2 Copyright (C) 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010, 2011, 2012
3 Free Software Foundation, Inc.
4 Contributed by Dorit Naishlos <dorit@il.ibm.com> and
5 Ira Rosen <irar@il.ibm.com>
7 This file is part of GCC.
9 GCC is free software; you can redistribute it and/or modify it under
10 the terms of the GNU General Public License as published by the Free
11 Software Foundation; either version 3, or (at your option) any later
14 GCC is distributed in the hope that it will be useful, but WITHOUT ANY
15 WARRANTY; without even the implied warranty of MERCHANTABILITY or
16 FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
19 You should have received a copy of the GNU General Public License
20 along with GCC; see the file COPYING3. If not see
21 <http://www.gnu.org/licenses/>. */
25 #include "coretypes.h"
29 #include "basic-block.h"
30 #include "tree-pretty-print.h"
31 #include "gimple-pretty-print.h"
32 #include "tree-flow.h"
33 #include "tree-dump.h"
35 #include "cfglayout.h"
40 #include "diagnostic-core.h"
41 #include "tree-chrec.h"
42 #include "tree-scalar-evolution.h"
43 #include "tree-vectorizer.h"
46 /* Loop Vectorization Pass.
48 This pass tries to vectorize loops.
50 For example, the vectorizer transforms the following simple loop:
52 short a[N]; short b[N]; short c[N]; int i;
58 as if it was manually vectorized by rewriting the source code into:
60 typedef int __attribute__((mode(V8HI))) v8hi;
61 short a[N]; short b[N]; short c[N]; int i;
62 v8hi *pa = (v8hi*)a, *pb = (v8hi*)b, *pc = (v8hi*)c;
65 for (i=0; i<N/8; i++){
72 The main entry to this pass is vectorize_loops(), in which
73 the vectorizer applies a set of analyses on a given set of loops,
74 followed by the actual vectorization transformation for the loops that
75 had successfully passed the analysis phase.
76 Throughout this pass we make a distinction between two types of
77 data: scalars (which are represented by SSA_NAMES), and memory references
78 ("data-refs"). These two types of data require different handling both
79 during analysis and transformation. The types of data-refs that the
80 vectorizer currently supports are ARRAY_REFS which base is an array DECL
81 (not a pointer), and INDIRECT_REFS through pointers; both array and pointer
82 accesses are required to have a simple (consecutive) access pattern.
86 The driver for the analysis phase is vect_analyze_loop().
87 It applies a set of analyses, some of which rely on the scalar evolution
88 analyzer (scev) developed by Sebastian Pop.
90 During the analysis phase the vectorizer records some information
91 per stmt in a "stmt_vec_info" struct which is attached to each stmt in the
92 loop, as well as general information about the loop as a whole, which is
93 recorded in a "loop_vec_info" struct attached to each loop.
97 The loop transformation phase scans all the stmts in the loop, and
98 creates a vector stmt (or a sequence of stmts) for each scalar stmt S in
99 the loop that needs to be vectorized. It inserts the vector code sequence
100 just before the scalar stmt S, and records a pointer to the vector code
101 in STMT_VINFO_VEC_STMT (stmt_info) (stmt_info is the stmt_vec_info struct
102 attached to S). This pointer will be used for the vectorization of following
103 stmts which use the def of stmt S. Stmt S is removed if it writes to memory;
104 otherwise, we rely on dead code elimination for removing it.
106 For example, say stmt S1 was vectorized into stmt VS1:
109 S1: b = x[i]; STMT_VINFO_VEC_STMT (stmt_info (S1)) = VS1
112 To vectorize stmt S2, the vectorizer first finds the stmt that defines
113 the operand 'b' (S1), and gets the relevant vector def 'vb' from the
114 vector stmt VS1 pointed to by STMT_VINFO_VEC_STMT (stmt_info (S1)). The
115 resulting sequence would be:
118 S1: b = x[i]; STMT_VINFO_VEC_STMT (stmt_info (S1)) = VS1
120 S2: a = b; STMT_VINFO_VEC_STMT (stmt_info (S2)) = VS2
122 Operands that are not SSA_NAMEs, are data-refs that appear in
123 load/store operations (like 'x[i]' in S1), and are handled differently.
127 Currently the only target specific information that is used is the
128 size of the vector (in bytes) - "TARGET_VECTORIZE_UNITS_PER_SIMD_WORD".
129 Targets that can support different sizes of vectors, for now will need
130 to specify one value for "TARGET_VECTORIZE_UNITS_PER_SIMD_WORD". More
131 flexibility will be added in the future.
133 Since we only vectorize operations which vector form can be
134 expressed using existing tree codes, to verify that an operation is
135 supported, the vectorizer checks the relevant optab at the relevant
136 machine_mode (e.g, optab_handler (add_optab, V8HImode)). If
137 the value found is CODE_FOR_nothing, then there's no target support, and
138 we can't vectorize the stmt.
140 For additional information on this project see:
141 http://gcc.gnu.org/projects/tree-ssa/vectorization.html
144 /* Function vect_determine_vectorization_factor
146 Determine the vectorization factor (VF). VF is the number of data elements
147 that are operated upon in parallel in a single iteration of the vectorized
148 loop. For example, when vectorizing a loop that operates on 4byte elements,
149 on a target with vector size (VS) 16byte, the VF is set to 4, since 4
150 elements can fit in a single vector register.
