aboutsummaryrefslogtreecommitdiff
path: root/gcc/testsuite/gcc.dg/vect
AgeCommit message (Collapse)AuthorFilesLines
2010-04-26tree-vectorizer.h (struct _stmt_vec_info): Add new field to determine if the ↵Ira Rosen1-0/+54
statement is vectorizable... * tree-vectorizer.h (struct _stmt_vec_info): Add new field to determine if the statement is vectorizable, and a macro to access it. * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Skip statements that can't be vectorized. If the analysis fails, mark the statement as unvectorizable if vectorizing basic block. (vect_compute_data_refs_alignment): Likewise. (vect_verify_datarefs_alignment): Skip statements marked as unvectorizable. Add print. (vect_analyze_group_access): Skip statements that can't be vectorized. If the analysis fails, mark the statement as unvectorizable if vectorizing basic block. (vect_analyze_data_ref_accesses, vect_analyze_data_refs): Likewise. * tree-vect-stmts.c (vectorizable_store): Fix the number of generated stmts for SLP. (new_stmt_vec_info): Initialize the new field. * tree-vect-slp.c (vect_build_slp_tree): Fail to vectorize statements marked as unvectorizable. From-SVN: r158719
2010-04-22re PR tree-optimization/43842 (ice in vect_create_epilog_for_reduction)Ira Rosen1-0/+55
PR tree-optimization/43842 * tree-vect-loop.c (vect_create_epilog_for_reduction): Handle loop unrolling in update of exit phis. Fix comment. * tree-vect-slp.c (vect_analyze_slp): Check that there are at least two reduction statements in the loop before starting SLP analysis. From-SVN: r158650
2010-04-19re PR tree-optimization/37027 (SLP loop vectorization missing support for ↵Ira Rosen7-0/+351
reductions) PR tree-optimization/37027 * tree-vectorizer.h (struct _loop_vec_info): Add new field reductions and macro to access it. (vectorizable_reduction): Add argument. (vect_get_slp_defs): Likewise. * tree-vect-loop.c (vect_analyze_scalar_cycles_1): Collect reduction statements for possible use in SLP. (new_loop_vec_info): Initialize LOOP_VINFO_REDUCTIONS. (destroy_loop_vec_info): Free LOOP_VINFO_REDUCTIONS. (vect_create_epilog_for_reduction): Handle SLP. Modify documentation, add new argument. (vectorizable_reduction): Likewise. * tree-vect-stmts.c (vect_get_vec_defs): Update call to vect_get_slp_defs. (vectorizable_type_demotion, vectorizable_type_promotion, vectorizable_store): Likewise. (vect_analyze_stmt): Update call to vectorizable_reduction. (vect_transform_stmt): Likewise. * tree-vect-slp.c (vect_get_and_check_slp_defs): Handle reduction. (vect_build_slp_tree): Fix indentation. Check that there are no loads from different interleaving chains in same node. (vect_slp_rearrange_stmts): New function. (vect_supported_load_permutation_p): Allow load permutations for reductions. Call vect_slp_rearrange_stmts() to rearrange statements inside SLP nodes if necessary. (vect_analyze_slp_instance): Handle reductions. (vect_analyze_slp): Try to build SLP instances originating from groups of reductions. (vect_detect_hybrid_slp_stmts): Skip reduction statements. (vect_get_constant_vectors): Create initial vectors for reductions according to reduction code. Add new argument. (vect_get_slp_defs): Add new argument, pass it to vect_get_constant_vectors. (vect_schedule_slp_instance): Remove SLP tree root statements. From-SVN: r158506
2010-04-13tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Only add RW ↵Richard Guenther1-1/+1
dependence for dependence distance zero. 2010-04-13 Richard Guenther <rguenther@suse.de> * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Only add RW dependence for dependence distance zero. Adjust maximal vectorization factor according to dependences. Move alignment handling ... (vect_find_same_alignment_drs): ... here. New function. (vect_analyze_data_ref_dependences): Adjust. (vect_analyze_data_refs_alignment): Call vect_find_same_alignment_drs. (vect_analyze_data_refs): Adjust minimal vectorization factor according to data references. * tree-vect-loop.c (vect_analyze_loop): Analyze data-ref dependences before determining the vectorization factor. Analyze alignment after determining the vectorization factor. * tree-vect-slp.c ((vect_slp_analyze_bb): Analyze data-ref dependences before alignment. * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust prototype. (vect_analyze_data_refs): Likewise. (MAX_VECTORIZATION_FACTOR): New define. * gcc.dg/vect/no-vfa-vect-depend-1.c: Adjust. From-SVN: r158259
