_p, vect_recog_vector_vector_shift_pattern, check_bool_pattern): Likewise. * tree-vect-stmts.c (process_use, vect_get_vec_def_for_operand, vectorizable_call, vectorizable_conversion, vectorizable_assignment, vectorizable_shift, vectorizable_operation, vectorizable_store, vectorizable_load): Likewise. (vect_is_simple_cond): Add an argument, pass it to vect_is_simple_use_1. (vectorizable_condition): Update calls to vect_is_simple_cond, vect_is_simple_use. (vect_is_simple_use): Add an argument, the statement in which OPERAND is used. Check that if OPERAND's def stmt is a double reduction phi node, the use is a phi node too. (vect_is_simple_use_1): Add an argument, pass it to vect_is_simple_use. * tree-vect-slp.c (vect_get_and_check_slp_defs): Update a call to vect_is_simple_use. From-SVN: r183902
2011-12-23Delete VEC_INTERLEAVE_*_EXPR.Jakub Jelinek1-0/+1
* tree.def (VEC_INTERLEAVE_HIGH_EXPR, VEC_INTERLEAVE_LOW_EXPR): Remove. * gimple-pretty-print.c (dump_binary_rhs): Don't handle VEC_INTERLEAVE_HIGH_EXPR and VEC_INTERLEAVE_LOW_EXPR. * expr.c (expand_expr_real_2): Likewise. * tree-cfg.c (verify_gimple_assign_binary): Likewise. * cfgexpand.c (expand_debug_expr): Likewise. * tree-inline.c (estimate_operator_cost): Likewise. * tree-pretty-print.c (dump_generic_node): Likewise. * tree-vect-generic.c (expand_vector_operations_1): Likewise. * fold-const.c (fold_binary_loc): Likewise. * doc/generic.texi (VEC_INTERLEAVE_HIGH_EXPR, VEC_INTERLEAVE_LOW_EXPR): Remove documentation. * optabs.c (optab_for_tree_code): Don't handle VEC_INTERLEAVE_HIGH_EXPR and VEC_INTERLEAVE_LOW_EXPR. (expand_binop, init_optabs): Remove vec_interleave_high_optab and vec_interleave_low_optab. * genopinit.c (optabs): Likewise. * optabs.h (OTI_vec_interleave_high, OTI_vec_interleave_low): Remove. (vec_interleave_high_optab, vec_interleave_low_optab): Remove. * doc/md.texi (vec_interleave_high, vec_interleave_low): Remove documentation. * tree-vect-stmts.c (gen_perm_mask): Renamed to... (vect_gen_perm_mask): ... this. No longer static. (perm_mask_for_reverse, vectorizable_load): Adjust callers. * tree-vectorizer.h (vect_gen_perm_mask): New prototype. * tree-vect-data-refs.c (vect_strided_store_supported): Don't try VEC_INTERLEAVE_*_EXPR, use can_vec_perm_p instead of can_vec_perm_for_code_p. (vect_permute_store_chain): Generate VEC_PERM_EXPR with interleaving masks instead of VEC_INTERLEAVE_HIGH_EXPR and VEC_INTERLEAVE_LOW_EXPR. * config/i386/i386.c (expand_vec_perm_interleave2): If expand_vec_perm_interleave3 would handle it, return false. (expand_vec_perm_broadcast_1): Don't use vec_interleave_*_optab. From-SVN: r182663
2011-12-15tree-vectorizer.h (struct _stmt_vec_info): Remove pattern_def_stmt field, ↵Jakub Jelinek1-4/+4
add pattern_def_seq. * tree-vectorizer.h (struct _stmt_vec_info): Remove pattern_def_stmt field, add pattern_def_seq. (STMT_VINFO_PATTERN_DEF_STMT): Remove. (STMT_VINFO_PATTERN_DEF_SEQ): Define. (NUM_PATTERNS): Bump to 10. * tree-vect-loop.c (vect_determine_vectorization_factor, vect_transform_loop): Adjust for pattern def changing from a single gimple stmt to gimple_seq. * tree-vect-stmts.c (vect_analyze_stmt, new_stmt_vec_info, free_stmt_vec_info): Likewise. * tree-vect-patterns.c (vect_recog_over_widening_pattern, vect_recog_vector_vector_shift_pattern, vect_recog_mixed_size_cond_pattern, adjust_bool_pattern_cast, adjust_bool_pattern, vect_mark_pattern_stmts): Likewise. (vect_recog_sdivmod_pow2_pattern): New function. (vect_vect_recog_func_ptrs): Add it. * config/i386/sse.md (vcond<V_256:mode><VI_256:mode>, vcond<V_128:mode><VI124_128:mode>, vcond<VI8F_128:mode>v2di): Use general_operand instead of nonimmediate_operand for operand 5 and no predicate for operands 1 and 2. * config/i386/i386.c (ix86_expand_int_vcond): Optimize x < 0 ? -1 : 0 and x < 0 ? 