152 We currently support vectorization of loops in which all types operated upon
153 are of the same size. Therefore this function currently sets VF according to
154 the size of the types operated upon, and fails if there are multiple sizes
157 VF is also the factor by which the loop iterations are strip-mined, e.g.:
164 for (i=0; i<N; i+=VF){
165 a[i:VF] = b[i:VF] + c[i:VF];
170 vect_determine_vectorization_factor (loop_vec_info loop_vinfo)
172 struct loop *loop = LOOP_VINFO_LOOP (loop_vinfo);
173 basic_block *bbs = LOOP_VINFO_BBS (loop_vinfo);
174 int nbbs = loop->num_nodes;
175 gimple_stmt_iterator si;
176 unsigned int vectorization_factor = 0;
181 stmt_vec_info stmt_info;
184 gimple stmt, pattern_stmt = NULL;
185 gimple_seq pattern_def_seq = NULL;
186 gimple_stmt_iterator pattern_def_si = gsi_start (NULL);
187 bool analyze_pattern_stmt = false;
189 if (vect_print_dump_info (REPORT_DETAILS))
190 fprintf (vect_dump, "=== vect_determine_vectorization_factor ===");
192 for (i = 0; i < nbbs; i++)
194 basic_block bb = bbs[i];
196 for (si = gsi_start_phis (bb); !gsi_end_p (si); gsi_next (&si))
199 stmt_info = vinfo_for_stmt (phi);
200 if (vect_print_dump_info (REPORT_DETAILS))
202 fprintf (vect_dump, "==> examining phi: ");
203 print_gimple_stmt (vect_dump, phi, 0, TDF_SLIM);
206 gcc_assert (stmt_info);
208 if (STMT_VINFO_RELEVANT_P (stmt_info))
210 gcc_assert (!STMT_VINFO_VECTYPE (stmt_info));
211 scalar_type = TREE_TYPE (PHI_RESULT (phi));
213 if (vect_print_dump_info (REPORT_DETAILS))
215 fprintf (vect_dump, "get vectype for scalar type: ");
216 print_generic_expr (vect_dump, scalar_type, TDF_SLIM);
219 vectype = get_vectype_for_scalar_type (scalar_type);
222 if (vect_print_dump_info (REPORT_UNVECTORIZED_LOCATIONS))
225 "not vectorized: unsupported data-type ");
226 print_generic_expr (vect_dump, scalar_type, TDF_SLIM);
230 STMT_VINFO_VECTYPE (stmt_info) = vectype;
232 if (vect_print_dump_info (REPORT_DETAILS))
234 fprintf (vect_dump, "vectype: ");
235 print_generic_expr (vect_dump, vectype, TDF_SLIM);
238 nunits = TYPE_VECTOR_SUBPARTS (vectype);
239 if (vect_print_dump_info (REPORT_DETAILS))
240 fprintf (vect_dump, "nunits = %d", nunits);
242 if (!vectorization_factor
243 || (nunits > vectorization_factor))
244 vectorization_factor = nunits;
248 for (si = gsi_start_bb (bb); !gsi_end_p (si) || analyze_pattern_stmt;)
252 if (analyze_pattern_stmt)
255 stmt = gsi_stmt (si);
257 stmt_info = vinfo_for_stmt (stmt);
259 if (vect_print_dump_info (REPORT_DETAILS))
261 fprintf (vect_dump, "==> examining statement: ");
262 print_gimple_stmt (vect_dump, stmt, 0, TDF_SLIM);
265 gcc_assert (stmt_info);
267 /* Skip stmts which do not need to be vectorized. */
268 if (!STMT_VINFO_RELEVANT_P (stmt_info)
269 && !STMT_VINFO_LIVE_P (stmt_info))
271 if (STMT_VINFO_IN_PATTERN_P (stmt_info)
272 && (pattern_stmt = STMT_VINFO_RELATED_STMT (stmt_info))
273 && (STMT_VINFO_RELEVANT_P (vinfo_for_stmt (pattern_stmt))
274 || STMT_VINFO_LIVE_P (vinfo_for_stmt (pattern_stmt))))
277 stmt_info = vinfo_for_stmt (pattern_stmt);
278 if (vect_print_dump_info (REPORT_DETAILS))
280 fprintf (vect_dump, "==> examining pattern statement: ");
281 print_gimple_stmt (vect_dump, stmt, 0, TDF_SLIM);
286 if (vect_print_dump_info (REPORT_DETAILS))
287 fprintf (vect_dump, "skip.");
292 else if (STMT_VINFO_IN_PATTERN_P (stmt_info)
293 && (pattern_stmt = STMT_VINFO_RELATED_STMT (stmt_info))
294 && (STMT_VINFO_RELEVANT_P (vinfo_for_stmt (pattern_stmt))
295 || STMT_VINFO_LIVE_P (vinfo_for_stmt (pattern_stmt))))
296 analyze_pattern_stmt = true;
298 /* If a pattern statement has def stmts, analyze them too. */
299 if (is_pattern_stmt_p (stmt_info))
301 if (pattern_def_seq == NULL)
303 pattern_def_seq = STMT_VINFO_PATTERN_DEF_SEQ (stmt_info);
304 pattern_def_si = gsi_start (pattern_def_seq);
306 else if (!gsi_end_p (pattern_def_si))
307 gsi_next (&pattern_def_si);
308 if (pattern_def_seq != NULL)
310 gimple pattern_def_stmt = NULL;
311 stmt_vec_info pattern_def_stmt_info = NULL;
313 while (!gsi_end_p (pattern_def_si))
315 pattern_def_stmt = gsi_stmt (pattern_def_si);
316 pattern_def_stmt_info
317 = vinfo_for_stmt (pattern_def_stmt);
318 if (STMT_VINFO_RELEVANT_P (pattern_def_stmt_info)
319 || STMT_VINFO_LIVE_P (pattern_def_stmt_info))
321 gsi_next (&pattern_def_si);
324 if (!gsi_end_p (pattern_def_si))
326 if (vect_print_dump_info (REPORT_DETAILS))
329 "==> examining pattern def stmt: ");
330 print_gimple_stmt (vect_dump, pattern_def_stmt, 0,
334 stmt = pattern_def_stmt;
335 stmt_info = pattern_def_stmt_info;
339 pattern_def_si = gsi_start (NULL);
340 analyze_pattern_stmt = false;
344 analyze_pattern_stmt = false;
347 if (gimple_get_lhs (stmt) == NULL_TREE)
349 if (vect_print_dump_info (REPORT_UNVECTORIZED_LOCATIONS))
351 fprintf (vect_dump, "not vectorized: irregular stmt.");
352 print_gimple_stmt (vect_dump, stmt, 0, TDF_SLIM);
357 if (VECTOR_MODE_P (TYPE_MODE (gimple_expr_type (stmt))))
359 if (vect_print_dump_info (REPORT_UNVECTORIZED_LOCATIONS))
361 fprintf (vect_dump, "not vectorized: vector stmt in loop:");
362 print_gimple_stmt (vect_dump, stmt, 0, TDF_SLIM);
367 if (STMT_VINFO_VECTYPE (stmt_info))
369 /* The only case when a vectype had been already set is for stmts
370 that contain a dataref, or for "pattern-stmts" (stmts