2010-04-06re PR tree-optimization/32824 (Missed reduction vectorizer after store to ↵Changpeng Fang1-0/+15
global is LIM'd) Fix PR32824. 2010-04-06 Changpeng Fang <changpeng.fang@amd.com> PR middle-end/32824 * passes.c (init_optimization_passes): Move pass_lim before pass_copy_prop and pass_dce_loop. * gcc.dg/vect/pr32824: New. From-SVN: r158018
2010-04-04Don't use uint8_t.Sebastian Pop1-4/+4
2010-04-04 Sebastian Pop <sebastian.pop@amd.com> * gcc.dg/vect/pr43430-1.c: Don't use uint8_t. From-SVN: r157959
2010-03-30Replace type != type comparisons with types_compatible_p.Sebastian Pop1-0/+39
2010-03-30 Sebastian Pop <sebastian.pop@amd.com> PR middle-end/43430 * tree-vect-slp.c (vect_get_and_check_slp_defs): Replace type pointer comparisons with types_compatible_p. * tree-vect-stmts.c (vectorizable_call): Same. (vectorizable_condition): Same. * gcc.dg/vect/pr43430-1.c: New. From-SVN: r157833
2010-02-16re PR tree-optimization/43074 (ICE in vectorizable_reduction, at ↵Ira Rosen1-0/+16
tree-vect-loop.c:3491) PR tree-optimization/43074 * tree-vectorizer.h (VECTORIZABLE_CYCLE_DEF): New. * tree-vect-loop.c (vect_analyze_loop_operations): Add vectorizable cycles in hybrid SLP check. * tree-vect-slp.c (vect_detect_hybrid_slp_stmts): Likewise. From-SVN: r156800
2010-01-14re PR tree-optimization/42709 (error: type mismatch in pointer plus expression)Ira Rosen1-0/+28
PR tree-optimization/42709 * tree-vect-slp.c (vect_get_constant_vectors): Use constant's type as scalar type in creation of constant vector operand. From-SVN: r155882
2010-01-09re PR debug/42604 (VTA ICE in the vectorizer)Alexandre Oliva2-0/+45
gcc/ChangeLog: PR debug/42604 PR debug/42395 * tree-vect-loop-manip.c (adjust_info): New type. (adjust_vec): New pointer to vector. (adjust_debug_stmts_now, adjust_vec_debug_stmts): New. (adjust_debug_stmts, adjust_phi_and_debug_stmts): New. (slpeel_update_phis_for_duplicate_loop): Use them. (slpeel_update_phi_nodes_for_guard1): Likewise. (slpeel_update_phi_nodes_for_guard2): Likewise. (slpeel_tree_peel_loop_to_edge): Likewise. (vect_update_ivs_after_vectorizer): Likewise. gcc/testsuite/ChangeLog: PR debug/42604 PR debug/42395 * gcc.dg/vect/pr42604.c: New. * gcc.dg/vect/pr42395.c: New. From-SVN: r155761
2009-12-30re PR tree-optimization/41956 (Segfault in vectorizer)Ira Rosen1-0/+13
PR tree-optimization/41956 * tree-vect-slp.c (vect_supported_load_permutation_p): Add check that the load indices differ. From-SVN: r155523
2009-11-29re PR middle-end/42193 (454.calculix in SPEC CPU 2006 failed to compile at -O3)Ira Rosen1-0/+28
PR tree-optimization/42193 * tree-vect-stmts.c (vectorizable_operation): Set vectorization factor to 1 in case of basic block SLP. (vectorizable_load): Likewise. From-SVN: r154738
2009-11-25i386-builtin-types.awk (DEF_VECTOR_TYPE): Allow an optional 3rd argument to ↵Richard Henderson1-2/+2
define the mode. * config/i386/i386-builtin-types.awk (DEF_VECTOR_TYPE): Allow an optional 3rd argument to define the mode. * config/i386/i386-builtin-types.def (UQI, UHI, USI, UDI): New. (V2UDI, V4USI, V8UHI, V16UQI): New. (V4SF_FTYPE_V4SF_V4SF_V4SI, V2UDI_FTYPE_V2UDI_V2UDI_V2UDI, V4USI_FTYPE_V4USI_V4USI_V4USI, V8UHI_FTYPE_V8UHI_V8UHI_V8UHI, V16UQI_FTYPE_V16UQI_V16UQI_V16UQI): New. * config/i386/i386-modes.def: Rearrange for double-wide AVX. * config/i386/i386-protos.h (ix86_expand_vec_extract_even_odd): New. * config/i386/i386.c (IX86_BUILTIN_VEC_PERM_*): New. (bdesc_args): Add the builtin definitions to match. (ix86_expand_builtin): Expand them. (ix86_builtin_vectorization_cost): Rename from x86_builtin_vectorization_cost. (ix86_vectorize_builtin_vec_perm, struct