1 : 0 into vector arithmetic resp. logical shift. * gcc.dg/vect/vect-sdivmod-1.c: New test. From-SVN: r182388
2011-12-13Revert two unintended commitsRichard Henderson1-1/+0
From-SVN: r182301
2011-12-13Delete VEC_INTERLEAVE_*_EXPR.Richard Henderson1-0/+1
* tree.def (VEC_INTERLEAVE_HIGH_EXPR, VEC_INTERLEAVE_LOW_EXPR): Remove. * gimple-pretty-print.c (dump_binary_rhs): Don't handle VEC_INTERLEAVE_HIGH_EXPR and VEC_INTERLEAVE_LOW_EXPR. * expr.c (expand_expr_real_2): Likewise. * tree-cfg.c (verify_gimple_assign_binary): Likewise. * cfgexpand.c (expand_debug_expr): Likewise. * tree-inline.c (estimate_operator_cost): Likewise. * tree-pretty-print.c (dump_generic_node): Likewise. * tree-vect-generic.c (expand_vector_operations_1): Likewise. * fold-const.c (fold_binary_loc): Likewise. * doc/generic.texi (VEC_INTERLEAVE_HIGH_EXPR, VEC_INTERLEAVE_LOW_EXPR): Remove documentation. * optabs.c (optab_for_tree_code): Don't handle VEC_INTERLEAVE_HIGH_EXPR and VEC_INTERLEAVE_LOW_EXPR. (expand_binop, init_optabs): Remove vec_interleave_high_optab and vec_interleave_low_optab. * genopinit.c (optabs): Likewise. * optabs.h (OTI_vec_interleave_high, OTI_vec_interleave_low): Remove. (vec_interleave_high_optab, vec_interleave_low_optab): Remove. * doc/md.texi (vec_interleave_high, vec_interleave_low): Remove documentation. * tree-vect-stmts.c (gen_perm_mask): Renamed to... (vect_gen_perm_mask): ... this. No longer static. (perm_mask_for_reverse, vectorizable_load): Adjust callers. * tree-vectorizer.h (vect_gen_perm_mask): New prototype. * tree-vect-data-refs.c (vect_strided_store_supported): Don't try VEC_INTERLEAVE_*_EXPR, use can_vec_perm_p instead of can_vec_perm_for_code_p. (vect_permute_store_chain): Generate VEC_PERM_EXPR with interleaving masks instead of VEC_INTERLEAVE_HIGH_EXPR and VEC_INTERLEAVE_LOW_EXPR. * config/i386/i386.c (expand_vec_perm_interleave2): If expand_vec_perm_interleave3 would handle it, return false. (expand_vec_perm_broadcast_1): Don't use vec_interleave_*_optab. From-SVN: r182298
2011-11-07re PR tree-optimization/50789 (Gather vectorization)Jakub Jelinek1-0/+6
PR tree-optimization/50789 * tree-vect-stmts.c (process_use): Add force argument, avoid exist_non_indexing_operands_for_use_p check if true. (vect_mark_stmts_to_be_vectorized): Adjust callers. Handle STMT_VINFO_GATHER_P. (gen_perm_mask): New function. (perm_mask_for_reverse): Use it. (reverse_vec_element): Rename to... (permute_vec_elements): ... this. Add Y and MASK_VEC arguments, generalize for any permutations. (vectorizable_load): Adjust caller. Handle STMT_VINFO_GATHER_P. * target.def (TARGET_VECTORIZE_BUILTIN_GATHER): New hook. * doc/tm.texi.in (TARGET_VECTORIZE_BUILTIN_GATHER): Document it. * doc/tm.texi: Regenerate. * tree-data-ref.c (initialize_data_dependence_relation, compute_self_dependence): No longer static. * tree-data-ref.h (initialize_data_dependence_relation, compute_self_dependence): New prototypes. * tree-vect-data-refs.c (vect_check_gather): New