371 generated by the vectorizer to represent/replace a certain
373 gcc_assert (STMT_VINFO_DATA_REF (stmt_info)
374 || is_pattern_stmt_p (stmt_info)
375 || !gsi_end_p (pattern_def_si));
376 vectype = STMT_VINFO_VECTYPE (stmt_info);
380 gcc_assert (!STMT_VINFO_DATA_REF (stmt_info));
381 scalar_type = TREE_TYPE (gimple_get_lhs (stmt));
382 if (vect_print_dump_info (REPORT_DETAILS))
384 fprintf (vect_dump, "get vectype for scalar type: ");
385 print_generic_expr (vect_dump, scalar_type, TDF_SLIM);
387 vectype = get_vectype_for_scalar_type (scalar_type);
390 if (vect_print_dump_info (REPORT_UNVECTORIZED_LOCATIONS))
393 "not vectorized: unsupported data-type ");
394 print_generic_expr (vect_dump, scalar_type, TDF_SLIM);
399 STMT_VINFO_VECTYPE (stmt_info) = vectype;
402 /* The vectorization factor is according to the smallest
403 scalar type (or the largest vector size, but we only
404 support one vector size per loop). */
405 scalar_type = vect_get_smallest_scalar_type (stmt, &dummy,
407 if (vect_print_dump_info (REPORT_DETAILS))
409 fprintf (vect_dump, "get vectype for scalar type: ");
410 print_generic_expr (vect_dump, scalar_type, TDF_SLIM);
412 vf_vectype = get_vectype_for_scalar_type (scalar_type);
415 if (vect_print_dump_info (REPORT_UNVECTORIZED_LOCATIONS))
418 "not vectorized: unsupported data-type ");
419 print_generic_expr (vect_dump, scalar_type, TDF_SLIM);
424 if ((GET_MODE_SIZE (TYPE_MODE (vectype))
425 != GET_MODE_SIZE (TYPE_MODE (vf_vectype))))
427 if (vect_print_dump_info (REPORT_UNVECTORIZED_LOCATIONS))
430 "not vectorized: different sized vector "
431 "types in statement, ");
432 print_generic_expr (vect_dump, vectype, TDF_SLIM);
433 fprintf (vect_dump, " and ");
434 print_generic_expr (vect_dump, vf_vectype, TDF_SLIM);
439 if (vect_print_dump_info (REPORT_DETAILS))
441 fprintf (vect_dump, "vectype: ");
442 print_generic_expr (vect_dump, vf_vectype, TDF_SLIM);
445 nunits = TYPE_VECTOR_SUBPARTS (vf_vectype);
446 if (vect_print_dump_info (REPORT_DETAILS))
447 fprintf (vect_dump, "nunits = %d", nunits);
449 if (!vectorization_factor
450 || (nunits > vectorization_factor))
451 vectorization_factor = nunits;
453 if (!analyze_pattern_stmt && gsi_end_p (pattern_def_si))
455 pattern_def_seq = NULL;
461 /* TODO: Analyze cost. Decide if worth while to vectorize. */
462 if (vect_print_dump_info (REPORT_DETAILS))
463 fprintf (vect_dump, "vectorization factor = %d", vectorization_factor);
464 if (vectorization_factor <= 1)
466 if (vect_print_dump_info (REPORT_UNVECTORIZED_LOCATIONS))
467 fprintf (vect_dump, "not vectorized: unsupported data-type");
470 LOOP_VINFO_VECT_FACTOR (loop_vinfo) = vectorization_factor;
476 /* Function vect_is_simple_iv_evolution.
478 FORNOW: A simple evolution of an induction variables in the loop is
479 considered a polynomial evolution with constant step. */
482 vect_is_simple_iv_evolution (unsigned loop_nb, tree access_fn, tree * init,
487 tree evolution_part = evolution_part_in_loop_num (access_fn, loop_nb);
489 /* When there is no evolution in this loop, the evolution function
491 if (evolution_part == NULL_TREE)
494 /* When the evolution is a polynomial of degree >= 2
495 the evolution function is not "simple". */
496 if (tree_is_chrec (evolution_part))
499 step_expr = evolution_part;
500 init_expr = unshare_expr (initial_condition_in_loop_num (access_fn, loop_nb));
502 if (vect_print_dump_info (REPORT_DETAILS))
504 fprintf (vect_dump, "step: ");
505 print_generic_expr (vect_dump, step_expr, TDF_SLIM);
506 fprintf (vect_dump, ", init: ");
507 print_generic_expr (vect_dump, init_expr, TDF_SLIM);
513 if (TREE_CODE (step_expr) != INTEGER_CST)
515 if (vect_print_dump_info (REPORT_DETAILS))
516 fprintf (vect_dump, "step unknown.");
523 /* Function vect_analyze_scalar_cycles_1.
525 Examine the cross iteration def-use cycles of scalar variables
526 in LOOP. LOOP_VINFO represents the loop that is now being
527 considered for vectorization (can be LOOP, or an outer-loop
531 vect_analyze_scalar_cycles_1 (loop_vec_info loop_vinfo, struct loop *loop)
533 basic_block bb = loop->header;
535 VEC(gimple,heap) *worklist = VEC_alloc (gimple, heap, 64);
536 gimple_stmt_iterator gsi;
539 if (vect_print_dump_info (REPORT_DETAILS))
540 fprintf (vect_dump, "=== vect_analyze_scalar_cycles ===");
542 /* First - identify all inductions. Reduction detection assumes that all the
543 inductions have been identified, therefore, this order must not be
545 for (gsi = gsi_start_phis (bb); !gsi_end_p (gsi); gsi_next (&gsi))
547 gimple phi = gsi_stmt (gsi);
548 tree access_fn = NULL;
549 tree def = PHI_RESULT (phi);
550 stmt_vec_info stmt_vinfo = vinfo_for_stmt (phi);
552 if (vect_print_dump_info (REPORT_DETAILS))
554 fprintf (vect_dump, "Analyze phi: ");
555 print_gimple_stmt (vect_dump, phi, 0, TDF_SLIM);
558 /* Skip virtual phi's. The data dependences that are associated with
559 virtual defs/uses (i.e., memory accesses) are analyzed elsewhere. */
560 if (!is_gimple_reg (SSA_NAME_VAR (def)))
563 STMT_VINFO_DEF_TYPE (stmt_vinfo) = vect_unknown_def_type;
565 /* Analyze the evolution function. */
566 access_fn = analyze_scalar_evolution (loop, def);
569 STRIP_NOPS (access_fn);
570 if (vect_print_dump_info (REPORT_DETAILS))
