URL
https://opencores.org/ocsvn/openrisc/openrisc/trunk
Subversion Repositories openrisc
[/] [openrisc/] [trunk/] [gnu-old/] [gcc-4.2.2/] [gcc/] [tree-vect-analyze.c] - Rev 856
Go to most recent revision | Compare with Previous | Blame | View Log
/* Analysis Utilities for Loop Vectorization. Copyright (C) 2003,2004,2005,2006, 2007 Free Software Foundation, Inc. Contributed by Dorit Naishlos <dorit@il.ibm.com> This file is part of GCC. GCC is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 3, or (at your option) any later version. GCC is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with GCC; see the file COPYING3. If not see <http://www.gnu.org/licenses/>. */ #include "config.h" #include "system.h" #include "coretypes.h" #include "tm.h" #include "ggc.h" #include "tree.h" #include "target.h" #include "basic-block.h" #include "diagnostic.h" #include "tree-flow.h" #include "tree-dump.h" #include "timevar.h" #include "cfgloop.h" #include "expr.h" #include "optabs.h" #include "params.h" #include "tree-chrec.h" #include "tree-data-ref.h" #include "tree-scalar-evolution.h" #include "tree-vectorizer.h" /* Main analysis functions. */ static loop_vec_info vect_analyze_loop_form (struct loop *); static bool vect_analyze_data_refs (loop_vec_info); static bool vect_mark_stmts_to_be_vectorized (loop_vec_info); static void vect_analyze_scalar_cycles (loop_vec_info); static bool vect_analyze_data_ref_accesses (loop_vec_info); static bool vect_analyze_data_ref_dependences (loop_vec_info); static bool vect_analyze_data_refs_alignment (loop_vec_info); static bool vect_compute_data_refs_alignment (loop_vec_info); static bool vect_enhance_data_refs_alignment (loop_vec_info); static bool vect_analyze_operations (loop_vec_info); static bool vect_determine_vectorization_factor (loop_vec_info); /* Utility functions for the analyses. */ static bool exist_non_indexing_operands_for_use_p (tree, tree); static void vect_mark_relevant (VEC(tree,heap) **, tree, bool, bool); static bool vect_stmt_relevant_p (tree, loop_vec_info, bool *, bool *); static tree vect_get_loop_niters (struct loop *, tree *); static bool vect_analyze_data_ref_dependence (struct data_dependence_relation *, loop_vec_info); static bool vect_compute_data_ref_alignment (struct data_reference *); static bool vect_analyze_data_ref_access (struct data_reference *); static bool vect_can_advance_ivs_p (loop_vec_info); static void vect_update_misalignment_for_peel (struct data_reference *, struct data_reference *, int npeel); /* Function vect_determine_vectorization_factor Determine the vectorization factor (VF). VF is the number of data elements that are operated upon in parallel in a single iteration of the vectorized loop. For example, when vectorizing a loop that operates on 4byte elements, on a target with vector size (VS) 16byte, the VF is set to 4, since 4 elements can fit in a single vector register. We currently support vectorization of loops in which all types operated upon are of the same size. Therefore this function currently sets VF according to the size of the types operated upon, and fails if there are multiple sizes in the loop. VF is also the factor by which the loop iterations are strip-mined, e.g.: original loop: for (i=0; i<N; i++){ a[i] = b[i] + c[i]; } vectorized loop: for (i=0; i<N; i+=VF){ a[i:VF] = b[i:VF] + c[i:VF]; } */ static bool vect_determine_vectorization_factor (loop_vec_info loop_vinfo) { struct loop *loop = LOOP_VINFO_LOOP (loop_vinfo); basic_block *bbs = LOOP_VINFO_BBS (loop_vinfo); int nbbs = loop->num_nodes; block_stmt_iterator si; unsigned int vectorization_factor = 0; int i; tree scalar_type; if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "=== vect_determine_vectorization_factor ==="); for (i = 0; i < nbbs; i++) { basic_block bb = bbs[i]; for (si = bsi_start (bb); !bsi_end_p (si); bsi_next (&si)) { tree stmt = bsi_stmt (si); unsigned int nunits; stmt_vec_info stmt_info = vinfo_for_stmt (stmt); tree vectype; if (vect_print_dump_info (REPORT_DETAILS)) { fprintf (vect_dump, "==> examining statement: "); print_generic_expr (vect_dump, stmt, TDF_SLIM); } gcc_assert (stmt_info); /* skip stmts which do not need to be vectorized. */ if (!STMT_VINFO_RELEVANT_P (stmt_info) && !STMT_VINFO_LIVE_P (stmt_info)) { if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "skip."); continue; } if (VECTOR_MODE_P (TYPE_MODE (TREE_TYPE (stmt)))) { if (vect_print_dump_info (REPORT_UNVECTORIZED_LOOPS)) { fprintf (vect_dump, "not vectorized: vector stmt in loop:"); print_generic_expr (vect_dump, stmt, TDF_SLIM); } return false; } if (STMT_VINFO_VECTYPE (stmt_info)) { vectype = STMT_VINFO_VECTYPE (stmt_info); scalar_type = TREE_TYPE (vectype); } else { if (STMT_VINFO_DATA_REF (stmt_info)) scalar_type = TREE_TYPE (DR_REF (STMT_VINFO_DATA_REF (stmt_info))); else if (TREE_CODE (stmt) == MODIFY_EXPR) scalar_type = TREE_TYPE (TREE_OPERAND (stmt, 0)); else scalar_type = TREE_TYPE (stmt); if (vect_print_dump_info (REPORT_DETAILS)) { fprintf (vect_dump, "get vectype for scalar type: "); print_generic_expr (vect_dump, scalar_type, TDF_SLIM); } vectype = get_vectype_for_scalar_type (scalar_type); if (!vectype) { if (vect_print_dump_info (REPORT_UNVECTORIZED_LOOPS)) { fprintf (vect_dump, "not vectorized: unsupported data-type "); print_generic_expr (vect_dump, scalar_type, TDF_SLIM); } return false; } STMT_VINFO_VECTYPE (stmt_info) = vectype; } if (vect_print_dump_info (REPORT_DETAILS)) { fprintf (vect_dump, "vectype: "); print_generic_expr (vect_dump, vectype, TDF_SLIM); } nunits = TYPE_VECTOR_SUBPARTS (vectype); if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "nunits = %d", nunits); if (vectorization_factor) { /* FORNOW: don't allow mixed units. This restriction will be relaxed in the future. */ if (nunits != vectorization_factor) { if (vect_print_dump_info (REPORT_UNVECTORIZED_LOOPS)) fprintf (vect_dump, "not vectorized: mixed data-types"); return false; } } else vectorization_factor = nunits; gcc_assert (GET_MODE_SIZE (TYPE_MODE (scalar_type)) * vectorization_factor == UNITS_PER_SIMD_WORD); } } /* TODO: Analyze cost. Decide if worth while to vectorize. */ if (vectorization_factor <= 1) { if (vect_print_dump_info (REPORT_UNVECTORIZED_LOOPS)) fprintf (vect_dump, "not vectorized: unsupported data-type"); return false; } LOOP_VINFO_VECT_FACTOR (loop_vinfo) = vectorization_factor; return true; } /* Function vect_analyze_operations. Scan the loop stmts and make sure they are all vectorizable. */ static bool vect_analyze_operations (loop_vec_info loop_vinfo) { struct loop *loop = LOOP_VINFO_LOOP (loop_vinfo); basic_block *bbs = LOOP_VINFO_BBS (loop_vinfo); int nbbs = loop->num_nodes; block_stmt_iterator si; unsigned int vectorization_factor = 0; int i; bool ok; tree phi; stmt_vec_info stmt_info; bool need_to_vectorize = false; if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "=== vect_analyze_operations ==="); gcc_assert (LOOP_VINFO_VECT_FACTOR (loop_vinfo)); vectorization_factor = LOOP_VINFO_VECT_FACTOR (loop_vinfo); for (i = 0; i < nbbs; i++) { basic_block bb = bbs[i]; for (phi = phi_nodes (bb); phi; phi = PHI_CHAIN (phi)) { stmt_info = vinfo_for_stmt (phi); if (vect_print_dump_info (REPORT_DETAILS)) { fprintf (vect_dump, "examining phi: "); print_generic_expr (vect_dump, phi, TDF_SLIM); } gcc_assert (stmt_info); if (STMT_VINFO_LIVE_P (stmt_info)) { /* FORNOW: not yet supported. */ if (vect_print_dump_info (REPORT_UNVECTORIZED_LOOPS)) fprintf (vect_dump, "not vectorized: value used after loop."); return false; } if (STMT_VINFO_RELEVANT_P (stmt_info)) { /* Most likely a reduction-like computation that is used in the loop. */ if (vect_print_dump_info (REPORT_UNVECTORIZED_LOOPS)) fprintf (vect_dump, "not vectorized: unsupported pattern."); return false; } } for (si = bsi_start (bb); !bsi_end_p (si); bsi_next (&si)) { tree stmt = bsi_stmt (si); stmt_vec_info stmt_info = vinfo_for_stmt (stmt); if (vect_print_dump_info (REPORT_DETAILS)) { fprintf (vect_dump, "==> examining statement: "); print_generic_expr (vect_dump, stmt, TDF_SLIM); } gcc_assert (stmt_info); /* skip stmts which do not need to be vectorized. this is expected to include: - the COND_EXPR which is the loop exit condition - any LABEL_EXPRs in the loop - computations that are used only for array indexing or loop control */ if (!STMT_VINFO_RELEVANT_P (stmt_info) && !STMT_VINFO_LIVE_P (stmt_info)) { if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "irrelevant."); continue; } if (STMT_VINFO_RELEVANT_P (stmt_info)) { gcc_assert (!VECTOR_MODE_P (TYPE_MODE (TREE_TYPE (stmt)))); gcc_assert (STMT_VINFO_VECTYPE (stmt_info)); ok = (vectorizable_operation (stmt, NULL, NULL) || vectorizable_assignment (stmt, NULL, NULL) || vectorizable_load (stmt, NULL, NULL) || vectorizable_store (stmt, NULL, NULL) || vectorizable_condition (stmt, NULL, NULL)); if (!ok) { if (vect_print_dump_info (REPORT_UNVECTORIZED_LOOPS)) { fprintf (vect_dump, "not vectorized: relevant stmt not supported: "); print_generic_expr (vect_dump, stmt, TDF_SLIM); } return false; } need_to_vectorize = true; } if (STMT_VINFO_LIVE_P (stmt_info)) { ok = vectorizable_reduction (stmt, NULL, NULL); if (ok) need_to_vectorize = true; else ok = vectorizable_live_operation (stmt, NULL, NULL); if (!ok) { if (vect_print_dump_info (REPORT_UNVECTORIZED_LOOPS)) { fprintf (vect_dump, "not vectorized: live stmt not supported: "); print_generic_expr (vect_dump, stmt, TDF_SLIM); } return false; } } } /* stmts in bb */ } /* bbs */ /* TODO: Analyze cost. Decide if worth while to vectorize. */ /* All operations in the loop are either irrelevant (deal with loop control, or dead), or only used outside the loop and can be moved out of the loop (e.g. invariants, inductions). The loop can be optimized away by scalar optimizations. We're better off not touching this loop. */ if (!need_to_vectorize) { if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "All the computation can be taken out of the loop."); if (vect_print_dump_info (REPORT_UNVECTORIZED_LOOPS)) fprintf (vect_dump, "not vectorized: redundant loop. no profit to vectorize."); return false; } if (LOOP_VINFO_NITERS_KNOWN_P (loop_vinfo) && vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "vectorization_factor = %d, niters = " HOST_WIDE_INT_PRINT_DEC, vectorization_factor, LOOP_VINFO_INT_NITERS (loop_vinfo)); if (LOOP_VINFO_NITERS_KNOWN_P (loop_vinfo) && LOOP_VINFO_INT_NITERS (loop_vinfo) < vectorization_factor) { if (vect_print_dump_info (REPORT_UNVECTORIZED_LOOPS)) fprintf (vect_dump, "not vectorized: iteration count too small."); return false; } if (!LOOP_VINFO_NITERS_KNOWN_P (loop_vinfo) || LOOP_VINFO_INT_NITERS (loop_vinfo) % vectorization_factor != 0 || LOOP_PEELING_FOR_ALIGNMENT (loop_vinfo)) { if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "epilog loop required."); if (!vect_can_advance_ivs_p (loop_vinfo)) { if (vect_print_dump_info (REPORT_UNVECTORIZED_LOOPS)) fprintf (vect_dump, "not vectorized: can't create epilog loop 1."); return false; } if (!slpeel_can_duplicate_loop_p (loop, loop->single_exit)) { if (vect_print_dump_info (REPORT_UNVECTORIZED_LOOPS)) fprintf (vect_dump, "not vectorized: can't create epilog loop 2."); return false; } } return true; } /* Function exist_non_indexing_operands_for_use_p USE is one of the uses attached to STMT. Check if USE is used in STMT for anything other than indexing an array. */ static bool exist_non_indexing_operands_for_use_p (tree use, tree stmt) { tree operand; stmt_vec_info stmt_info = vinfo_for_stmt (stmt); /* USE corresponds to some operand in STMT. If there is no data reference in STMT, then any operand that corresponds to USE is not indexing an array. */ if (!STMT_VINFO_DATA_REF (stmt_info)) return true; /* STMT has a data_ref. FORNOW this means that its of one of the following forms: -1- ARRAY_REF = var -2- var = ARRAY_REF (This should have been verified in analyze_data_refs). 'var' in the second case corresponds to a def, not a use, so USE cannot correspond to any operands that are not used for array indexing. Therefore, all we need to check is if STMT falls into the first case, and whether var corresponds to USE. */ if (TREE_CODE (TREE_OPERAND (stmt, 0)) == SSA_NAME) return false; operand = TREE_OPERAND (stmt, 1); if (TREE_CODE (operand) != SSA_NAME) return false; if (operand == use) return true; return false; } /* Function vect_analyze_scalar_cycles. Examine the cross iteration def-use cycles of scalar variables, by analyzing the loop (scalar) PHIs; Classify each cycle as one of the following: invariant, induction, reduction, unknown. Some forms of scalar cycles are not yet supported. Example1: reduction: (unsupported yet) loop1: for (i=0; i<N; i++) sum += a[i]; Example2: induction: (unsupported yet) loop2: for (i=0; i<N; i++) a[i] = i; Note: the following loop *is* vectorizable: loop3: for (i=0; i<N; i++) a[i] = b[i]; even though it has a def-use cycle caused by the induction variable i: loop: i_2 = PHI (i_0, i_1) a[i_2] = ...; i_1 = i_2 + 1; GOTO loop; because the def-use cycle in loop3 is considered "not relevant" - i.e., it does not need to be vectorized because it is only used for array indexing (see 'mark_stmts_to_be_vectorized'). The def-use cycle in loop2 on the other hand is relevant (it is being written to memory). */ static void vect_analyze_scalar_cycles (loop_vec_info loop_vinfo) { tree phi; struct loop *loop = LOOP_VINFO_LOOP (loop_vinfo); basic_block bb = loop->header; tree dummy; if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "=== vect_analyze_scalar_cycles ==="); for (phi = phi_nodes (bb); phi; phi = PHI_CHAIN (phi)) { tree access_fn = NULL; tree def = PHI_RESULT (phi); stmt_vec_info stmt_vinfo = vinfo_for_stmt (phi); tree reduc_stmt; if (vect_print_dump_info (REPORT_DETAILS)) { fprintf (vect_dump, "Analyze phi: "); print_generic_expr (vect_dump, phi, TDF_SLIM); } /* Skip virtual phi's. The data dependences that are associated with virtual defs/uses (i.e., memory accesses) are analyzed elsewhere. */ if (!is_gimple_reg (SSA_NAME_VAR (def))) { if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "virtual phi. skip."); continue; } STMT_VINFO_DEF_TYPE (stmt_vinfo) = vect_unknown_def_type; /* Analyze the evolution function. */ access_fn = analyze_scalar_evolution (loop, def); if (!access_fn) continue; if (vect_print_dump_info (REPORT_DETAILS)) { fprintf (vect_dump, "Access function of PHI: "); print_generic_expr (vect_dump, access_fn, TDF_SLIM); } if (vect_is_simple_iv_evolution (loop->num, access_fn, &dummy, &dummy)) { if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "Detected induction."); STMT_VINFO_DEF_TYPE (stmt_vinfo) = vect_induction_def; continue; } /* TODO: handle invariant phis */ reduc_stmt = vect_is_simple_reduction (loop, phi); if (reduc_stmt) { if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "Detected reduction."); STMT_VINFO_DEF_TYPE (stmt_vinfo) = vect_reduction_def; STMT_VINFO_DEF_TYPE (vinfo_for_stmt (reduc_stmt)) = vect_reduction_def; } else if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "Unknown def-use cycle pattern."); } return; } /* Function vect_analyze_data_ref_dependence. Return TRUE if there (might) exist a dependence between a memory-reference DRA and a memory-reference DRB. */ static bool vect_analyze_data_ref_dependence (struct data_dependence_relation *ddr, loop_vec_info loop_vinfo) { unsigned int i; struct loop *loop = LOOP_VINFO_LOOP (loop_vinfo); int vectorization_factor = LOOP_VINFO_VECT_FACTOR (loop_vinfo); struct data_reference *dra = DDR_A (ddr); struct data_reference *drb = DDR_B (ddr); stmt_vec_info stmtinfo_a = vinfo_for_stmt (DR_STMT (dra)); stmt_vec_info stmtinfo_b = vinfo_for_stmt (DR_STMT (drb)); lambda_vector dist_v; unsigned int loop_depth; if (DDR_ARE_DEPENDENT (ddr) == chrec_known) return false; if (DDR_ARE_DEPENDENT (ddr) == chrec_dont_know) { if (vect_print_dump_info (REPORT_UNVECTORIZED_LOOPS)) { fprintf (vect_dump, "not vectorized: can't determine dependence between "); print_generic_expr (vect_dump, DR_REF (dra), TDF_SLIM); fprintf (vect_dump, " and "); print_generic_expr (vect_dump, DR_REF (drb), TDF_SLIM); } return true; } if (DDR_NUM_DIST_VECTS (ddr) == 0) { if (vect_print_dump_info (REPORT_UNVECTORIZED_LOOPS)) { fprintf (vect_dump, "not vectorized: bad dist vector for "); print_generic_expr (vect_dump, DR_REF (dra), TDF_SLIM); fprintf (vect_dump, " and "); print_generic_expr (vect_dump, DR_REF (drb), TDF_SLIM); } return true; } loop_depth = index_in_loop_nest (loop->num, DDR_LOOP_NEST (ddr)); for (i = 0; VEC_iterate (lambda_vector, DDR_DIST_VECTS (ddr), i, dist_v); i++) { int dist = dist_v[loop_depth]; if (vect_print_dump_info (REPORT_DR_DETAILS)) fprintf (vect_dump, "dependence distance = %d.", dist); /* Same loop iteration. */ if (dist % vectorization_factor == 0) { /* Two references with distance zero have the same alignment. */ VEC_safe_push (dr_p, heap, STMT_VINFO_SAME_ALIGN_REFS (stmtinfo_a), drb); VEC_safe_push (dr_p, heap, STMT_VINFO_SAME_ALIGN_REFS (stmtinfo_b), dra); if (vect_print_dump_info (REPORT_ALIGNMENT)) fprintf (vect_dump, "accesses have the same alignment."); if (vect_print_dump_info (REPORT_DR_DETAILS)) { fprintf (vect_dump, "dependence distance modulo vf == 0 between "); print_generic_expr (vect_dump, DR_REF (dra), TDF_SLIM); fprintf (vect_dump, " and "); print_generic_expr (vect_dump, DR_REF (drb), TDF_SLIM); } continue; } if (abs (dist) >= vectorization_factor) { /* Dependence distance does not create dependence, as far as vectorization is concerned, in this case. */ if (vect_print_dump_info (REPORT_DR_DETAILS)) fprintf (vect_dump, "dependence distance >= VF."); continue; } if (vect_print_dump_info (REPORT_UNVECTORIZED_LOOPS)) { fprintf (vect_dump, "not vectorized: possible dependence between data-refs "); print_generic_expr (vect_dump, DR_REF (dra), TDF_SLIM); fprintf (vect_dump, " and "); print_generic_expr (vect_dump, DR_REF (drb), TDF_SLIM); } return true; } return false; } /* Function vect_analyze_data_ref_dependences. Examine all the data references in the loop, and make sure there do not exist any data dependences between them. */ static bool vect_analyze_data_ref_dependences (loop_vec_info loop_vinfo) { unsigned int i; VEC (ddr_p, heap) *ddrs = LOOP_VINFO_DDRS (loop_vinfo); struct data_dependence_relation *ddr; if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "=== vect_analyze_dependences ==="); for (i = 0; VEC_iterate (ddr_p, ddrs, i, ddr); i++) if (vect_analyze_data_ref_dependence (ddr, loop_vinfo)) return false; return true; } /* Function vect_compute_data_ref_alignment Compute the misalignment of the data reference DR. Output: 1. If during the misalignment computation it is found that the data reference cannot be vectorized then false is returned. 2. DR_MISALIGNMENT (DR) is defined. FOR NOW: No analysis is actually performed. Misalignment is calculated only for trivial cases. TODO. */ static bool vect_compute_data_ref_alignment (struct data_reference *dr) { tree stmt = DR_STMT (dr); stmt_vec_info stmt_info = vinfo_for_stmt (stmt); tree ref = DR_REF (dr); tree vectype; tree base, base_addr; bool base_aligned; tree misalign; tree aligned_to, alignment; if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "vect_compute_data_ref_alignment:"); /* Initialize misalignment to unknown. */ DR_MISALIGNMENT (dr) = -1; misalign = DR_OFFSET_MISALIGNMENT (dr); aligned_to = DR_ALIGNED_TO (dr); base_addr = DR_BASE_ADDRESS (dr); base = build_fold_indirect_ref (base_addr); vectype = STMT_VINFO_VECTYPE (stmt_info); alignment = ssize_int (TYPE_ALIGN (vectype)/BITS_PER_UNIT); if ((aligned_to && tree_int_cst_compare (aligned_to, alignment) < 0) || !misalign) { if (vect_print_dump_info (REPORT_DETAILS)) { fprintf (vect_dump, "Unknown alignment for access: "); print_generic_expr (vect_dump, base, TDF_SLIM); } return true; } if ((DECL_P (base) && tree_int_cst_compare (ssize_int (DECL_ALIGN_UNIT (base)), alignment) >= 0) || (TREE_CODE (base_addr) == SSA_NAME && tree_int_cst_compare (ssize_int (TYPE_ALIGN_UNIT (TREE_TYPE ( TREE_TYPE (base_addr)))), alignment) >= 0)) base_aligned = true; else base_aligned = false; if (!base_aligned) { /* Do not change the alignment of global variables if flag_section_anchors is enabled. */ if (!vect_can_force_dr_alignment_p (base, TYPE_ALIGN (vectype)) || (TREE_STATIC (base) && flag_section_anchors)) { if (vect_print_dump_info (REPORT_DETAILS)) { fprintf (vect_dump, "can't force alignment of ref: "); print_generic_expr (vect_dump, ref, TDF_SLIM); } return true; } /* Force the alignment of the decl. NOTE: This is the only change to the code we make during the analysis phase, before deciding to vectorize the loop. */ if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "force alignment"); DECL_ALIGN (base) = TYPE_ALIGN (vectype); DECL_USER_ALIGN (base) = 1; } /* At this point we assume that the base is aligned. */ gcc_assert (base_aligned || (TREE_CODE (base) == VAR_DECL && DECL_ALIGN (base) >= TYPE_ALIGN (vectype))); /* Modulo alignment. */ misalign = size_binop (TRUNC_MOD_EXPR, misalign, alignment); if (!host_integerp (misalign, 1)) { /* Negative or overflowed misalignment value. */ if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "unexpected misalign value"); return false; } DR_MISALIGNMENT (dr) = TREE_INT_CST_LOW (misalign); if (vect_print_dump_info (REPORT_DETAILS)) { fprintf (vect_dump, "misalign = %d bytes of ref ", DR_MISALIGNMENT (dr)); print_generic_expr (vect_dump, ref, TDF_SLIM); } return true; } /* Function vect_compute_data_refs_alignment Compute the misalignment of data references in the loop. Return FALSE if a data reference is found that cannot be vectorized. */ static bool vect_compute_data_refs_alignment (loop_vec_info loop_vinfo) { VEC (data_reference_p, heap) *datarefs = LOOP_VINFO_DATAREFS (loop_vinfo); struct data_reference *dr; unsigned int i; for (i = 0; VEC_iterate (data_reference_p, datarefs, i, dr); i++) if (!vect_compute_data_ref_alignment (dr)) return false; return true; } /* Function