function. (vect_analyze_data_refs): Detect possible gather load data refs. * tree-vectorizer.h (struct _stmt_vec_info): Add gather_p field. (STMT_VINFO_GATHER_P): Define. (vect_check_gather): New prototype. * config/i386/i386-builtin-types.def: Add types for alternate gather builtins. * config/i386/sse.md (AVXMODE48P_DI): Remove. (VEC_GATHER_MODE): Rename mode_attr to... (VEC_GATHER_IDXSI): ... this. (VEC_GATHER_IDXDI, VEC_GATHER_SRCDI): New mode_attrs. (avx2_gathersi<mode>, *avx2_gathersi<mode>): Use <VEC_GATHER_IDXSI> instead of <VEC_GATHER_MODE>. (avx2_gatherdi<mode>): Use <VEC_GATHER_IDXDI> instead of <<AVXMODE48P_DI> and <VEC_GATHER_SRCDI> instead of VEC_GATHER_MODE on src and mask operands. (*avx2_gatherdi<mode>): Likewise. Use VEC_GATHER_MODE iterator instead of AVXMODE48P_DI. (avx2_gatherdi<mode>256, *avx2_gatherdi<mode>256): Removed. * config/i386/i386.c (enum ix86_builtins): Add IX86_BUILTIN_GATHERALTSIV4DF, IX86_BUILTIN_GATHERALTDIV8SF, IX86_BUILTIN_GATHERALTSIV4DI and IX86_BUILTIN_GATHERALTDIV8SI. (ix86_init_mmx_sse_builtins): Create those builtins. (ix86_expand_builtin): Handle those builtins and adjust expansions of other gather builtins. (ix86_vectorize_builtin_gather): New function. (TARGET_VECTORIZE_BUILTIN_GATHER): Define. * gcc.target/i386/avx2-gather-1.c: New test. * gcc.target/i386/avx2-gather-2.c: New test. * gcc.target/i386/avx2-gather-3.c: New test. * gcc.target/i386/avx2-gather-4.c: New test. From-SVN: r181089
2011-11-06tree-vectorizer.h (vectorizable_condition): Add argument.Ira Rosen1-1/+1
* tree-vectorizer.h (vectorizable_condition): Add argument. * tree-vect-loop.c (vectorizable_reduction): Fail for condition in SLP. Update calls to vectorizable_condition. * tree-vect-stmts.c (vect_is_simple_cond): Add basic block info to the arguments. Pass it to vect_is_simple_use_1. (vectorizable_condition): Add slp_node to the arguments. Support vectorization of basic blocks. Fail for reduction in SLP. Update calls to vect_is_simple_cond and vect_is_simple_use. Support SLP: call vect_get_slp_defs to get vector operands. (vect_analyze_stmt): Update calls to vectorizable_condition. (vect_transform_stmt): Likewise. * tree-vect-slp.c (vect_create_new_slp_node): Handle COND_EXPR. (vect_get_and_check_slp_defs): Handle COND_EXPR. Allow pattern def stmts. (vect_build_slp_tree): Handle COND_EXPR. (vect_analyze_slp_instance): Push pattern statements to root node. (vect_get_constant_vectors): Fix comments. Handle COND_EXPR. From-SVN: r181026
2011-11-03re PR tree-optimization/50912 (gimple assertion failure at gimple.h:1940 ↵Ira Rosen1-10/+31
with -msse2) PR tree-optimization/50912 * tree-vectorizer.h (slp_void_p): New. (struct _slp_tree): Replace left and right with children. Update documentation. (struct _slp_oprnd_info): New. (vect_get_vec_defs): Declare. (vect_get_slp_defs): Update arguments. * tree-vect-loop.c (vect_create_epilog_for_reduction): Call vect_get_vec_defs instead of vect_get_slp_defs. (vectorizable_reduction): Likewise. * tree-vect-stmts.c (vect_get_vec_defs): Remove static, add argument. Update call to vect_get_slp_defs. (vectorizable_conversion): Update call to vect_get_vec_defs. (vectorizable_assignment, vectorizable_shift, vectorizable_operation): Likewise. (vectorizable_type_demotion): Call vect_get_vec_defs instead of vect_get_slp_defs. (vectorizable_type_promotion, vectorizable_store): Likewise. (vect_analyze_stmt): Fix