572 fprintf (vect_dump, "Access function of PHI: ");
573 print_generic_expr (vect_dump, access_fn, TDF_SLIM);
575 STMT_VINFO_LOOP_PHI_EVOLUTION_PART (stmt_vinfo)
576 = evolution_part_in_loop_num (access_fn, loop->num);
580 || !vect_is_simple_iv_evolution (loop->num, access_fn, &dumy, &dumy))
582 VEC_safe_push (gimple, heap, worklist, phi);
586 gcc_assert (STMT_VINFO_LOOP_PHI_EVOLUTION_PART (stmt_vinfo) != NULL_TREE);
588 if (vect_print_dump_info (REPORT_DETAILS))
589 fprintf (vect_dump, "Detected induction.");
590 STMT_VINFO_DEF_TYPE (stmt_vinfo) = vect_induction_def;
594 /* Second - identify all reductions and nested cycles. */
595 while (VEC_length (gimple, worklist) > 0)
597 gimple phi = VEC_pop (gimple, worklist);
598 tree def = PHI_RESULT (phi);
599 stmt_vec_info stmt_vinfo = vinfo_for_stmt (phi);
603 if (vect_print_dump_info (REPORT_DETAILS))
605 fprintf (vect_dump, "Analyze phi: ");
606 print_gimple_stmt (vect_dump, phi, 0, TDF_SLIM);
609 gcc_assert (is_gimple_reg (SSA_NAME_VAR (def)));
610 gcc_assert (STMT_VINFO_DEF_TYPE (stmt_vinfo) == vect_unknown_def_type);
612 nested_cycle = (loop != LOOP_VINFO_LOOP (loop_vinfo));
613 reduc_stmt = vect_force_simple_reduction (loop_vinfo, phi, !nested_cycle,
619 if (vect_print_dump_info (REPORT_DETAILS))
620 fprintf (vect_dump, "Detected double reduction.");
622 STMT_VINFO_DEF_TYPE (stmt_vinfo) = vect_double_reduction_def;
623 STMT_VINFO_DEF_TYPE (vinfo_for_stmt (reduc_stmt)) =
624 vect_double_reduction_def;
630 if (vect_print_dump_info (REPORT_DETAILS))
631 fprintf (vect_dump, "Detected vectorizable nested cycle.");
633 STMT_VINFO_DEF_TYPE (stmt_vinfo) = vect_nested_cycle;
634 STMT_VINFO_DEF_TYPE (vinfo_for_stmt (reduc_stmt)) =
639 if (vect_print_dump_info (REPORT_DETAILS))
640 fprintf (vect_dump, "Detected reduction.");
642 STMT_VINFO_DEF_TYPE (stmt_vinfo) = vect_reduction_def;
643 STMT_VINFO_DEF_TYPE (vinfo_for_stmt (reduc_stmt)) =
645 /* Store the reduction cycles for possible vectorization in
647 VEC_safe_push (gimple, heap,
648 LOOP_VINFO_REDUCTIONS (loop_vinfo),
654 if (vect_print_dump_info (REPORT_DETAILS))
655 fprintf (vect_dump, "Unknown def-use cycle pattern.");
658 VEC_free (gimple, heap, worklist);
662 /* Function vect_analyze_scalar_cycles.
664 Examine the cross iteration def-use cycles of scalar variables, by
665 analyzing the loop-header PHIs of scalar variables. Classify each
666 cycle as one of the following: invariant, induction, reduction, unknown.
667 We do that for the loop represented by LOOP_VINFO, and also to its
668 inner-loop, if exists.
669 Examples for scalar cycles:
684 vect_analyze_scalar_cycles (loop_vec_info loop_vinfo)
686 struct loop *loop = LOOP_VINFO_LOOP (loop_vinfo);
688 vect_analyze_scalar_cycles_1 (loop_vinfo, loop);
690 /* When vectorizing an outer-loop, the inner-loop is executed sequentially.
691 Reductions in such inner-loop therefore have different properties than
692 the reductions in the nest that gets vectorized:
693 1. When vectorized, they are executed in the same order as in the original
694 scalar loop, so we can't change the order of computation when
696 2. FIXME: Inner-loop reductions can be used in the inner-loop, so the
697 current checks are too strict. */
700 vect_analyze_scalar_cycles_1 (loop_vinfo, loop->inner);
703 /* Function vect_get_loop_niters.
705 Determine how many iterations the loop is executed.
706 If an expression that represents the number of iterations
707 can be constructed, place it in NUMBER_OF_ITERATIONS.
708 Return the loop exit condition. */
711 vect_get_loop_niters (struct loop *loop, tree *number_of_iterations)
715 if (vect_print_dump_info (REPORT_DETAILS))
716 fprintf (vect_dump, "=== get_loop_niters ===");
718 niters = number_of_exit_cond_executions (loop);
720 if (niters != NULL_TREE
721 && niters != chrec_dont_know)
723 *number_of_iterations = niters;
725 if (vect_print_dump_info (REPORT_DETAILS))
727 fprintf (vect_dump, "==> get_loop_niters:" );
728 print_generic_expr (vect_dump, *number_of_iterations, TDF_SLIM);
732 return get_loop_exit_condition (loop);
736 /* Function bb_in_loop_p
738 Used as predicate for dfs order traversal of the loop bbs. */
741 bb_in_loop_p (const_basic_block bb, const void *data)
743 const struct loop *const loop = (const struct loop *)data;
744 if (flow_bb_inside_loop_p (loop, bb))
750 /* Function new_loop_vec_info.
752 Create and initialize a new loop_vec_info struct for LOOP, as well as
753 stmt_vec_info structs for all the stmts in LOOP. */
756 new_loop_vec_info (struct loop *loop)
760 gimple_stmt_iterator si;
761 unsigned int i, nbbs;
763 res = (loop_vec_info) xcalloc (1, sizeof (struct _loop_vec_info));
764 LOOP_VINFO_LOOP (res) = loop;
766 bbs = get_loop_body (loop);
768 /* Create/Update stmt_info for all stmts in the loop. */
769 for (i = 0; i < loop->num_nodes; i++)
771 basic_block bb = bbs[i];
773 /* BBs in a nested inner-loop will have been already processed (because
774 we will have called vect_analyze_loop_form for any nested inner-loop).
775 Therefore, for stmts in an inner-loop we just want to update the
776 STMT_VINFO_LOOP_VINFO field of their stmt_info to point to the new
777 loop_info of the outer-loop we are currently considering to vectorize
778 (instead of the loop_info of the inner-loop).