expand_vec_perm_d, doublesize_vector_mode, expand_vselect, expand_vselect_vconcat, expand_vec_perm_blend, expand_vec_perm_vpermil, expand_vec_perm_pshufb, expand_vec_perm_1, expand_vec_perm_pshuflw_pshufhw, expand_vec_perm_palignr, expand_vec_perm_interleave2, expand_vec_perm_pshufb2, expand_vec_perm_even_odd_1, expand_vec_perm_even_odd, ix86_expand_vec_perm_builtin_1, extract_vec_perm_cst, ix86_expand_vec_perm_builtin, ix86_vectorize_builtin_vec_perm_ok, ix86_expand_vec_extract_even_odd, TARGET_VECTORIZE_BUILTIN_VEC_PERM, TARGET_VECTORIZE_BUILTIN_VEC_PERM_OK): New. * sse.md (SSEMODE_EO): New. (vec_extract_even<mode>): Use SSEMODE_EO and ix86_expand_vec_extract_even_odd. (vec_extract_odd<mode>): Likewise. (mulv16qi3, vec_pack_trunc_v8hi, vec_pack_trunc_v4si, vec_pack_trunc_v2di): Use ix86_expand_vec_extract_even_odd. testsuite/ * gcc.dg/vect/slp-21.c: Succeed with vect_extract_even_odd too. * lib/target-supports.exp (check_effective_target_vect_extract_even_odd): Add x86. * gcc.target/i386/isa-check.h: New. * gcc.target/i386/vperm-2-2.inc, gcc.target/i386/vperm-4-1.inc, gcc.target/i386/vperm-4-2.inc, gcc.target/i386/vperm-v2df.c, gcc.target/i386/vperm-v2di.c, gcc.target/i386/vperm-v4sf-1.c, gcc.target/i386/vperm-v4sf-2.c, gcc.target/i386/vperm-v4si-1.c, gcc.target/i386/vperm-v4si-2.c, gcc.target/i386/vperm-v4si-2x.c, gcc.target/i386/vperm.pl: New files. From-SVN: r154667
2009-11-25pr36493.c: Call check_vect.Richard Henderson4-6/+35
* gcc.dg/vect/pr36493.c: Call check_vect. * gcc.dg/vect/pr37539.c: Likewise. * gcc.dg/vect/vect-nest-cycle-3.c: Call check_vect earlier. * tree-vect.h (check_vect): Use cpuid for x86. From-SVN: r154666
2009-11-18re PR debug/41926 ([VTA] internal compiler error: verify_ssa failed)Alexandre Oliva1-0/+20
gcc/ChangeLog: PR debug/41926 * tree-vect-loop.c (vect_loop_kill_debug_uses): New. (vect_transform_loop): Call it. gcc/testsuite/ChangeLog: PR debug/41926 * gcc.dg/vect/vect-debug-pr41926.c: New. From-SVN: r154281
2009-11-10vect-multitypes-5.c: XFAIL on SPARC 32-bit.Eric Botcazou4-5/+5
* gcc.dg/vect/vect-multitypes-5.c: XFAIL on SPARC 32-bit. * gcc.dg/vect/vect-multitypes-6.c: Likewise. * gcc.dg/vect/vect-multitypes-12.c: Likewise. * gcc.dg/vect/slp-multitypes-3.c: Likewise. From-SVN: r154070
2009-11-04Another fix to the misaligned store vectorizer patchRevital Eres14-17/+17
From-SVN: r153892
2009-10-27Fix PR40648 -- Fix misaligned store vectorizer patchRevital Eres40-90/+70
From-SVN: r153590
2009-09-25vect.exp: Append extra parameters as separate parameters for ↵Nathan Froyd1-1/+1
alignment-sensitive... * gcc.dg/vect/vect.exp: Append extra parameters as separate parameters for alignment-sensitive -fsection-anchors tests. From-SVN: r152182
2009-09-25Vectorizer testsuite adjustments for P7Revital Eres1-0/+4
From-SVN: r152157
2009-09-08large-size-array-2.c: Fix target requirement.Ozkan Sezer3-3/+9
2009-09-08 Ozkan Sezer <sezeroz@gmail.com> * gcc.dg/large-size-array-2.c: Fix target requirement. * gcc.dg/large-size-array-4.c: Likewise. * gcc.dg/c90-const-expr-2.c: Define ZERO as 0LL for _WIN64. * gcc.dg/c99-const-expr-2.c: Likewise. * gcc.dg/c90-const-expr-6.c: Typedef ptrdiff_t with __extension__. * gcc.dg/c90-const-expr-9.c: Typedef size_t with __extension__. * gcc.dg/c90-const-expr-10.c: Likewise. * gcc.dg/torture/pr39074.c: Use intptr_t typedef'ed as __PTRDIFF_TYPE__ instead of long. * gcc.dg/tree-ssa/pr33920.c: Likewise. * gcc.dg/vect/pr33846.c: Likewise. * gcc.dg/vect/pr33833.c: Use uintptr_t, typedef'ed as unsigned __PTRDIFF_TYPE__, instead of unsigned long. * gcc.dg/vect/O1-pr33854.c: Use __SIZE_TYPE__ instead of unsigned long. From-SVN: r151533
2009-08-18vect.exp: Add new stanza for aligned-section-anchors-* tests.Nathan Froyd2-2/+7
* gcc.dg/vect/vect.exp: Add new stanza for aligned-section-anchors-* tests. * gcc.dg/vect/section-anchors-nest-1.c: Rename to... * gcc.dg/vect/aligned-section-anchors-nest-1.c: ...this. Remove dg-options. Fix dg-final clause. From-SVN: r150880
2009-08-09re PR tree-optimization/41008 (ICE in vect_is_simple_reduction, at ↵Ira Rosen1-0/+23
tree-vect-loop.c:1708) PR tree-optimization/41008 * tree-vect-loop.c (vect_is_simple_reduction): Get operands from condition only in case it's a comparison. Adjust checks. From-SVN: r150591
2009-07-31Fix patterns in lim and dceloop testcases.Sebastian Pop1-2/+2