vect_update_misalignment_for_peel DR - the data reference whose misalignment is to be adjusted. DR_PEEL - the data reference whose misalignment is being made zero in the vector loop by the peel. NPEEL - the number of iterations in the peel loop if the misalignment of DR_PEEL is known at compile time. */ static void vect_update_misalignment_for_peel (struct data_reference *dr, struct data_reference *dr_peel, int npeel) { unsigned int i; int drsize; VEC(dr_p,heap) *same_align_drs; struct data_reference *current_dr; if (known_alignment_for_access_p (dr) && DR_MISALIGNMENT (dr) == DR_MISALIGNMENT (dr_peel)) { DR_MISALIGNMENT (dr) = 0; return; } /* It can be assumed that the data refs with the same alignment as dr_peel are aligned in the vector loop. */ same_align_drs = STMT_VINFO_SAME_ALIGN_REFS (vinfo_for_stmt (DR_STMT (dr_peel))); for (i = 0; VEC_iterate (dr_p, same_align_drs, i, current_dr); i++) { if (current_dr != dr) continue; gcc_assert (DR_MISALIGNMENT (dr) == DR_MISALIGNMENT (dr_peel)); DR_MISALIGNMENT (dr) = 0; return; } if (known_alignment_for_access_p (dr) && known_alignment_for_access_p (dr_peel)) { drsize = GET_MODE_SIZE (TYPE_MODE (TREE_TYPE (DR_REF (dr)))); DR_MISALIGNMENT (dr) += npeel * drsize; DR_MISALIGNMENT (dr) %= UNITS_PER_SIMD_WORD; return; } DR_MISALIGNMENT (dr) = -1; } /* Function vect_verify_datarefs_alignment Return TRUE if all data references in the loop can be handled with respect to alignment. */ static bool vect_verify_datarefs_alignment (loop_vec_info loop_vinfo) { VEC (data_reference_p, heap) *datarefs = LOOP_VINFO_DATAREFS (loop_vinfo); struct data_reference *dr; enum dr_alignment_support supportable_dr_alignment; unsigned int i; for (i = 0; VEC_iterate (data_reference_p, datarefs, i, dr); i++) { supportable_dr_alignment = vect_supportable_dr_alignment (dr); if (!supportable_dr_alignment) { if (vect_print_dump_info (REPORT_UNVECTORIZED_LOOPS)) { if (DR_IS_READ (dr)) fprintf (vect_dump, "not vectorized: unsupported unaligned load."); else fprintf (vect_dump, "not vectorized: unsupported unaligned store."); } return false; } if (supportable_dr_alignment != dr_aligned && vect_print_dump_info (REPORT_ALIGNMENT)) fprintf (vect_dump, "Vectorizing an unaligned access."); } return true; } /* Function vector_alignment_reachable_p Return true if vector alignment for DR is reachable by peeling a few loop iterations. Return false otherwise. */ static bool vector_alignment_reachable_p (struct data_reference *dr) { tree stmt = DR_STMT (dr); stmt_vec_info stmt_info = vinfo_for_stmt (stmt); tree vectype = STMT_VINFO_VECTYPE (stmt_info); /* If misalignment is known at the compile time then allow peeling only if natural alignment is reachable through peeling. */ if (known_alignment_for_access_p (dr) && !aligned_access_p (dr)) { HOST_WIDE_INT elmsize = int_cst_value (TYPE_SIZE_UNIT (TREE_TYPE (vectype))); if (vect_print_dump_info (REPORT_DETAILS)) { fprintf (vect_dump, "data size =" HOST_WIDE_INT_PRINT_DEC, elmsize); fprintf (vect_dump, ". misalignment = %d. ", DR_MISALIGNMENT (dr)); } if (DR_MISALIGNMENT (dr) % elmsize) { if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "data size does not divide the misalignment.\n"); return false; } } if (!known_alignment_for_access_p (dr)) { tree type = (TREE_TYPE (DR_REF (dr))); tree ba = DR_BASE_OBJECT (dr); bool is_packed = false; if (ba) is_packed = contains_packed_reference (ba); if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "Unknown misalignment, is_packed = %d",is_packed); if (targetm.vectorize.vector_alignment_reachable (type, is_packed)) return true; else return false; } return true; } /* Function vect_enhance_data_refs_alignment This pass will use loop versioning and loop peeling in order to enhance the alignment of data references in the loop. FOR NOW: we assume that whatever versioning/peeling takes place, only the original loop is to be vectorized; Any other loops that are created by the transformations performed in this pass - are not supposed to be vectorized. This restriction will be relaxed. This pass will require a cost model to guide it whether to apply peeling or versioning or a combination of the two. For example, the scheme that intel uses when given a loop with several memory accesses, is as follows: choose one memory access ('p') which alignment you want to force by doing peeling. Then, either (1) generate a loop in which 'p' is aligned and all other accesses are not necessarily aligned, or (2) use loop versioning to generate one loop in which all accesses are aligned, and another loop in which only 'p' is necessarily aligned. ("Automatic Intra-Register Vectorization for the Intel Architecture", Aart J.C. Bik, Milind Girkar, Paul M. Grey and Ximmin Tian, International Journal of Parallel Programming, Vol. 30, No. 2, April 2002.) Devising a cost model is the most critical aspect of this work. It will guide us on which access to peel for, whether to use loop versioning, how many versions to create, etc. The cost model will probably consist of generic considerations as well as target specific considerations (on powerpc for example, misaligned stores are more painful than misaligned loads). Here are the general steps involved in alignment enhancements: -- original loop, before alignment analysis: for (i=0; i<N; i++){ x = q[i]; # DR_MISALIGNMENT(q) = unknown p[i] = y; # DR_MISALIGNMENT(p) = unknown } -- After vect_compute_data_refs_alignment: for (i=0; i<N; i++){ x = q[i]; # DR_MISALIGNMENT(q) = 3 p[i] = y; # DR_MISALIGNMENT(p) = unknown } -- Possibility 1: we do loop versioning: if (p is aligned) { for (i=0; i<N; i++){ # loop 1A x = q[i]; # DR_MISALIGNMENT(q) = 3 p[i] = y; # DR_MISALIGNMENT(p) = 0 } } else { for (i=0; i<N; i++){ # loop 1B x = q[i]; # DR_MISALIGNMENT(q) = 3 p[i] = y; # DR_MISALIGNMENT(p) = unaligned } } -- Possibility 2: we do loop peeling: for (i = 0; i < 3; i++){ # (scalar loop, not to be vectorized). x = q[i]; p[i] = y; } for (i = 3; i < N; i++){ # loop 2A x = q[i]; # DR_MISALIGNMENT(q) = 0 p[i] = y; # DR_MISALIGNMENT(p) = unknown } -- Possibility 3: combination of loop peeling and versioning: for (i = 0; i < 3; i++){ # (scalar loop, not to be vectorized). x = q[i]; p[i] = y; } if (p is aligned) { for (i = 3; i<N; i++){ # loop 3A x = q[i]; # DR_MISALIGNMENT(q) = 0 p[i] = y; # DR_MISALIGNMENT(p) = 0 } } else { for (i = 3; i<N; i++){ # loop 3B x = q[i]; # DR_MISALIGNMENT(q) = 0 p[i] = y; # DR_MISALIGNMENT(p) = unaligned } } These loops are later passed to loop_transform to be vectorized. The vectorizer will use the alignment information to guide the transformation (whether to generate regular loads/stores, or with special handling for misalignment). */ static bool vect_enhance_data_refs_alignment (loop_vec_info loop_vinfo) { VEC (data_reference_p, heap) *datarefs = LOOP_VINFO_DATAREFS (loop_vinfo); enum dr_alignment_support supportable_dr_alignment; struct data_reference *dr0 = NULL; struct data_reference *dr; unsigned int i; bool do_peeling = false; bool do_versioning = false; bool stat; /* While cost model enhancements are expected in the future, the high level view of the code at this time is as follows: A) If there is a misaligned write then see if peeling to align this write can make all data references satisfy vect_supportable_dr_alignment. If so, update data structures as needed and return