779 For stmts in other BBs we need to create a stmt_info from scratch. */
780 if (bb->loop_father != loop)
783 gcc_assert (loop->inner && bb->loop_father == loop->inner);
784 for (si = gsi_start_phis (bb); !gsi_end_p (si); gsi_next (&si))
786 gimple phi = gsi_stmt (si);
787 stmt_vec_info stmt_info = vinfo_for_stmt (phi);
788 loop_vec_info inner_loop_vinfo =
789 STMT_VINFO_LOOP_VINFO (stmt_info);
790 gcc_assert (loop->inner == LOOP_VINFO_LOOP (inner_loop_vinfo));
791 STMT_VINFO_LOOP_VINFO (stmt_info) = res;
793 for (si = gsi_start_bb (bb); !gsi_end_p (si); gsi_next (&si))
795 gimple stmt = gsi_stmt (si);
796 stmt_vec_info stmt_info = vinfo_for_stmt (stmt);
797 loop_vec_info inner_loop_vinfo =
798 STMT_VINFO_LOOP_VINFO (stmt_info);
799 gcc_assert (loop->inner == LOOP_VINFO_LOOP (inner_loop_vinfo));
800 STMT_VINFO_LOOP_VINFO (stmt_info) = res;
805 /* bb in current nest. */
806 for (si = gsi_start_phis (bb); !gsi_end_p (si); gsi_next (&si))
808 gimple phi = gsi_stmt (si);
809 gimple_set_uid (phi, 0);
810 set_vinfo_for_stmt (phi, new_stmt_vec_info (phi, res, NULL));
813 for (si = gsi_start_bb (bb); !gsi_end_p (si); gsi_next (&si))
815 gimple stmt = gsi_stmt (si);
816 gimple_set_uid (stmt, 0);
817 set_vinfo_for_stmt (stmt, new_stmt_vec_info (stmt, res, NULL));
822 /* CHECKME: We want to visit all BBs before their successors (except for
823 latch blocks, for which this assertion wouldn't hold). In the simple
824 case of the loop forms we allow, a dfs order of the BBs would the same
825 as reversed postorder traversal, so we are safe. */
828 bbs = XCNEWVEC (basic_block, loop->num_nodes);
829 nbbs = dfs_enumerate_from (loop->header, 0, bb_in_loop_p,
830 bbs, loop->num_nodes, loop);
831 gcc_assert (nbbs == loop->num_nodes);
833 LOOP_VINFO_BBS (res) = bbs;
834 LOOP_VINFO_NITERS (res) = NULL;
835 LOOP_VINFO_NITERS_UNCHANGED (res) = NULL;
836 LOOP_VINFO_COST_MODEL_MIN_ITERS (res) = 0;
837 LOOP_VINFO_VECTORIZABLE_P (res) = 0;
838 LOOP_PEELING_FOR_ALIGNMENT (res) = 0;
839 LOOP_VINFO_VECT_FACTOR (res) = 0;
840 LOOP_VINFO_LOOP_NEST (res) = VEC_alloc (loop_p, heap, 3);
841 LOOP_VINFO_DATAREFS (res) = VEC_alloc (data_reference_p, heap, 10);
842 LOOP_VINFO_DDRS (res) = VEC_alloc (ddr_p, heap, 10 * 10);
843 LOOP_VINFO_UNALIGNED_DR (res) = NULL;
844 LOOP_VINFO_MAY_MISALIGN_STMTS (res) =
845 VEC_alloc (gimple, heap,
846 PARAM_VALUE (PARAM_VECT_MAX_VERSION_FOR_ALIGNMENT_CHECKS));
847 LOOP_VINFO_MAY_ALIAS_DDRS (res) =
848 VEC_alloc (ddr_p, heap,
849 PARAM_VALUE (PARAM_VECT_MAX_VERSION_FOR_ALIAS_CHECKS));
850 LOOP_VINFO_STRIDED_STORES (res) = VEC_alloc (gimple, heap, 10);
851 LOOP_VINFO_REDUCTIONS (res) = VEC_alloc (gimple, heap, 10);
852 LOOP_VINFO_REDUCTION_CHAINS (res) = VEC_alloc (gimple, heap, 10);
853 LOOP_VINFO_SLP_INSTANCES (res) = VEC_alloc (slp_instance, heap, 10);
854 LOOP_VINFO_SLP_UNROLLING_FACTOR (res) = 1;
855 LOOP_VINFO_PEELING_HTAB (res) = NULL;
856 LOOP_VINFO_PEELING_FOR_GAPS (res) = false;
862 /* Function destroy_loop_vec_info.
864 Free LOOP_VINFO struct, as well as all the stmt_vec_info structs of all the
865 stmts in the loop. */
868 destroy_loop_vec_info (loop_vec_info loop_vinfo, bool clean_stmts)
873 gimple_stmt_iterator si;
875 VEC (slp_instance, heap) *slp_instances;
876 slp_instance instance;
881 loop = LOOP_VINFO_LOOP (loop_vinfo);
883 bbs = LOOP_VINFO_BBS (loop_vinfo);
884 nbbs = loop->num_nodes;
888 free (LOOP_VINFO_BBS (loop_vinfo));
889 free_data_refs (LOOP_VINFO_DATAREFS (loop_vinfo));
890 free_dependence_relations (LOOP_VINFO_DDRS (loop_vinfo));
891 VEC_free (loop_p, heap, LOOP_VINFO_LOOP_NEST (loop_vinfo));
892 VEC_free (gimple, heap, LOOP_VINFO_MAY_MISALIGN_STMTS (loop_vinfo));
893 VEC_free (ddr_p, heap, LOOP_VINFO_MAY_ALIAS_DDRS (loop_vinfo));
900 for (j = 0; j < nbbs; j++)
902 basic_block bb = bbs[j];
903 for (si = gsi_start_phis (bb); !gsi_end_p (si); gsi_next (&si))
904 free_stmt_vec_info (gsi_stmt (si));
906 for (si = gsi_start_bb (bb); !gsi_end_p (si); )
908 gimple stmt = gsi_stmt (si);
909 /* Free stmt_vec_info. */
910 free_stmt_vec_info (stmt);
915 free (LOOP_VINFO_BBS (loop_vinfo));
916 free_data_refs (LOOP_VINFO_DATAREFS (loop_vinfo));
917 free_dependence_relations (LOOP_VINFO_DDRS (loop_vinfo));
918 VEC_free (loop_p, heap, LOOP_VINFO_LOOP_NEST (loop_vinfo));
919 VEC_free (gimple, heap, LOOP_VINFO_MAY_MISALIGN_STMTS (loop_vinfo));
920 VEC_free (ddr_p, heap, LOOP_VINFO_MAY_ALIAS_DDRS (loop_vinfo));
921 slp_instances = LOOP_VINFO_SLP_INSTANCES (loop_vinfo);
922 FOR_EACH_VEC_ELT (slp_instance, slp_instances, j, instance)
923 vect_free_slp_instance (instance);
925 VEC_free (slp_instance, heap, LOOP_VINFO_SLP_INSTANCES (loop_vinfo));
926 VEC_free (gimple, heap, LOOP_VINFO_STRIDED_STORES (loop_vinfo));
927 VEC_free (gimple, heap, LOOP_VINFO_REDUCTIONS (loop_vinfo));
928 VEC_free (gimple, heap, LOOP_VINFO_REDUCTION_CHAINS (loop_vinfo));
930 if (LOOP_VINFO_PEELING_HTAB (loop_vinfo))
931 htab_delete (LOOP_VINFO_PEELING_HTAB (loop_vinfo));
938 /* Function vect_analyze_loop_1.