typo. * tree-vect-slp.c (vect_free_slp_tree): Update SLP tree traversal. (vect_print_slp_tree, vect_mark_slp_stmts, vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts, vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations, vect_schedule_slp_instance): Likewise. (vect_create_new_slp_node): New. (vect_create_oprnd_info, vect_free_oprnd_info): Likewise. (vect_get_and_check_slp_defs): Pass information about defs using oprnds_info, allow any number of operands. (vect_build_slp_tree): Likewise. Update calls to vect_get_and_check_slp_defs. Fix comments. (vect_analyze_slp_instance): Move node creation to vect_create_new_slp_node. (vect_get_slp_defs): Allow any number of operands. From-SVN: r180819
2011-10-30tree-vectorizer.h (NUM_PATTERNS): Bump to 9.Jakub Jelinek1-1/+1
* tree-vectorizer.h (NUM_PATTERNS): Bump to 9. * tree-vect-patterns.c (vect_recog_vector_vector_shift_pattern): New function. (vect_vect_recog_func_ptrs): Add it. From-SVN: r180677
2011-10-23re PR tree-optimization/50819 (missed SLP vectorization)Ira Rosen1-1/+1
PR tree-optimization/50819 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Remove the last argument. * tree-vect-loop.c (vect_analyze_loop_2): Update call to vect_analyze_data_ref_dependences. * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Remove the last argument. Check load-after-store dependence for unknown dependencies in basic blocks. (vect_analyze_data_ref_dependences): Update call to vect_analyze_data_ref_dependences. * tree-vect-patterns.c (vect_recog_widen_shift_pattern): Fix typo. * tree-vect-slp.c (vect_bb_vectorizable_with_dependencies): Remove. (vect_slp_analyze_bb_1): Update call to vect_analyze_data_ref_dependences. Don't call vect_bb_vectorizable_with_dependencies. From-SVN: r180334
2011-10-18md.texi (vec_widen_ushiftl_hi, [...]): Document.Ira Rosen1-1/+1
* doc/md.texi (vec_widen_ushiftl_hi, vec_widen_ushiftl_lo, vec_widen_sshiftl_hi, vec_widen_sshiftl_lo): Document. * tree-pretty-print.c (dump_generic_node): Handle WIDEN_LSHIFT_EXPR, VEC_WIDEN_LSHIFT_HI_EXPR and VEC_WIDEN_LSHIFT_LO_EXPR. (op_code_prio): Likewise. (op_symbol_code): Handle WIDEN_LSHIFT_EXPR. * optabs.c (optab_for_tree_code): Handle VEC_WIDEN_LSHIFT_HI_EXPR and VEC_WIDEN_LSHIFT_LO_EXPR. (init-optabs): Initialize optab codes for vec_widen_u/sshiftl_hi/lo. * optabs.h (enum optab_index): Add OTI_vec_widen_u/sshiftl_hi/lo. * genopinit.c (optabs): Initialize the new optabs. * expr.c (expand_expr_real_2): Handle VEC_WIDEN_LSHIFT_HI_EXPR and VEC_WIDEN_LSHIFT_LO_EXPR. * gimple-pretty-print.c (dump_binary_rhs): Likewise. * tree-vectorizer.h (NUM_PATTERNS): Increase to 8. * tree.def (WIDEN_LSHIFT_EXPR, VEC_WIDEN_LSHIFT_HI_EXPR, VEC_WIDEN_LSHIFT_LO_EXPR): New. * cfgexpand.c (expand_debug_expr): Handle new tree codes. * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add vect_recog_widen_shift_pattern. (vect_handle_widen_mult_by_const): Rename... (vect_handle_widen_op_by_const): ...to this. Handle shifts. Add a new argument, update documentation. (vect_recog_widen_mult_pattern): Assume that only second operand can be constant. Update call to vect_handle_widen_op_by_const. (vect_recog_over_widening_pattern): Fix typo. (vect_recog_widen_shift_pattern): New. * tree-vect-stmts.c (vectorizable_type_promotion): Handle widening shifts. (supportable_widening_operation): Likewise. * tree-inline.c (estimate_operator_cost): Handle new tree codes. * tree-vect-generic.c (expand_vector_operations_1): Likewise. * tree-cfg.c (verify_gimple_assign_binary): Likewise. * config/arm/neon.md (neon_vec_<US>shiftl_<mode>): New. (vec_widen_<US>shiftl_lo_<mode>, neon_vec_<US>shiftl_hi_<mode>, vec_widen_<US>shiftl_hi_<mode>, neon_vec_<US>shift_left_<mode>): Likewise. * config/arm/predicates.md (const_neon_scalar_shift_amount_operand): New. * config/arm/iterators.md (V_innermode): New. * tree-vect-slp.c (vect_build_slp_tree): Require same shift operand for widening shift. From-SVN: r180128
2011-10-16re PR tree-optimization/50596 (Problems in vectorization of condition ↵Jakub Jelinek1-1/+1
expression) PR tree-optimization/50596 * tree-vectorizer.h (NUM_PATTERNS): Increase to 7. * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add vect_recog_bool_pattern. (check_bool_pattern, adjust_bool_pattern_cast, adjust_bool_pattern, vect_recog_bool_pattern): New functions. * gcc.dg/vect/vect-cond-9.c: New test. From-SVN: r180057
2011-10-07re PR middle-end/50650 (ICE in vect_is_simple_use_1, at tree-vect-stmts.c:5689)Jakub Jelinek1-1/+0
PR tree-optimization/50650 * tree-vect-patterns.c (vect_recog_mixed_size_cond_pattern): Don't call vect_is_simple_cond here, instead fail if cond_expr isn't COMPARISON_CLASS_P or if get_vectype_for_scalar_type returns NULL for cond_expr's first operand. * tree-vect-stmts.c (vect_is_simple_cond): Static again. * tree-vectorizer.h (vect_is_simple_cond): Remove prototype. * gcc.c-torture/compile/pr50650.c: New test. From-SVN: r179653
2011-10-06re PR tree-optimization/50596 (Problems in vectorization of condition ↵Jakub Jelinek1-2/+3
expression) PR tree-optimization/50596 * tree-vectorizer.h (vect_is_simple_cond): New prototype. (NUM_PATTERNS): Change to 6. * tree-vect-patterns.c (vect_recog_mixed_size_cond_pattern): New function. (vect_vect_recog_func_ptrs): Add vect_recog_mixed_size_cond_pattern. (vect_mark_pattern_stmts): Don't create stmt_vinfo for def_stmt if it already has one, and don't set STMT_VINFO_VECTYPE in it if it is already set. * tree-vect-stmts.c (vect_mark_stmts_to_be_vectorized): Handle COND_EXPR in pattern stmts. (vect_is_simple_cond): No longer static. * lib/target-supports.exp (check_effective_target_vect_cond_mixed): New. * gcc.dg/vect/vect-cond-8.c: New test. From-SVN: r179626
2011-08-04tree-vectorizer.h (struct _stmt_vec_info): Add new field for pattern def ↵Ira Rosen1-1/+6
statement, and its access macro. * tree-vectorizer.h (struct _stmt_vec_info): Add new field for pattern def statement, and its access macro. (NUM_PATTERNS): Set to 5. * tree-vect-loop.c (vect_determine_vectorization_factor): Handle pattern def statement. (vect_transform_loop): Likewise. * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add new function vect_recog_over_widening_pattern (). (vect_operation_fits_smaller_type): New function. (vect_recog_over_widening_pattern, vect_mark_pattern_stmts): Likewise. (vect_pattern_recog_1): Move the code that marks pattern statements to vect_mark_pattern_stmts (), and call it. Update documentation. * tree-vect-stmts.c (vect_supportable_shift): New function. (vect_analyze_stmt): Handle pattern def statement. (new_stmt_vec_info): Initialize pattern def statement. From-SVN: r177409
2011-06-16tree-vectorizer.h (vect_recog_func_ptr): Change the first argument to be a ↵Ira Rosen1-1/+1