true. Note that at this time vect_supportable_dr_alignment is known to return false for a a misaligned write. B) If peeling wasn't possible and there is a data reference with an unknown misalignment that does not satisfy vect_supportable_dr_alignment then see if loop versioning checks can be used to make all data references satisfy vect_supportable_dr_alignment. If so, update data structures as needed and return true. C) If neither peeling nor versioning were successful then return false if any data reference does not satisfy vect_supportable_dr_alignment. D) Return true (all data references satisfy vect_supportable_dr_alignment). Note, Possibility 3 above (which is peeling and versioning together) is not being done at this time. */ /* (1) Peeling to force alignment. */ /* (1.1) Decide whether to perform peeling, and how many iterations to peel: Considerations: + How many accesses will become aligned due to the peeling - How many accesses will become unaligned due to the peeling, and the cost of misaligned accesses. - The cost of peeling (the extra runtime checks, the increase in code size). The scheme we use FORNOW: peel to force the alignment of the first misaligned store in the loop. Rationale: misaligned stores are not yet supported. TODO: Use a cost model. */ for (i = 0; VEC_iterate (data_reference_p, datarefs, i, dr); i++) if (!DR_IS_READ (dr) && !aligned_access_p (dr)) { do_peeling = vector_alignment_reachable_p (dr); if (do_peeling) dr0 = dr; if (!do_peeling && vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "vector alignment may not be reachable"); break; } /* Often peeling for alignment will require peeling for loop-bound, which in turn requires that we know how to adjust the loop ivs after the loop. */ if (!vect_can_advance_ivs_p (loop_vinfo)) do_peeling = false; if (do_peeling) { int mis; int npeel = 0; if (known_alignment_for_access_p (dr0)) { /* Since it's known at compile time, compute the number of iterations in the peeled loop (the peeling factor) for use in updating DR_MISALIGNMENT values. The peeling factor is the vectorization factor minus the misalignment as an element count. */ mis = DR_MISALIGNMENT (dr0); mis /= GET_MODE_SIZE (TYPE_MODE (TREE_TYPE (DR_REF (dr0)))); npeel = LOOP_VINFO_VECT_FACTOR (loop_vinfo) - mis; } /* Ensure that all data refs can be vectorized after the peel. */ for (i = 0; VEC_iterate (data_reference_p, datarefs, i, dr); i++) { int save_misalignment; if (dr == dr0) continue; save_misalignment = DR_MISALIGNMENT (dr); vect_update_misalignment_for_peel (dr, dr0, npeel); supportable_dr_alignment = vect_supportable_dr_alignment (dr); DR_MISALIGNMENT (dr) = save_misalignment; if (!supportable_dr_alignment) { do_peeling = false; break; } } if (do_peeling) { /* (1.2) Update the DR_MISALIGNMENT of each data reference DR_i. If the misalignment of DR_i is identical to that of dr0 then set DR_MISALIGNMENT (DR_i) to zero. If the misalignment of DR_i and dr0 are known at compile time then increment DR_MISALIGNMENT (DR_i) by the peeling factor times the element size of DR_i (MOD the vectorization factor times the size). Otherwise, the misalignment of DR_i must be set to unknown. */ for (i = 0; VEC_iterate (data_reference_p, datarefs, i, dr); i++) if (dr != dr0) vect_update_misalignment_for_peel (dr, dr0, npeel); LOOP_VINFO_UNALIGNED_DR (loop_vinfo) = dr0; LOOP_PEELING_FOR_ALIGNMENT (loop_vinfo) = DR_MISALIGNMENT (dr0); DR_MISALIGNMENT (dr0) = 0; if (vect_print_dump_info (REPORT_ALIGNMENT)) fprintf (vect_dump, "Alignment of access forced using peeling."); if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "Peeling for alignment will be applied."); stat = vect_verify_datarefs_alignment (loop_vinfo); gcc_assert (stat); return stat; } } /* (2) Versioning to force alignment. */ /* Try versioning if: 1) flag_tree_vect_loop_version is TRUE 2) optimize_size is FALSE 3) there is at least one unsupported misaligned data ref with an unknown misalignment, and 4) all misaligned data refs with a known misalignment are supported, and 5) the number of runtime alignment checks is within reason. */ do_versioning = flag_tree_vect_loop_version && (!optimize_size); if (do_versioning) { for (i = 0; VEC_iterate (data_reference_p, datarefs, i, dr); i++) { if (aligned_access_p (dr)) continue; supportable_dr_alignment = vect_supportable_dr_alignment (dr); if (!supportable_dr_alignment) { tree stmt; int mask; tree vectype; if (known_alignment_for_access_p (dr) || VEC_length (tree, LOOP_VINFO_MAY_MISALIGN_STMTS (loop_vinfo)) >= (unsigned) PARAM_VALUE (PARAM_VECT_MAX_VERSION_CHECKS)) { do_versioning = false; break; } stmt = DR_STMT (dr); vectype = STMT_VINFO_VECTYPE (vinfo_for_stmt (stmt)); gcc_assert (vectype); /* The rightmost bits of an aligned address must be zeros. Construct the mask needed for this test. For example, GET_MODE_SIZE for the vector mode V4SI is 16 bytes so the mask must be 15 = 0xf. */ mask = GET_MODE_SIZE (TYPE_MODE (vectype)) - 1; /* FORNOW: use the same mask to test all potentially unaligned references in the loop. The vectorizer currently supports a single vector size, see the reference to GET_MODE_NUNITS (TYPE_MODE (vectype)) where the vectorization factor is computed. */ gcc_assert (!LOOP_VINFO_PTR_MASK (loop_vinfo) || LOOP_VINFO_PTR_MASK (loop_vinfo) == mask); LOOP_VINFO_PTR_MASK (loop_vinfo) = mask; VEC_safe_push (tree, heap, LOOP_VINFO_MAY_MISALIGN_STMTS (loop_vinfo), DR_STMT (dr)); } } /* Versioning requires at least one misaligned data reference. */ if (VEC_length (tree, LOOP_VINFO_MAY_MISALIGN_STMTS (loop_vinfo)) == 0) do_versioning = false; else if (!do_versioning) VEC_truncate (tree, LOOP_VINFO_MAY_MISALIGN_STMTS (loop_vinfo), 0); } if (do_versioning) { VEC(tree,heap) *may_misalign_stmts = LOOP_VINFO_MAY_MISALIGN_STMTS (loop_vinfo); tree stmt; /* It can now be assumed that the data references in the statements in LOOP_VINFO_MAY_MISALIGN_STMTS will be aligned in the version of the loop being vectorized. */ for (i = 0; VEC_iterate (tree, may_misalign_stmts, i, stmt); i++) { stmt_vec_info stmt_info = vinfo_for_stmt (stmt); dr = STMT_VINFO_DATA_REF (stmt_info); DR_MISALIGNMENT (dr) = 0; if (vect_print_dump_info (REPORT_ALIGNMENT)) fprintf (vect_dump, "Alignment of access forced using versioning."); } if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "Versioning for alignment will be applied."); /* Peeling and versioning can't be done together at this time. */ gcc_assert (! (do_peeling && do_versioning)); stat = vect_verify_datarefs_alignment (loop_vinfo); gcc_assert (stat); return stat; } /* This point is reached if neither peeling nor versioning is being done. */ gcc_assert (! (do_peeling || do_versioning)); stat = vect_verify_datarefs_alignment (loop_vinfo); return stat; } /* Function vect_analyze_data_refs_alignment Analyze the alignment of the data-references in the loop. Return FALSE if a data reference is found that cannot be vectorized. */ static bool vect_analyze_data_refs_alignment (loop_vec_info loop_vinfo) { if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "=== vect_analyze_data_refs_alignment ==="); if (!vect_compute_data_refs_alignment (loop_vinfo)) { if (vect_print_dump_info (REPORT_UNVECTORIZED_LOOPS)) fprintf (vect_dump, "not vectorized: can't calculate alignment for data ref."); return false; } return true; } /* Function