940 Apply a set of analyses on LOOP, and create a loop_vec_info struct
941 for it. The different analyses will record information in the
942 loop_vec_info struct. This is a subset of the analyses applied in
943 vect_analyze_loop, to be applied on an inner-loop nested in the loop
944 that is now considered for (outer-loop) vectorization. */
947 vect_analyze_loop_1 (struct loop *loop)
949 loop_vec_info loop_vinfo;
951 if (vect_print_dump_info (REPORT_DETAILS))
952 fprintf (vect_dump, "===== analyze_loop_nest_1 =====");
954 /* Check the CFG characteristics of the loop (nesting, entry/exit, etc. */
956 loop_vinfo = vect_analyze_loop_form (loop);
959 if (vect_print_dump_info (REPORT_DETAILS))
960 fprintf (vect_dump, "bad inner-loop form.");
968 /* Function vect_analyze_loop_form.
970 Verify that certain CFG restrictions hold, including:
971 - the loop has a pre-header
972 - the loop has a single entry and exit
973 - the loop exit condition is simple enough, and the number of iterations
974 can be analyzed (a countable loop). */
977 vect_analyze_loop_form (struct loop *loop)
979 loop_vec_info loop_vinfo;
981 tree number_of_iterations = NULL;
982 loop_vec_info inner_loop_vinfo = NULL;
984 if (vect_print_dump_info (REPORT_DETAILS))
985 fprintf (vect_dump, "=== vect_analyze_loop_form ===");
987 /* Different restrictions apply when we are considering an inner-most loop,
988 vs. an outer (nested) loop.
989 (FORNOW. May want to relax some of these restrictions in the future). */
993 /* Inner-most loop. We currently require that the number of BBs is
994 exactly 2 (the header and latch). Vectorizable inner-most loops
1005 if (loop->num_nodes != 2)
1007 if (vect_print_dump_info (REPORT_BAD_FORM_LOOPS))
1008 fprintf (vect_dump, "not vectorized: control flow in loop.");
1012 if (empty_block_p (loop->header))
1014 if (vect_print_dump_info (REPORT_BAD_FORM_LOOPS))
1015 fprintf (vect_dump, "not vectorized: empty loop.");
1021 struct loop *innerloop = loop->inner;
1024 /* Nested loop. We currently require that the loop is doubly-nested,
1025 contains a single inner loop, and the number of BBs is exactly 5.
1026 Vectorizable outer-loops look like this:
1038 The inner-loop has the properties expected of inner-most loops
1039 as described above. */
1041 if ((loop->inner)->inner || (loop->inner)->next)
1043 if (vect_print_dump_info (REPORT_BAD_FORM_LOOPS))
1044 fprintf (vect_dump, "not vectorized: multiple nested loops.");
1048 /* Analyze the inner-loop. */
1049 inner_loop_vinfo = vect_analyze_loop_1 (loop->inner);
1050 if (!inner_loop_vinfo)
1052 if (vect_print_dump_info (REPORT_BAD_FORM_LOOPS))
1053 fprintf (vect_dump, "not vectorized: Bad inner loop.");
1057 if (!expr_invariant_in_loop_p (loop,
1058 LOOP_VINFO_NITERS (inner_loop_vinfo)))
1060 if (vect_print_dump_info (REPORT_BAD_FORM_LOOPS))
1062 "not vectorized: inner-loop count not invariant.");
1063 destroy_loop_vec_info (inner_loop_vinfo, true);
1067 if (loop->num_nodes != 5)
1069 if (vect_print_dump_info (REPORT_BAD_FORM_LOOPS))
1070 fprintf (vect_dump, "not vectorized: control flow in loop.");
1071 destroy_loop_vec_info (inner_loop_vinfo, true);
1075 gcc_assert (EDGE_COUNT (innerloop->header->preds) == 2);
1076 entryedge = EDGE_PRED (innerloop->header, 0);
1077 if (EDGE_PRED (innerloop->header, 0)->src == innerloop->latch)
1078 entryedge = EDGE_PRED (innerloop->header, 1);
1080 if (entryedge->src != loop->header
1081 || !single_exit (innerloop)
1082 || single_exit (innerloop)->dest != EDGE_PRED (loop->latch, 0)->src)
1084 if (vect_print_dump_info (REPORT_BAD_FORM_LOOPS))
1085 fprintf (vect_dump, "not vectorized: unsupported outerloop form.");
1086 destroy_loop_vec_info (inner_loop_vinfo, true);
1090 if (vect_print_dump_info (REPORT_DETAILS))
1091 fprintf (vect_dump, "Considering outer-loop vectorization.");
1094 if (!single_exit (loop)
1095 || EDGE_COUNT (loop->header->preds) != 2)
1097 if (vect_print_dump_info (REPORT_BAD_FORM_LOOPS))
1099 if (!single_exit (loop))
1100 fprintf (vect_dump, "not vectorized: multiple exits.");
1101 else if (EDGE_COUNT (loop->header->preds) != 2)
1102 fprintf (vect_dump, "not vectorized: too many incoming edges.");
1104 if (inner_loop_vinfo)
1105 destroy_loop_vec_info (inner_loop_vinfo, true);
1109 /* We assume that the loop exit condition is at the end of the loop. i.e,
1110 that the loop is represented as a do-while (with a proper if-guard
1111 before the loop if needed), where the loop header contains all the
1112 executable statements, and the latch is empty. */
1113 if (!empty_block_p (loop->latch)
1114 || !gimple_seq_empty_p (phi_nodes (loop->latch)))
1116 if (vect_print_dump_info (REPORT_BAD_FORM_LOOPS))
1117 fprintf (vect_dump, "not vectorized: unexpected loop form.");
1118 if (inner_loop_vinfo)
1119 destroy_loop_vec_info (inner_loop_vinfo, true);
1123 /* Make sure there exists a single-predecessor exit bb: */