VEC of statements. * tree-vectorizer.h (vect_recog_func_ptr): Change the first argument to be a VEC of statements. * tree-vect-loop.c (vect_determine_vectorization_factor): Remove the assert that pattern statements have to have their vector type set. * tree-vect-patterns.c (vect_recog_widen_sum_pattern): Change the first argument to be a VEC of statements. Update documentation. (vect_recog_dot_prod_pattern, vect_recog_pow_pattern): Likewise. (vect_handle_widen_mult_by_const): New function. (vect_recog_widen_mult_pattern): Change the first argument to be a VEC of statements. Update documentation. Check that the constant is INTEGER_CST. Support multiplication by a constant that fits an intermediate type - call vect_handle_widen_mult_by_const. (vect_pattern_recog_1): Update vect_recog_func_ptr and its call. Handle additional pattern statements if necessary. From-SVN: r175102
2011-06-07tree-vectorizer.h (vect_recog_func_ptr): Make last argument to be a pointer.Ira Rosen1-1/+1
* tree-vectorizer.h (vect_recog_func_ptr): Make last argument to be a pointer. * tree-vect-patterns.c (vect_recog_widen_sum_pattern, vect_recog_widen_mult_pattern, vect_recog_dot_prod_pattern, vect_recog_pow_pattern): Likewise. (vect_pattern_recog_1): Remove declaration. (widened_name_p): Remove declaration. Add new argument to specify whether to check that both types are either signed or unsigned. (vect_recog_widen_mult_pattern): Update documentation. Handle unsigned patterns and multiplication by constants. (vect_pattern_recog_1): Update vect_recog_func references. Use statement information from the statement returned from pattern detection functions. (vect_pattern_recog): Update vect_recog_func reference. * tree-vect-stmts.c (vectorizable_type_promotion): For widening multiplication by a constant use the type of the other operand. From-SVN: r174735
2011-05-26re PR tree-optimization/49038 (-ftree-vectorise introduces reads past end of ↵Ira Rosen1-0/+6
array) PR tree-optimization/49038 * tree-vect-loop-manip.c (vect_generate_tmps_on_preheader): Ensure at least one epilogue iteration if required by data accesses with gaps. * tree-vectorizer.h (struct _loop_vec_info): Add new field to mark loops that require peeling for gaps. * tree-vect-loop.c (new_loop_vec_info): Initialize new field. (vect_get_known_peeling_cost): Take peeling for gaps into account. (vect_transform_loop): Generate epilogue if required by data access with gaps. * tree-vect-data-refs.c (vect_analyze_group_access): Mark the loop as requiring an epilogue if there are gaps in the end of the strided group. From-SVN: r174265
2011-05-18re PR tree-optimization/41881 (Complete unrolling (inner) versus ↵Ira Rosen1-0/+5
vectorization of reduction) PR tree-optimization/41881 * tree-vectorizer.h (struct _loop_vec_info): Add new field reduction_chains along with a macro for its access. * tree-vect-loop.c (new_loop_vec_info): Initialize reduction chains. (destroy_loop_vec_info): Free reduction chains. (vect_analyze_loop_2): Return false if vect_analyze_slp() returns false. (vect_is_slp_reduction): New function. (vect_is_simple_reduction_1): Call vect_is_slp_reduction. (vect_create_epilog_for_reduction): Support SLP reduction chains. * tree-vect-slp.c (vect_get_and_check_slp_defs): Allow different definition types for reduction chains. (vect_supported_load_permutation_p): Don't allow permutations for reduction chains. (vect_analyze_slp_instance): Support reduction chains. (vect_analyze_slp): Try to build SLP instance from reduction chains. (vect_get_constant_vectors): Handle reduction chains. (vect_schedule_slp_instance): Mark the first statement of the reduction chain as reduction. From-SVN: r173856