vect_analyze_data_ref_access. Analyze the access pattern of the data-reference DR. For now, a data access has to be consecutive to be considered vectorizable. */ static bool vect_analyze_data_ref_access (struct data_reference *dr) { tree step = DR_STEP (dr); tree scalar_type = TREE_TYPE (DR_REF (dr)); if (!step || tree_int_cst_compare (step, TYPE_SIZE_UNIT (scalar_type))) { if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "not consecutive access"); return false; } return true; } /* Function vect_analyze_data_ref_accesses. Analyze the access pattern of all the data references in the loop. FORNOW: the only access pattern that is considered vectorizable is a simple step 1 (consecutive) access. FORNOW: handle only arrays and pointer accesses. */ static bool vect_analyze_data_ref_accesses (loop_vec_info loop_vinfo) { unsigned int i; VEC (data_reference_p, heap) *datarefs = LOOP_VINFO_DATAREFS (loop_vinfo); struct data_reference *dr; if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "=== vect_analyze_data_ref_accesses ==="); for (i = 0; VEC_iterate (data_reference_p, datarefs, i, dr); i++) if (!vect_analyze_data_ref_access (dr)) { if (vect_print_dump_info (REPORT_UNVECTORIZED_LOOPS)) fprintf (vect_dump, "not vectorized: complicated access pattern."); return false; } return true; } /* Function vect_analyze_data_refs. Find all the data references in the loop. The general structure of the analysis of data refs in the vectorizer is as follows: 1- vect_analyze_data_refs(loop): call compute_data_dependences_for_loop to find and analyze all data-refs in the loop and their dependences. 2- vect_analyze_dependences(): apply dependence testing using ddrs. 3- vect_analyze_drs_alignment(): check that ref_stmt.alignment is ok. 4- vect_analyze_drs_access(): check that ref_stmt.step is ok. */ static bool vect_analyze_data_refs (loop_vec_info loop_vinfo) { struct loop *loop = LOOP_VINFO_LOOP (loop_vinfo); unsigned int i; VEC (data_reference_p, heap) *datarefs; struct data_reference *dr; tree scalar_type; if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "=== vect_analyze_data_refs ==="); compute_data_dependences_for_loop (loop, false, &LOOP_VINFO_DATAREFS (loop_vinfo), &LOOP_VINFO_DDRS (loop_vinfo)); /* Go through the data-refs, check that the analysis succeeded. Update pointer from stmt_vec_info struct to DR and vectype. */ datarefs = LOOP_VINFO_DATAREFS (loop_vinfo); for (i = 0; VEC_iterate (data_reference_p, datarefs, i, dr); i++) { tree stmt; stmt_vec_info stmt_info; if (!dr || !DR_REF (dr)) { if (vect_print_dump_info (REPORT_UNVECTORIZED_LOOPS)) fprintf (vect_dump, "not vectorized: unhandled data-ref "); return false; } /* Update DR field in stmt_vec_info struct. */ stmt = DR_STMT (dr); stmt_info = vinfo_for_stmt (stmt); if (STMT_VINFO_DATA_REF (stmt_info)) { if (vect_print_dump_info (REPORT_UNVECTORIZED_LOOPS)) { fprintf (vect_dump, "not vectorized: more than one data ref in stmt: "); print_generic_expr (vect_dump, stmt, TDF_SLIM); } return false; } STMT_VINFO_DATA_REF (stmt_info) = dr; /* Check that analysis of the data-ref succeeded. */ if (!DR_BASE_ADDRESS (dr) || !DR_OFFSET (dr) || !DR_INIT (dr) || !DR_STEP (dr)) { if (vect_print_dump_info (REPORT_UNVECTORIZED_LOOPS)) { fprintf (vect_dump, "not vectorized: data ref analysis failed "); print_generic_expr (vect_dump, stmt, TDF_SLIM); } return false; } if (!DR_MEMTAG (dr)) { if (vect_print_dump_info (REPORT_UNVECTORIZED_LOOPS)) { fprintf (vect_dump, "not vectorized: no memory tag for "); print_generic_expr (vect_dump, DR_REF (dr), TDF_SLIM); } return false; } /* Set vectype for STMT. */ scalar_type = TREE_TYPE (DR_REF (dr)); STMT_VINFO_VECTYPE (stmt_info) = get_vectype_for_scalar_type (scalar_type); if (!STMT_VINFO_VECTYPE (stmt_info)) { if (vect_print_dump_info (REPORT_UNVECTORIZED_LOOPS)) { fprintf (vect_dump, "not vectorized: no vectype for stmt: "); print_generic_expr (vect_dump, stmt, TDF_SLIM); fprintf (vect_dump, " scalar_type: "); print_generic_expr (vect_dump, scalar_type, TDF_DETAILS); } return false; } } return true; } /* Utility functions used by vect_mark_stmts_to_be_vectorized. */ /* Function vect_mark_relevant. Mark STMT as "relevant for vectorization" and add it to WORKLIST. */ static void vect_mark_relevant (VEC(tree,heap) **worklist, tree stmt, bool relevant_p, bool live_p) { stmt_vec_info stmt_info = vinfo_for_stmt (stmt); bool save_relevant_p = STMT_VINFO_RELEVANT_P (stmt_info); bool save_live_p = STMT_VINFO_LIVE_P (stmt_info); if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "mark relevant %d, live %d.",relevant_p, live_p); if (STMT_VINFO_IN_PATTERN_P (stmt_info)) { tree pattern_stmt; /* This is the last stmt in a sequence that was detected as a pattern that can potentially be vectorized. Don't mark the stmt as relevant/live because it's not going to vectorized. Instead mark the pattern-stmt that replaces it. */ if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "last stmt in pattern. don't mark relevant/live."); pattern_stmt = STMT_VINFO_RELATED_STMT (stmt_info); stmt_info = vinfo_for_stmt (pattern_stmt); gcc_assert (STMT_VINFO_RELATED_STMT (stmt_info) == stmt); save_relevant_p = STMT_VINFO_RELEVANT_P (stmt_info); save_live_p = STMT_VINFO_LIVE_P (stmt_info); stmt = pattern_stmt; } STMT_VINFO_LIVE_P (stmt_info) |= live_p; STMT_VINFO_RELEVANT_P (stmt_info) |= relevant_p; if (TREE_CODE (stmt) == PHI_NODE) /* Don't put phi-nodes in the worklist. Phis that are marked relevant or live will fail vectorization later on. */ return; if (STMT_VINFO_RELEVANT_P (stmt_info) == save_relevant_p && STMT_VINFO_LIVE_P (stmt_info) == save_live_p) { if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "already marked relevant/live."); return; } VEC_safe_push (tree, heap, *worklist, stmt); } /* Function vect_stmt_relevant_p. Return true if STMT in loop that is represented by LOOP_VINFO is "relevant for vectorization". A stmt is considered "relevant for vectorization" if: - it has uses outside the loop. - it has vdefs (it alters memory). - control stmts in the loop (except for the exit condition). CHECKME: what other side effects would the vectorizer allow? */ static bool vect_stmt_relevant_p (tree stmt, loop_vec_info loop_vinfo, bool *relevant_p, bool *live_p) { struct loop *loop = LOOP_VINFO_LOOP (loop_vinfo); ssa_op_iter op_iter; imm_use_iterator imm_iter; use_operand_p use_p; def_operand_p def_p; *relevant_p = false; *live_p = false; /* cond stmt other than loop exit cond. */ if (is_ctrl_stmt (stmt) && (stmt != LOOP_VINFO_EXIT_COND (loop_vinfo))) *relevant_p = true; /* changing memory. */ if (TREE_CODE (stmt) != PHI_NODE) if (!ZERO_SSA_OPERANDS (stmt, SSA_OP_VIRTUAL_DEFS)) { if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "vec_stmt_relevant_p: stmt has vdefs."); *relevant_p = true; } /* uses outside the loop. */ FOR_EACH_PHI_OR_STMT_DEF (def_p, stmt, op_iter, SSA_OP_DEF) { FOR_EACH_IMM_USE_FAST (use_p, imm_iter, DEF_FROM_PTR (def_p)) { basic_block bb = bb_for_stmt (USE_STMT (use_p)); if (!flow_bb_inside_loop_p (loop, bb)) { if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "vec_stmt_relevant_p: used out of loop."); /* We expect all such uses to be in the loop exit phis (because of loop closed form) */ gcc_assert (TREE_CODE (USE_STMT (use_p)) == PHI_NODE); gcc_assert (bb == loop->single_exit->dest); *live_p = true; } } } return (*live_p || *relevant_p); } /* Function vect_mark_stmts_to_be_vectorized. Not all stmts in the loop need to be vectorized. For example: for i... for j... 1. T0 = i + j 2. T1 = a[T0] 3. j = j + 1 Stmt 1 and 3 do not need to be vectorized, because loop control and addressing of vectorized data-refs are handled differently. This pass detects such stmts. */ static bool vect_mark_stmts_to_be_vectorized (loop_vec_info loop_vinfo) { VEC(tree,heap) *worklist; struct loop *loop = LOOP_VINFO_LOOP (loop_vinfo); basic_block *bbs = LOOP_VINFO_BBS (loop_vinfo); unsigned int nbbs = loop->num_nodes; block_stmt_iterator si; tree stmt, use; stmt_ann_t ann; ssa_op_iter iter; unsigned int i; stmt_vec_info stmt_vinfo; basic_block bb; tree phi; bool relevant_p, live_p; tree def, def_stmt; enum vect_def_type dt; if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "=== vect_mark_stmts_to_be_vectorized ==="); worklist = VEC_alloc (tree, heap, 64); /* 1. Init worklist. */ bb = loop->header; for (phi = phi_nodes (bb); phi; phi = PHI_CHAIN (phi)) { if (vect_print_dump_info (REPORT_DETAILS)) { fprintf (vect_dump, "init: phi relevant? "); print_generic_expr (vect_dump, phi, TDF_SLIM); } if (vect_stmt_relevant_p (phi, loop_vinfo, &relevant_p, &live_p)) vect_mark_relevant (&worklist, phi, relevant_p, live_p); } for (i = 0; i < nbbs; i++) { bb = bbs[i]; for (si = bsi_start (bb); !bsi_end_p (si); bsi_next (&si)) { stmt = bsi_stmt (si); if (vect_print_dump_info (REPORT_DETAILS)) { fprintf (vect_dump, "init: stmt relevant? "); print_generic_expr (vect_dump, stmt, TDF_SLIM); } if (vect_stmt_relevant_p (stmt, loop_vinfo, &relevant_p, &live_p)) vect_mark_relevant (&worklist, stmt, relevant_p, live_p); } } /* 2. Process_worklist */ while (VEC_length (tree, worklist) > 0) { stmt = VEC_pop (tree, worklist); if (vect_print_dump_info (REPORT_DETAILS)) { fprintf (vect_dump, "worklist: examine stmt: "); print_generic_expr (vect_dump, stmt, TDF_SLIM); } /* Examine the USEs of STMT. For each ssa-name USE thta is defined in the loop, mark the stmt that defines it (DEF_STMT) as relevant/irrelevant and live/dead according to the liveness and relevance properties of STMT. */ gcc_assert (TREE_CODE (stmt) != PHI_NODE); ann = stmt_ann (stmt); stmt_vinfo = vinfo_for_stmt (stmt); relevant_p = STMT_VINFO_RELEVANT_P (stmt_vinfo); live_p = STMT_VINFO_LIVE_P (stmt_vinfo); /* Generally, the liveness and relevance properties of STMT are propagated to the DEF_STMTs of its USEs: STMT_VINFO_LIVE_P (DEF_STMT_info) <-- live_p STMT_VINFO_RELEVANT_P (DEF_STMT_info) <-- relevant_p Exceptions: (case 1) If USE is used only for address computations (e.g. array indexing), which does not need to be directly vectorized, then the liveness/relevance of the respective DEF_STMT is left unchanged. (case 2) If STMT has been identified as defining a reduction variable, then we have two cases: (case 2.1) The last use of STMT is the reduction-variable, which is defined by a loop-header-phi. We don't want to mark the phi as live or relevant (because it does not need to be vectorized, it is handled as part of the vectorization of the reduction), so in this case we skip the call to vect_mark_relevant. (case 2.2) The rest of the uses of STMT are defined in the loop body. For the def_stmt of these uses we want to set liveness/relevance as follows: STMT_VINFO_LIVE_P (DEF_STMT_info) <-- false STMT_VINFO_RELEVANT_P (DEF_STMT_info) <-- true because even though STMT is classified as live (since it defines a value that is used across loop iterations) and irrelevant (since it is not used inside the loop), it will be vectorized, and therefore the corresponding DEF_STMTs need to marked as relevant. */ /* case 2.2: */ if (STMT_VINFO_DEF_TYPE (stmt_vinfo) == vect_reduction_def) { gcc_assert (!relevant_p && live_p); relevant_p = true; live_p = false; } FOR_EACH_SSA_TREE_OPERAND (use, stmt, iter, SSA_OP_USE) { /* case 1: we are only interested in uses that need to be vectorized. Uses that are used for address computation are not considered relevant. */ if (!exist_non_indexing_operands_for_use_p (use, stmt)) continue; if (!vect_is_simple_use (use, loop_vinfo, &def_stmt, &def, &dt)) { if (vect_print_dump_info (REPORT_UNVECTORIZED_LOOPS)) fprintf (vect_dump, "not vectorized: unsupported use in stmt."); VEC_free (tree, heap, worklist); return false; } if (!def_stmt || IS_EMPTY_STMT (def_stmt)) continue; if (vect_print_dump_info (REPORT_DETAILS)) { fprintf (vect_dump, "worklist: examine use %d: ", i); print_generic_expr (vect_dump, use, TDF_SLIM); } bb = bb_for_stmt (def_stmt); if (!flow_bb_inside_loop_p (loop, bb)) continue; /* case 2.1: the reduction-use does not mark the defining-phi as relevant. */ if (STMT_VINFO_DEF_TYPE (stmt_vinfo) == vect_reduction_def && TREE_CODE (def_stmt) == PHI_NODE) continue; vect_mark_relevant (&worklist, def_stmt, relevant_p, live_p); } } /* while worklist */ VEC_free (tree, heap, worklist); return true; } /* Function vect_can_advance_ivs_p In case the number of iterations that LOOP iterates is unknown at compile time, an epilog loop will be generated, and the loop induction variables (IVs) will be "advanced" to the value they are supposed to take just before the epilog loop. Here we check that the access function of the loop IVs and the expression that represents the loop bound are simple enough. These restrictions will be relaxed in the future. */ static bool vect_can_advance_ivs_p (loop_vec_info loop_vinfo) { struct loop *loop = LOOP_VINFO_LOOP (loop_vinfo); basic_block bb = loop->header; tree phi; /* Analyze phi functions of the loop header. */ if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "=== vect_can_advance_ivs_p ==="); for (phi = phi_nodes (bb); phi; phi = PHI_CHAIN (phi)) { tree access_fn = NULL; tree evolution_part; if (vect_print_dump_info (REPORT_DETAILS)) { fprintf (vect_dump, "Analyze phi: "); print_generic_expr (vect_dump, phi, TDF_SLIM); } /* Skip virtual phi's. The data dependences that are associated with virtual defs/uses (i.e., memory accesses) are analyzed elsewhere. */ if (!is_gimple_reg (SSA_NAME_VAR (PHI_RESULT (phi)))) { if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "virtual phi. skip."); continue; } /* Skip reduction phis. */ if (STMT_VINFO_DEF_TYPE (vinfo_for_stmt (phi)) == vect_reduction_def) { if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "reduc phi. skip."); continue; } /* Analyze the evolution function. */ access_fn = instantiate_parameters (loop, analyze_scalar_evolution (loop, PHI_RESULT (phi))); if (!access_fn) { if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "No Access function."); return false; } if (vect_print_dump_info (REPORT_DETAILS)) { fprintf (vect_dump, "Access function of PHI: "); print_generic_expr (vect_dump, access_fn, TDF_SLIM); } evolution_part = evolution_part_in_loop_num (access_fn, loop->num); if (evolution_part == NULL_TREE) { if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "No evolution."); return false; } /* FORNOW: We do not transform initial conditions of IVs which evolution functions are a polynomial of degree >= 2. */ if (tree_is_chrec (evolution_part)) return false; } return true; } /* Function vect_get_loop_niters. Determine how many iterations the loop is executed. If an expression that represents the number of iterations can be constructed, place