2009-07-30 Sebastian Pop <sebastian.pop@amd.com> * gcc.dg/tree-ssa/20050314-1.c: Fix patterns for lim and dceloop. * gcc.dg/tree-ssa/loop-32.c: Same. * gcc.dg/tree-ssa/loop-33.c: Same. * gcc.dg/tree-ssa/loop-34.c: Same. * gcc.dg/tree-ssa/loop-35.c: Same. * gcc.dg/tree-ssa/loop-7.c: Same. * gcc.dg/tree-ssa/pr23109.c: Same. * gcc.dg/tree-ssa/restrict-2.c: Same. * gcc.dg/tree-ssa/restrict-3.c: Same. * gcc.dg/tree-ssa/ssa-lim-1.c: Same. * gcc.dg/tree-ssa/ssa-lim-2.c: Same. * gcc.dg/tree-ssa/ssa-lim-3.c: Same. * gcc.dg/tree-ssa/ssa-lim-6.c: Same. * gcc.dg/tree-ssa/structopt-1.c: Same. * gcc.dg/vect/dump-tree-dceloop-pr26359.c: Same. * gfortran.dg/pr32921.f: Same. From-SVN: r150304
2009-07-29re PR middle-end/40830 (gcc.dg/vect/vect-pre-interact.c doesn't work on ↵Michael Matz1-1/+1
Linux/ia32) PR middle-end/40830 * gcc.dg/vect/vect-pre-interact.c: XFAIL for no_align targets. From-SVN: r150215
2009-07-27no-scevccp-noreassoc-outer-2.c: Extended array 'a' so that indexing no ↵Simon Baldwin1-1/+1
longer runs off array end. * gcc.dg/vect/no-scevccp-noreassoc-outer-2.c: Extended array 'a' so that indexing no longer runs off array end. From-SVN: r150109
2009-07-23re PR middle-end/40830 (gcc.dg/vect/vect-pre-interact.c doesn't work on ↵Michael Matz1-2/+1
Linux/ia32) PR middle-end/40830 * gcc.dg/vect/vect-pre-interact.c: Require vect_float, and don't set options ourself. From-SVN: r149999
2009-07-22re PR tree-optimization/35229 (Vectorizer doesn't support dependence ↵Michael Matz1-0/+17
created by predictive commoning or PRE) PR tree-optimization/35229 PR tree-optimization/39300 * tree-ssa-pre.c (includes): Include tree-scalar-evolution.h. (inhibit_phi_insertion): New function. (insert_into_preds_of_block): Call it for REFERENCEs. (init_pre): Initialize and finalize scalar evolutions. * Makefile.in (tree-ssa-pre.o): Depend on tree-scalar-evolution.h . testsuite/ * gcc.dg/vect/vect-pre-interact.c: New test. From-SVN: r149942
2009-07-21tree-vectorizer.c (increase_alignment): Handle nested arrays.Paul Brook1-0/+35
2009-07-21 Paul Brook <paul@codesourcery.com> gcc/ * tree-vectorizer.c (increase_alignment): Handle nested arrays. Terminate debug dump with newline. gcc/testsuite/ * gcc.dg/vect/section-anchors-nest-1.c: New test. * lib/target-supports.exp (check_effective_target_section_anchors): Add arm*-*-*. From-SVN: r149853
2009-07-20tree-vectorizer.h (vectorizable_condition): Add parameters.Ira Rosen6-0/+358
* tree-vectorizer.h (vectorizable_condition): Add parameters. * tree-vect-loop.c (vect_is_simple_reduction): Support COND_EXPR. (get_initial_def_for_reduction): Likewise. (vectorizable_reduction): Skip the check of first operand in case of COND_EXPR. Add check that it is outer loop vectorization if nested cycle was detected. Call vectorizable_condition() for COND_EXPR. If reduction epilogue cannot be created do not fail for nested cycles (if it is not double reduction). Assert that there is only one type in the loop in case of COND_EXPR. Call vectorizable_condition() to vectorize COND_EXPR. * tree-vect-stmts.c (vectorizable_condition): Update comment. Add parameters. Allow nested cycles if called from vectorizable_reduction(). Use reduction vector variable if provided. (vect_analyze_stmt): Call vectorizable_reduction() before vectorizable_condition(). (vect_transform_stmt): Update call to vectorizable_condition(). From-SVN: r149806
2009-07-12tree-parloops.c (loop_parallel_p): Call vect_is_simple_reduction with ↵Ira Rosen8-1/+411
additional argument. * tree-parloops.c (loop_parallel_p): Call vect_is_simple_reduction with additional argument. * tree-vectorizer.h (enum vect_def_type): Add vect_double_reduction_def. (vect_is_simple_reduction): Add argument. * tree-vect-loop.c (vect_determine_vectorization_factor): Fix indentation. (vect_analyze_scalar_cycles_1): Detect double reduction. Call vect_is_simple_reduction with additional argument. (vect_analyze_loop_operations): Handle exit phi nodes in case of double reduction. (reduction_code_for_scalar_code): Handle additional codes by returning ERROR_MARK for them. Fix comment and