2011-05-18tree-vect-loop-manip.c (vect_create_cond_for_alias_checks): Use new names ↵Ira Rosen1-24/+24
for group elements access. * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks): Use new names for group elements access. * tree-vectorizer.h (struct _stmt_vec_info): Use interleaving info for reduction chains as well. Remove data reference and interleaving related words from the fields names. * tree-vect-loop.c (vect_transform_loop): Use new names for group elements access. * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain, vect_insert_into_interleaving_chain, vect_update_interleaving_chain, vect_update_interleaving_chain, vect_same_range_drs, vect_analyze_data_ref_dependence, vect_update_misalignment_for_peel, vect_verify_datarefs_alignment, vector_alignment_reachable_p, vect_peeling_hash_get_lowest_cost, vect_enhance_data_refs_alignment, vect_analyze_group_access, vect_analyze_data_ref_access, vect_create_data_ref_ptr, vect_transform_strided_load, vect_record_strided_load_vectors): Likewise. * tree-vect-stmts.c (vect_model_simple_cost, vect_model_store_cost, vect_model_load_cost, vectorizable_store, vectorizable_load, vect_remove_stores, new_stmt_vec_info): Likewise. * tree-vect-slp.c (vect_build_slp_tree, vect_supported_slp_permutation_p, vect_analyze_slp_instance): Likewise. From-SVN: r173855
2011-05-03md.texi (vec_load_lanes, [...]): Document.Richard Sandiford1-3/+6
gcc/ * doc/md.texi (vec_load_lanes, vec_store_lanes): Document. * optabs.h (COI_vec_load_lanes, COI_vec_store_lanes): New convert_optab_index values. (vec_load_lanes_optab, vec_store_lanes_optab): New convert optabs. * genopinit.c (optabs): Initialize the new optabs. * internal-fn.def (LOAD_LANES, STORE_LANES): New internal functions. * internal-fn.c (get_multi_vector_move, expand_LOAD_LANES) (expand_STORE_LANES): New functions. * tree.h (build_array_type_nelts): Declare. * tree.c (build_array_type_nelts): New function. * tree-vectorizer.h (vect_model_store_cost): Add a bool argument. (vect_model_load_cost): Likewise. (vect_store_lanes_supported, vect_load_lanes_supported) (vect_record_strided_load_vectors): Declare. * tree-vect-data-refs.c (vect_lanes_optab_supported_p) (vect_store_lanes_supported, vect_load_lanes_supported): New functions. (vect_transform_strided_load): Split out statement recording into... (vect_record_strided_load_vectors): ...this new function. * tree-vect-stmts.c (create_vector_array, read_vector_array) (write_vector_array, create_array_ref): New functions. (vect_model_store_cost): Add store_lanes_p argument. (vect_model_load_cost): Add load_lanes_p argument. (vectorizable_store): Try to use store-lanes functions for interleaved stores. (vectorizable_load): Likewise load-lanes and loads. * tree-vect-slp.c (vect_get_and_check_slp_defs): Update call to vect_model_store_cost. (vect_build_slp_tree): Likewise vect_model_load_cost. From-SVN: r173291
2011-04-28re PR tree-optimization/48765 (ICE in vect_transform_stmt)Ira Rosen1-1/+1