it in NUMBER_OF_ITERATIONS. Return the loop exit condition. */ static tree vect_get_loop_niters (struct loop *loop, tree *number_of_iterations) { tree niters; if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "=== get_loop_niters ==="); niters = number_of_iterations_in_loop (loop); if (niters != NULL_TREE && niters != chrec_dont_know) { *number_of_iterations = niters; if (vect_print_dump_info (REPORT_DETAILS)) { fprintf (vect_dump, "==> get_loop_niters:" ); print_generic_expr (vect_dump, *number_of_iterations, TDF_SLIM); } } return get_loop_exit_condition (loop); } /* Function vect_analyze_loop_form. Verify the following restrictions (some may be relaxed in the future): - it's an inner-most loop - number of BBs = 2 (which are the loop header and the latch) - the loop has a pre-header - the loop has a single entry and exit - the loop exit condition is simple enough, and the number of iterations can be analyzed (a countable loop). */ static loop_vec_info vect_analyze_loop_form (struct loop *loop) { loop_vec_info loop_vinfo; tree loop_cond; tree number_of_iterations = NULL; if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "=== vect_analyze_loop_form ==="); if (loop->inner) { if (vect_print_dump_info (REPORT_OUTER_LOOPS)) fprintf (vect_dump, "not vectorized: nested loop."); return NULL; } if (!loop->single_exit || loop->num_nodes != 2 || EDGE_COUNT (loop->header->preds) != 2) { if (vect_print_dump_info (REPORT_BAD_FORM_LOOPS)) { if (!loop->single_exit) fprintf (vect_dump, "not vectorized: multiple exits."); else if (loop->num_nodes != 2) fprintf (vect_dump, "not vectorized: too many BBs in loop."); else if (EDGE_COUNT (loop->header->preds) != 2) fprintf (vect_dump, "not vectorized: too many incoming edges."); } return NULL; } /* We assume that the loop exit condition is at the end of the loop. i.e, that the loop is represented as a do-while (with a proper if-guard before the loop if needed), where the loop header contains all the executable statements, and the latch is empty. */ if (!empty_block_p (loop->latch) || phi_nodes (loop->latch)) { if (vect_print_dump_info (REPORT_BAD_FORM_LOOPS)) fprintf (vect_dump, "not vectorized: unexpected loop form."); return NULL; } /* Make sure there exists a single-predecessor exit bb: */ if (!single_pred_p (loop->single_exit->dest)) { edge e = loop->single_exit; if (!(e->flags & EDGE_ABNORMAL)) { split_loop_exit_edge (e); if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "split exit edge."); } else { if (vect_print_dump_info (REPORT_BAD_FORM_LOOPS)) fprintf (vect_dump, "not vectorized: abnormal loop exit edge."); return NULL; } } if (empty_block_p (loop->header)) { if (vect_print_dump_info (REPORT_BAD_FORM_LOOPS)) fprintf (vect_dump, "not vectorized: empty loop."); return NULL; } loop_cond = vect_get_loop_niters (loop, &number_of_iterations); if (!loop_cond) { if (vect_print_dump_info (REPORT_BAD_FORM_LOOPS)) fprintf (vect_dump, "not vectorized: complicated exit condition."); return NULL; } if (!number_of_iterations) { if (vect_print_dump_info (REPORT_BAD_FORM_LOOPS)) fprintf (vect_dump, "not vectorized: number of iterations cannot be computed."); return NULL; } if (chrec_contains_undetermined (number_of_iterations)) { if (vect_print_dump_info (REPORT_BAD_FORM_LOOPS)) fprintf (vect_dump, "Infinite number of iterations."); return false; } loop_vinfo = new_loop_vec_info (loop); LOOP_VINFO_NITERS (loop_vinfo) = number_of_iterations; if (!LOOP_VINFO_NITERS_KNOWN_P (loop_vinfo)) { if (vect_print_dump_info (REPORT_DETAILS)) { fprintf (vect_dump, "Symbolic number of iterations is "); print_generic_expr (vect_dump, number_of_iterations, TDF_DETAILS); } } else if (LOOP_VINFO_INT_NITERS (loop_vinfo) == 0) { if (vect_print_dump_info (REPORT_UNVECTORIZED_LOOPS)) fprintf (vect_dump, "not vectorized: number of iterations = 0."); return NULL; } LOOP_VINFO_EXIT_COND (loop_vinfo) = loop_cond; return loop_vinfo; } /* Function vect_analyze_loop. Apply a set of analyses on LOOP, and create a loop_vec_info struct for it. The different analyses will record information in the loop_vec_info struct. */ loop_vec_info vect_analyze_loop (struct loop *loop) { bool ok; loop_vec_info loop_vinfo; if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "===== analyze_loop_nest ====="); /* Check the CFG characteristics of the loop (nesting, entry/exit, etc. */ loop_vinfo = vect_analyze_loop_form (loop); if (!loop_vinfo) { if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "bad loop form."); return NULL; } /* Find all data references in the loop (which correspond to vdefs/vuses) and analyze their evolution in the loop. FORNOW: Handle only simple, array references, which alignment can be forced, and aligned pointer-references. */ ok = vect_analyze_data_refs (loop_vinfo); if (!ok) { if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "bad data references."); destroy_loop_vec_info (loop_vinfo); return NULL; } /* Classify all cross-iteration scalar data-flow cycles. Cross-iteration cycles caused by virtual phis are analyzed separately. */ vect_analyze_scalar_cycles (loop_vinfo); vect_pattern_recog (loop_vinfo); /* Data-flow analysis to detect stmts that do not need to be vectorized. */ ok = vect_mark_stmts_to_be_vectorized (loop_vinfo); if (!ok) { if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "unexpected pattern."); destroy_loop_vec_info (loop_vinfo); return NULL; } /* Analyze the alignment of the data-refs in the loop. Fail if a data reference is found that cannot be vectorized. */ ok = vect_analyze_data_refs_alignment (loop_vinfo); if (!ok) { if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "bad data alignment."); destroy_loop_vec_info (loop_vinfo); return NULL; } ok = vect_determine_vectorization_factor (loop_vinfo); if (!ok) { if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "can't determine vectorization factor."); destroy_loop_vec_info (loop_vinfo); return NULL; } /* Analyze data dependences between the data-refs in the loop. FORNOW: fail at the first data dependence that we encounter. */ ok = vect_analyze_data_ref_dependences (loop_vinfo); if (!ok) { if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "bad data dependence."); destroy_loop_vec_info (loop_vinfo); return NULL; } /* Analyze the access patterns of the data-refs in the loop (consecutive, complex, etc.). FORNOW: Only handle consecutive access pattern. */ ok = vect_analyze_data_ref_accesses (loop_vinfo); if (!ok) { if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "bad data access."); destroy_loop_vec_info (loop_vinfo); return NULL; } /* This pass will decide on using loop versioning and/or loop peeling in order to enhance the alignment of data references in the loop. */ ok = vect_enhance_data_refs_alignment (loop_vinfo); if (!ok) { if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "bad data alignment."); destroy_loop_vec_info (loop_vinfo); return NULL; } /* Scan all the operations in the loop and make sure they are vectorizable. */ ok = vect_analyze_operations (loop_vinfo); if (!ok) { if (vect_print_dump_info (REPORT_DETAILS)) fprintf (vect_dump, "bad operation or unsupported loop bound."); destroy_loop_vec_info (loop_vinfo); return NULL; } LOOP_VINFO_VECTORIZABLE_P (loop_vinfo) = 1; return loop_vinfo; }
Go to most recent revision | Compare with Previous | Blame | View Log