indentation. (vect_is_simple_reduction): Fix comment, add argument to specify double reduction. Detect double reduction. (get_initial_def_for_induction): Fix indentation. (get_initial_def_for_reduction): Fix comment and indentation. Handle double reduction. Create initial definitions that do not require adjustment if ADJUSTMENT_DEF is NULL. Handle additional cases. (vect_create_epilog_for_reduction): Fix comment, add argument to handle double reduction. Use PLUS_EXPR in case of MINUS_EXPR in epilogue result extraction. Create double reduction phi node and replace relevant uses. (vectorizable_reduction): Call vect_is_simple_reduction with additional argument. Fix indentation. Update epilogue code treatment according to the changes in reduction_code_for_scalar_code. Check for double reduction. Call vect_create_epilog_for_reduction with additional argument. * tree-vect-stmts.c (process_use): Handle double reduction, update documentation. (vect_mark_stmts_to_be_vectorized): Handle double reduction. (vect_get_vec_def_for_operand): Likewise. From-SVN: r149526
2009-07-01re PR tree-optimization/19831 (Missing DSE/malloc/free optimization)Richard Guenther2-4/+8
2009-07-01 Richard Guenther <rguenther@suse.de> PR tree-optimization/19831 * tree-ssa-dce.c (propagate_necessity): Calls to functions that only act as barriers do not make any previous stores necessary. * tree-ssa-structalias.c (handle_lhs_call): Delay making HEAP variables global, do not add a constraint from nonlocal. (find_func_aliases): Handle escapes through return statements. (compute_points_to_sets): Make escaped HEAP variables global. * gcc.dg/tree-ssa/20041122-1.c: Enable TBAA, scan FRE dump, make allocated memory escape. Un-XFAIL. * gcc.dg/vect/pr21591.c: Make allocated memory escape. * gcc.dg/vect/pr31699.c: Likewise. * gcc.dg/tree-ssa/ssa-dce-7.c: New testcase. libmudflap/ * testsuite/libmudflap.c/fail11-frag.c: Make allocated memory escape. * testsuite/libmudflap.c/fail12-frag.c: Likewise. * testsuite/libmudflap.c/fail16-frag.c: Likewise. * testsuite/libmudflap.c/fail31-frag.c: Likewise. From-SVN: r149140
2009-06-30re PR tree-optimization/40542 (vectorizes access to volatile array)Ira Rosen1-0/+14
PR tree-optimization/40542 * tree-vect-stmts.c (vect_analyze_stmt): Don't vectorize volatile types. From-SVN: r149090
2009-06-27re PR testsuite/40567 (Revision 149002 caused many failures)Paolo Bonzini1-1/+1
2009-06-27 Paolo Bonzini <bonzini@gnu.org> PR testsuite/40567 * gcc.dg/vect/vect.exp: Fix lappend syntax. From-SVN: r149006
2009-06-26re PR testsuite/40565 (Extra failures)Richard Henderson1-2/+3
PR testsuite/40565 * gcc.dg/vect/vect.exp: Add -fno-ipa-cp-clone to -O3 tests. * gcc.target/i386/local.c: Pass a non-constant to T. From-SVN: r149002
2009-06-21fre-vce-1.c: Cleanup "fre" tree dump.Uros Bizjak2-1/+4
* gcc.dg/tree-ssa/fre-vce-1.c: Cleanup "fre" tree dump. * gcc.dg/vect/vect-shift-2.c: Cleanup "vect" tree dump. * gcc.dg/vect/vect.exp (VECT_SLP_CFLAGS): Initialize with original DEFAULT_VECTFLAGS. From-SVN: r148766
2009-06-21Fix 40359Ira Rosen3-3/+4
Co-Authored-By: Revital Eres <eres@il.ibm.com> From-SVN: r148762
2009-06-21re PR testsuite/40475 (gcc.dg/vect/vect-nest-cycle-[12].c)Ira Rosen2-2/+2
PR testsuite/40475 * gcc.dg/vect/vect-nest-cycle-1.c: Fail to vectorize on targets without misalignment support. * gcc.dg/vect/vect-nest-cycle-2.c: Likewise. From-SVN: r148758
2009-06-16tree-parloops.c (loop_parallel_p): Call vect_is_simple_reduction with ↵Ira Rosen8-10/+155
additional parameter. * tree-parloops.c (loop_parallel_p): Call vect_is_simple_reduction with additional parameter. * tree-vectorizer.h (enum vect_def_type): Add new value vect_nested_cycle. (enum vect_relevant): Add comments. (vect_is_simple_reduction): Add new argument. * tree-vect-loop.c (vect_analyze_scalar_cycles_1): Add comments. Detect nested cycles. (vect_is_simple_reduction): Update documentation, add an argument to distinguish inner-loop reduction from nested cycle, detect nested cycles, fix printings and indentation, don't swap operands in case of nested cycle. (get_initial_def_for_reduction): Handle subtraction. (vect_create_epilog_for_reduction): Add