1124 if (!single_pred_p (single_exit (loop)->dest))
1126 edge e = single_exit (loop);
1127 if (!(e->flags & EDGE_ABNORMAL))
1129 split_loop_exit_edge (e);
1130 if (vect_print_dump_info (REPORT_DETAILS))
1131 fprintf (vect_dump, "split exit edge.");
1135 if (vect_print_dump_info (REPORT_BAD_FORM_LOOPS))
1136 fprintf (vect_dump, "not vectorized: abnormal loop exit edge.");
1137 if (inner_loop_vinfo)
1138 destroy_loop_vec_info (inner_loop_vinfo, true);
1143 loop_cond = vect_get_loop_niters (loop, &number_of_iterations);
1146 if (vect_print_dump_info (REPORT_BAD_FORM_LOOPS))
1147 fprintf (vect_dump, "not vectorized: complicated exit condition.");
1148 if (inner_loop_vinfo)
1149 destroy_loop_vec_info (inner_loop_vinfo, true);
1153 if (!number_of_iterations)
1155 if (vect_print_dump_info (REPORT_BAD_FORM_LOOPS))
1157 "not vectorized: number of iterations cannot be computed.");
1158 if (inner_loop_vinfo)
1159 destroy_loop_vec_info (inner_loop_vinfo, true);
1163 if (chrec_contains_undetermined (number_of_iterations))
1165 if (vect_print_dump_info (REPORT_BAD_FORM_LOOPS))
1166 fprintf (vect_dump, "Infinite number of iterations.");
1167 if (inner_loop_vinfo)
1168 destroy_loop_vec_info (inner_loop_vinfo, true);
1172 if (!NITERS_KNOWN_P (number_of_iterations))
1174 if (vect_print_dump_info (REPORT_DETAILS))
1176 fprintf (vect_dump, "Symbolic number of iterations is ");
1177 print_generic_expr (vect_dump, number_of_iterations, TDF_DETAILS);
1180 else if (TREE_INT_CST_LOW (number_of_iterations) == 0)
1182 if (vect_print_dump_info (REPORT_UNVECTORIZED_LOCATIONS))
1183 fprintf (vect_dump, "not vectorized: number of iterations = 0.");
1184 if (inner_loop_vinfo)
1185 destroy_loop_vec_info (inner_loop_vinfo, false);
1189 loop_vinfo = new_loop_vec_info (loop);
1190 LOOP_VINFO_NITERS (loop_vinfo) = number_of_iterations;
1191 LOOP_VINFO_NITERS_UNCHANGED (loop_vinfo) = number_of_iterations;
1193 STMT_VINFO_TYPE (vinfo_for_stmt (loop_cond)) = loop_exit_ctrl_vec_info_type;
1195 /* CHECKME: May want to keep it around it in the future. */
1196 if (inner_loop_vinfo)
1197 destroy_loop_vec_info (inner_loop_vinfo, false);
1199 gcc_assert (!loop->aux);
1200 loop->aux = loop_vinfo;
1205 /* Get cost by calling cost target builtin. */
1208 vect_get_cost (enum vect_cost_for_stmt type_of_cost)
1210 tree dummy_type = NULL;
1213 return targetm.vectorize.builtin_vectorization_cost (type_of_cost,
1218 /* Function vect_analyze_loop_operations.
1220 Scan the loop stmts and make sure they are all vectorizable. */
1223 vect_analyze_loop_operations (loop_vec_info loop_vinfo, bool slp)
1225 struct loop *loop = LOOP_VINFO_LOOP (loop_vinfo);
1226 basic_block *bbs = LOOP_VINFO_BBS (loop_vinfo);
1227 int nbbs = loop->num_nodes;
1228 gimple_stmt_iterator si;
1229 unsigned int vectorization_factor = 0;
1232 stmt_vec_info stmt_info;
1233 bool need_to_vectorize = false;
1234 int min_profitable_iters;
1235 int min_scalar_loop_bound;
1237 bool only_slp_in_loop = true, ok;
1239 if (vect_print_dump_info (REPORT_DETAILS))
1240 fprintf (vect_dump, "=== vect_analyze_loop_operations ===");
1242 gcc_assert (LOOP_VINFO_VECT_FACTOR (loop_vinfo));
1243 vectorization_factor = LOOP_VINFO_VECT_FACTOR (loop_vinfo);
1246 /* If all the stmts in the loop can be SLPed, we perform only SLP, and
1247 vectorization factor of the loop is the unrolling factor required by
1248 the SLP instances. If that unrolling factor is 1, we say, that we
1249 perform pure SLP on loop - cross iteration parallelism is not
1251 for (i = 0; i < nbbs; i++)
1253 basic_block bb = bbs[i];
1254 for (si = gsi_start_bb (bb); !gsi_end_p (si); gsi_next (&si))
1256 gimple stmt = gsi_stmt (si);
1257 stmt_vec_info stmt_info = vinfo_for_stmt (stmt);
1258 gcc_assert (stmt_info);
1259 if ((STMT_VINFO_RELEVANT_P (stmt_info)
1260 || VECTORIZABLE_CYCLE_DEF (STMT_VINFO_DEF_TYPE (stmt_info)))
1261 && !PURE_SLP_STMT (stmt_info))
1262 /* STMT needs both SLP and loop-based vectorization. */
1263 only_slp_in_loop = false;
1267 if (only_slp_in_loop)
1268 vectorization_factor = LOOP_VINFO_SLP_UNROLLING_FACTOR (loop_vinfo);
1270 vectorization_factor = least_common_multiple (vectorization_factor,
1271 LOOP_VINFO_SLP_UNROLLING_FACTOR (loop_vinfo));
1273 LOOP_VINFO_VECT_FACTOR (loop_vinfo) = vectorization_factor;
1274 if (vect_print_dump_info (REPORT_DETAILS))
1275 fprintf (vect_dump, "Updating vectorization factor to %d ",
1276 vectorization_factor);
1279 for (i = 0; i < nbbs; i++)
1281 basic_block bb = bbs[i];
1283 for (si = gsi_start_phis (bb); !gsi_end_p (si); gsi_next (&si))
1285 phi = gsi_stmt (si);
1288 stmt_info = vinfo_for_stmt (phi);
1289 if (vect_print_dump_info (REPORT_DETAILS))
1291 fprintf (vect_dump, "examining phi: ");
1292 print_gimple_stmt (vect_dump, phi, 0, TDF_SLIM);
1295 /* Inner-loop loop-closed exit phi in outer-loop vectorization