PR tree-optimization/48765 * tree-vectorizer.h (vect_make_slp_decision): Return bool. * tree-vect-loop.c (vect_analyze_loop_operations): Add new argument to indicate if loop aware SLP is being used. Scan the statements and update the vectorization factor according to the type of vectorization before statement analysis. (vect_analyze_loop_2): Get a return value from vect_make_slp_decision, pass it to vect_analyze_loop_operations. (vectorizable_reduction): Set number of copies to 1 in case of pure SLP statement. * tree-vect-stmts.c (vectorizable_conversion, vectorizable_assignment, vectorizable_shift, vectorizable_operation, vectorizable_type_demotion, vectorizable_type_promotion, vectorizable_store, vectorizable_load): Likewise. (vectorizable_condition): Move the check that it is not SLP vectorization before the number of copies check. * tree-vect-slp.c (vect_make_slp_decision): Return TRUE if decided to vectorize the loop using SLP. From-SVN: r173132
2011-04-20Revert last patch.Richard Sandiford1-6/+3
From-SVN: r172763
2011-04-20md.texi (vec_load_lanes, [...]): Document.Richard Sandiford1-3/+6
gcc/ * doc/md.texi (vec_load_lanes, vec_store_lanes): Document. * optabs.h (COI_vec_load_lanes, COI_vec_store_lanes): New convert_optab_index values. (vec_load_lanes_optab, vec_store_lanes_optab): New convert optabs. * genopinit.c (optabs): Initialize the new optabs. * internal-fn.def (LOAD_LANES, STORE_LANES): New internal functions. * internal-fn.c (get_multi_vector_move, expand_LOAD_LANES) (expand_STORE_LANES): New functions. * tree.h (build_array_type_nelts): Declare. * tree.c (build_array_type_nelts): New function. * tree-vectorizer.h (vect_model_store_cost): Add a bool argument. (vect_model_load_cost): Likewise. (vect_store_lanes_supported, vect_load_lanes_supported) (vect_record_strided_load_vectors): Declare. * tree-vect-data-refs.c (vect_lanes_optab_supported_p) (vect_store_lanes_supported, vect_load_lanes_supported): New functions. (vect_transform_strided_load): Split out statement recording into... (vect_record_strided_load_vectors): ...this new function. * tree-vect-stmts.c (create_vector_array, read_vector_array) (write_vector_array, create_array_ref): New functions. (vect_model_store_cost): Add store_lanes_p argument. (vect_model_load_cost): Add load_lanes_p argument. (vectorizable_store): Try to use store-lanes functions for interleaved stores. (vectorizable_load): Likewise load-lanes and loads. * tree-vect-slp.c (vect_get_and_check_slp_defs) (vect_build_slp_tree): From-SVN: r172760
2011-04-14tree-vectorizer.h (vect_strided_store_supported): Add a HOST_WIDE_INT argument.Richard Sandiford1-6/+4
gcc/ * tree-vectorizer.h (vect_strided_store_supported): Add a HOST_WIDE_INT argument. (vect_strided_load_supported): Likewise. (vect_permute_store_chain): Return void. (vect_transform_strided_load): Likewise. (vect_permute_load_chain): Delete. * tree-vect-data-refs.c (vect_strided_store_supported): Take a count argument. Check that the count is a power of two. (vect_strided_load_supported): Likewise. (vect_permute_store_chain): Return void. Update after above changes. Assert that the access is supported. (vect_permute_load_chain): Likewise. (vect_transform_strided_load): Return void. * tree-vect-stmts.c (vectorizable_store): Update calls after above interface changes. (vectorizable_load): Likewise. (vect_analyze_stmt): Don't check for strided powers of two here. From-SVN: r172428
generated by cgit v1.1 at 2025-04-13 15:59:25 +0000