new argument to specify reduction variable. (vect_finalize_reduction): Handle subtraction, fix comments. (vectorizable_reduction): Handle nested cycles. In case of nested cycle keep track of the reduction variable position. Call vect_is_simple_reduction with additional parameter. Use original statement code in reduction epilogue for nested cycle. Call vect_create_epilog_for_reduction with additional parameter. * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Assert inner-loop vectorization. (vect_recog_widen_sum_pattern): Likewise. * tree-vect-stmts.c (process_use): Distinguish between nested cycles and reductions. (vect_mark_stmts_to_be_vectorized): Likewise. (vect_get_vec_def_for_operand): Handle nested cycles. From-SVN: r148518
2009-06-08re PR testsuite/40359 (Revision 148211 caused a lot of failures in the vect ↵Revital Eres4-10/+13
test suite.) PR40359 From-SVN: r148275
2009-06-05misaligned store supportRevital Eres30-71/+89
Co-Authored-By: Leehod Baruch <leehod@il.ibm.com> From-SVN: r148211
2009-06-04vect-42.c: Modify vect_no_align case.Steve Ellcey2-2/+2
2009-06-04 Steve Ellcey <sje@cup.hp.com> * gcc.dg/vect/vect-42.c: Modify vect_no_align case. * gcc.dg/vect/no-vfa-vect-43.c: Ditto. From-SVN: r148177
2009-05-31re PR testsuite/40244 (Revision 147829 caused extra failures)Ira Rosen2-10/+6
PR testsuite/40244 * gcc.dg/vect/bb-slp-4.c: Change the number of data accesses to 2. * gcc.dg/vect/bb-slp-10.c: Change the store misalignment to 1. From-SVN: r148010
2009-05-28re PR tree-optimization/40254 (SPEC2006 403.gcc miscompares)Ira Rosen1-0/+39
PR tree-optimization/40254 * tree-data-ref.c (dr_analyze_innermost): Take POFFSET into account in analysis of basic blocks. From-SVN: r147945
2009-05-25re PR tree-optimization/36327 (SCCVN should look through struct copies)Richard Guenther5-26/+55
2009-05-25 Richard Guenther <rguenther@suse.de> PR tree-optimization/36327 * tree-ssa-alias.c (walk_non_aliased_vuses): Add second walker callback for reference translation or lookup at the point of may-defs. * tree-ssa-alias.h (walk_non_aliased_vuses): Adjust prototype. * tree-ssa-sccvn.c (get_ref_from_reference_ops): Bail out for union COMPONENT_REFs. (vn_reference_lookup_3): New callback. Lookup from memset and CONSTRUCTOR assignment, translate through struct copies. (vn_reference_lookup_pieces): Make sure to not free the passed operands array. Adjust walk_non_aliased_vuses call. (vn_reference_lookup): Adjust walk_non_aliased_vuses call, make sure we do not leak memory. * gcc.dg/tree-ssa/ssa-fre-24.c: New testcase. * gcc.dg/tree-ssa/ssa-fre-25.c: Likewise. * gcc.dg/tree-ssa/sra-2.c: Disable FRE. * gcc.dg/vect/no-vfa-vect-43.c: Adjust. * gcc.dg/vect/vect-40.c: Likewise. * gcc.dg/vect/vect-42.c: Likewise. * gcc.dg/vect/vect-46.c: Likewise. * gcc.dg/vect/vect-76.c: Likewise. From-SVN: r147851
2009-05-25re PR tree-optimization/40238 (ICE in gimple_verify_flow_info, at ↵Ira Rosen1-0/+35
tree-cfg.c:4603) PR tree-optimization/40238 * tree-vect-stmts.c (vect_init_vector): Insert initialization statements after basic block's labels. * tree-vect-slp.c (vect_slp_transform_bb): Call destroy_bb_vec_info() to free the allocated memory. From-SVN: r147844
2009-05-24passes.texi (Tree-SSA passes): Document SLP pass.Ira Rosen23-0/+1231
* doc/passes.texi (Tree-SSA passes): Document SLP pass. * tree-pass.h (pass_slp_vectorize): New pass. * params.h (SLP_MAX_INSNS_IN_BB): Define. * timevar.def (TV_TREE_SLP_VECTORIZATION): Define. * tree-vectorizer.c (timevar.h): Include. (user_vect_verbosity_level): Declare. (vect_location): Fix comment. (vect_set_verbosity_level): Update user_vect_verbosity_level instead of vect_verbosity_level. (vect_set_dump_settings): Add an argument. Ignore user defined verbosity if dump flags require higher level of verbosity. Print to stderr only for loop vectorization. (vectorize_loops): Update call to vect_set_dump_settings. (execute_vect_slp): New function. (gate_vect_slp): Likewise. (struct gimple_opt_pass pass_slp_vectorize): New. * tree-vectorizer.h (struct _bb_vec_info): Define along macros to access its members. (vec_info_for_bb): New function. (struct _stmt_vec_info): Add bb_vinfo and a macro for its access. (VECTORIZATION_ENABLED): New macro. (SLP_ENABLED, SLP_DISABLED): Likewise. (vect_is_simple_use): Add bb_vec_info argument. (new_stmt_vec_info, vect_analyze_data_ref_dependences, vect_analyze_data_refs_alignment, vect_verify_datarefs_alignment, vect_analyze_data_ref_accesses, vect_analyze_data_refs, vect_schedule_slp, vect_analyze_slp): Likewise. (vect_analyze_stmt): Add slp_tree argument. (find_bb_location): Declare. (vect_slp_analyze_bb, vect_slp_transform_bb): Likewise. * tree-vect-loop.c (new_loop_vec_info): Adjust function calls. (vect_analyze_loop_operations, vect_analyze_loop, get_initial_def_for_induction, vect_create_epilog_for_reduction, vect_finalize_reduction, vectorizable_reduction, vectorizable_live_operation, vect_transform_loop): Likewise. * tree-data-ref.c (dr_analyze_innermost): Update comment, skip evolution analysis if analyzing a basic block. (dr_analyze_indices): Likewise. (initialize_data_dependence_relation): Skip the test whether the object is invariant for basic blocks. (compute_all_dependences): Skip dependence analysis for data references in basic blocks. (find_data_references_in_stmt): Don't fail in case of invariant access in basic block. (find_data_references_in_bb): New function. (find_data_references_in_loop): Move code to find_data_references_in_bb and add a call to it. (compute_data_dependences_for_bb): New function. * tree-data-ref.h (compute_data_dependences_for_bb): Declare. * tree-vect-data-refs.c (vect_check_interleaving): Adjust to the case that STEP is 0. (vect_analyze_data_ref_dependence): Check for interleaving in case of unknown dependence in basic block and fail in case of dependence in basic block. (vect_analyze_data_ref_dependences): Add bb_vinfo argument, get data dependence instances from either loop or basic block vectorization info. (vect_compute_data_ref_alignment): Check if it is loop vectorization before calling nested_in_vect_loop_p. (vect_compute_data_refs_alignment): Add bb_vinfo argument, get data dependence instances from either loop or basic block vectorization info. (vect_verify_datarefs_alignment): Likewise. (vect_enhance_data_refs_alignment): Adjust function calls. (vect_analyze_data_refs_alignment): Likewise. (vect_analyze_group_access): Fix printing. Skip different checks if DR_STEP is 0. Keep strided stores either in loop or basic block vectorization data structure. Fix indentation. (vect_analyze_data_ref_access): Fix comments, allow zero step in basic blocks. (vect_analyze_data_ref_accesses): Add bb_vinfo argument, get data dependence instances from either loop or basic block vectorization info. (vect_analyze_data_refs): Update comment. Call compute_data_dependences_for_bb to analyze basic blocks. (vect_create_addr_base_for_vector_ref): Check for outer loop only in case of loop vectorization. In case of basic block vectorization use data-ref itself as a base. (vect_create_data_ref_ptr): In case of basic block vectorization: don't advance the pointer, add new statements before the current statement. Adjust function calls. (vect_supportable_dr_alignment): Support only aligned accesses in basic block vectorization. * common.opt (ftree-slp-vectorize): New flag. * tree-vect-patterns.c (widened_name_p): Adjust function calls. (vect_pattern_recog_1): Likewise. * tree-vect-stmts.c (process_use): Likewise. (vect_init_vector): Add new statements in the beginning of the basic block in case of basic block SLP. (vect_get_vec_def_for_operand): Adjust function calls. (vect_finish_stmt_generation): Likewise. (vectorizable_call): Add assert that it is loop vectorization, adjust function calls. (vectorizable_conversion, vectorizable_assignment): Likewise. (vectorizable_operation): In case of basic block SLP, take vectorization factor from statement's type and skip the relevance check. Adjust function calls. (vectorizable_type_demotion): Add assert that it is loop vectorization, adjust function calls. (vectorizable_type_promotion): Likewise. (vectorizable_store): Check for outer loop only in case of loop vectorization. Adjust function calls. For basic blocks, skip the relevance check and don't advance pointers. (vectorizable_load): Likewise. (vectorizable_condition): Add assert that it is loop vectorization, adjust function calls. (vect_analyze_stmt): Add argument. In case of basic block SLP, check that it is not reduction, get vector type, call only supported functions, skip loop specific parts. (vect_transform_stmt): Check for outer loop only in case of loop vectorization. (new_stmt_vec_info): Add new argument and initialize bb_vinfo. (vect_is_simple_use): Fix comment, add new argument, fix conditions for external definition. * passes.c (pass_slp_vectorize): New pass. * tree-vect-slp.c (find_bb_location): New function. (vect_get_and_check_slp_defs): Add argument, adjust function calls, check for patterns only in loops. (vect_build_slp_tree): Add argument, adjust function calls, fail in case of multiple types in basic block SLP. (vect_mark_slp_stmts_relevant): New function. (vect_supported_load_permutation_p): Fix comment. (vect_analyze_slp_instance): Add argument. In case of basic block SLP, take vectorization factor from statement's type, check that unrolling factor is 1. Adjust function call. Save SLP instance in either loop or basic block vectorization structure. Return FALSE, if SLP failed. (vect_analyze_slp): Add argument. Get strided stores groups from either loop or basic block vectorization structure. Return FALSE if basic block SLP failed. (new_bb_vec_info): New function. (destroy_bb_vec_info, vect_slp_analyze_node_operations, vect_slp_analyze_operations, vect_slp_analyze_bb): Likewise. (vect_schedule_slp): Add argument. Get SLP instances from either loop or basic block vectorization structure. Set vectorization factor to be 1 for basic block SLP. (vect_slp_transform_bb): New function. * params.def (PARAM_SLP_MAX_INSNS_IN_BB): Define. From-SVN: r147829
2009-05-15align-main-1.c (check): Mark noinline.Jan Hubicka5-0/+5
* gcc.target/i386/align-main-1.c (check): Mark noinline. * gcc.target/i386/align-main-2.c (check): Mark noinline. * gcc.dg/ipa/ipa-4.c: Disable early inlining. * gcc.dg/vect/vect-iv-10.c (main1): Mark noinline. * gcc.dg/vect/costmodel/i386/costmodel-vect-33.c (main1): Mark noinline. * gcc.dg/vect/costmodel/x86_64/costmodel-vect-33.c (main1): Mark noinline. * gcc.dg/vect/pr31699.c (foo): Mark noinline. * gcc.dg/vect/pr18400.c (main1): Mark noinline. From-SVN: r147585
2009-05-12loop-36.c: Reduce amount of iterations to 2 so unrolling still happens.Jan Hubicka1-3/+2
* gcc.dg/tree-ssa/loop-36.c: Reduce amount of iterations to 2 so unrolling still happens. * gcc.dg/ipa/ipacost-1.c: Prevent inlining * gcc.dg/ipa/ipacost-2.c: Likewise. * gcc.dg/vect/slp-3.c: Loop is no longer unrolled. * tree-inline.c (estimate_operator_cost): Add operands; when division happens by constant, it is cheap. (estimate_num_insns): Loads and stores are not having cost of 0; EH magic stuff is cheap; when computing runtime cost of switch, use log2 base of amount of its cases; builtin_expect has cost of 0; compute cost for moving return value of call. (init_inline_once): Initialize time_based flags. * tree-inline.h (eni_weights_d): Add time_based flag. From-SVN: r147436
2009-05-11pr21829.c: Simplify matching since we now optimize better.Jan Hubicka6-10/+11
* gcc.dg/tree-ssa/pr21829.c: Simplify matching since we now optimize better. * gcc.dg/Wunreachable-8.c: Bogus warnings now come out at different places. * gcc.dg/vect/vect-92.c: Increase loop iteration count to prevent unroling. * gcc.dg/vect/vect-76.c: Likewise. * gcc.dg/vect/vect-70.c: Likewise. * gcc.dg/vect/vect-66.c: Likewise. * gcc.dg/vect/no-section-anchors-vect-66.c: Likewise. * gcc.dg/vect/slp-3.c: One of loops gets now fully unrolled. * tree-ssa-loop-ivcanon.c: Include target.h (struct loop_size): new structure. (constant_after_peeling): New predicate. (tree_estimate_loop_size): New function. (estimated_unrolled_size): Rewrite for new estimates. (try_unroll_loop_completely): Use new estimates. * Makefile.in (tree-ssa-loop-ivcanon.o): Add dependenc on target.h From-SVN: r147395
generated by cgit v1.1 at 2025-03-09 22:18:28 +0000