1296 (i.e., a phi in the tail of the outer-loop). */
1297 if (! is_loop_header_bb_p (bb))
1299 /* FORNOW: we currently don't support the case that these phis
1300 are not used in the outerloop (unless it is double reduction,
1301 i.e., this phi is vect_reduction_def), cause this case
1302 requires to actually do something here. */
1303 if ((!STMT_VINFO_RELEVANT_P (stmt_info)
1304 || STMT_VINFO_LIVE_P (stmt_info))
1305 && STMT_VINFO_DEF_TYPE (stmt_info)
1306 != vect_double_reduction_def)
1308 if (vect_print_dump_info (REPORT_DETAILS))
1310 "Unsupported loop-closed phi in outer-loop.");
1314 /* If PHI is used in the outer loop, we check that its operand
1315 is defined in the inner loop. */
1316 if (STMT_VINFO_RELEVANT_P (stmt_info))
1321 if (gimple_phi_num_args (phi) != 1)
1324 phi_op = PHI_ARG_DEF (phi, 0);
1325 if (TREE_CODE (phi_op) != SSA_NAME)
1328 op_def_stmt = SSA_NAME_DEF_STMT (phi_op);
1330 || !flow_bb_inside_loop_p (loop, gimple_bb (op_def_stmt))
1331 || !vinfo_for_stmt (op_def_stmt))
1334 if (STMT_VINFO_RELEVANT (vinfo_for_stmt (op_def_stmt))
1335 != vect_used_in_outer
1336 && STMT_VINFO_RELEVANT (vinfo_for_stmt (op_def_stmt))
1337 != vect_used_in_outer_by_reduction)
1344 gcc_assert (stmt_info);
1346 if (STMT_VINFO_LIVE_P (stmt_info))
1348 /* FORNOW: not yet supported. */
1349 if (vect_print_dump_info (REPORT_UNVECTORIZED_LOCATIONS))
1350 fprintf (vect_dump, "not vectorized: value used after loop.");
1354 if (STMT_VINFO_RELEVANT (stmt_info) == vect_used_in_scope
1355 && STMT_VINFO_DEF_TYPE (stmt_info) != vect_induction_def)
1357 /* A scalar-dependence cycle that we don't support. */
1358 if (vect_print_dump_info (REPORT_UNVECTORIZED_LOCATIONS))
1359 fprintf (vect_dump, "not vectorized: scalar dependence cycle.");
1363 if (STMT_VINFO_RELEVANT_P (stmt_info))
1365 need_to_vectorize = true;
1366 if (STMT_VINFO_DEF_TYPE (stmt_info) == vect_induction_def)
1367 ok = vectorizable_induction (phi, NULL, NULL);
1372 if (vect_print_dump_info (REPORT_UNVECTORIZED_LOCATIONS))
1375 "not vectorized: relevant phi not supported: ");
1376 print_gimple_stmt (vect_dump, phi, 0, TDF_SLIM);
1382 for (si = gsi_start_bb (bb); !gsi_end_p (si); gsi_next (&si))
1384 gimple stmt = gsi_stmt (si);
1385 if (!vect_analyze_stmt (stmt, &need_to_vectorize, NULL))
1390 /* All operations in the loop are either irrelevant (deal with loop
1391 control, or dead), or only used outside the loop and can be moved
1392 out of the loop (e.g. invariants, inductions). The loop can be
1393 optimized away by scalar optimizations. We're better off not
1394 touching this loop. */
1395 if (!need_to_vectorize)
1397 if (vect_print_dump_info (REPORT_DETAILS))
1399 "All the computation can be taken out of the loop.");
1400 if (vect_print_dump_info (REPORT_UNVECTORIZED_LOCATIONS))
1402 "not vectorized: redundant loop. no profit to vectorize.");
1406 if (LOOP_VINFO_NITERS_KNOWN_P (loop_vinfo)
1407 && vect_print_dump_info (REPORT_DETAILS))
1409 "vectorization_factor = %d, niters = " HOST_WIDE_INT_PRINT_DEC,
1410 vectorization_factor, LOOP_VINFO_INT_NITERS (loop_vinfo));
1412 if (LOOP_VINFO_NITERS_KNOWN_P (loop_vinfo)
1413 && (LOOP_VINFO_INT_NITERS (loop_vinfo) < vectorization_factor))
1415 if (vect_print_dump_info (REPORT_UNVECTORIZED_LOCATIONS))
1416 fprintf (vect_dump, "not vectorized: iteration count too small.");
1417 if (vect_print_dump_info (REPORT_DETAILS))
1418 fprintf (vect_dump,"not vectorized: iteration count smaller than "
1419 "vectorization factor.");
1423 /* Analyze cost. Decide if worth while to vectorize. */
1425 /* Once VF is set, SLP costs should be updated since the number of created
1426 vector stmts depends on VF. */
1427 vect_update_slp_costs_according_to_vf (loop_vinfo);
1429 min_profitable_iters = vect_estimate_min_profitable_iters (loop_vinfo);
1430 LOOP_VINFO_COST_MODEL_MIN_ITERS (loop_vinfo) = min_profitable_iters;
1432 if (min_profitable_iters < 0)
1434 if (vect_print_dump_info (REPORT_UNVECTORIZED_LOCATIONS))
1435 fprintf (vect_dump, "not vectorized: vectorization not profitable.");
1436 if (vect_print_dump_info (REPORT_DETAILS))
1437 fprintf (vect_dump, "not vectorized: vector version will never be "
1442 min_scalar_loop_bound = ((PARAM_VALUE (PARAM_MIN_VECT_LOOP_BOUND)
1443 * vectorization_factor) - 1);
1445 /* Use the cost model only if it is more conservative than user specified
1448 th = (unsigned) min_scalar_loop_bound;
1449 if (min_profitable_iters
1450 && (!min_scalar_loop_bound
1451 || min_profitable_iters > min_scalar_loop_bound))
1452 th = (unsigned) min_profitable_iters;
1454 if (LOOP_VINFO_NITERS_KNOWN_P (loop_vinfo)
1455 && LOOP_VINFO_INT_NITERS (loop_vinfo) <= th)
1457 if (vect_print_dump_info (REPORT_UNVECTORIZED_LOCATIONS))
1458 fprintf (vect_dump, "not vectorized: vectorization not "
1460 if (vect_print_dump_info (REPORT_DETAILS))
1461 fprintf (vect_dump, "not vectorized: iteration count smaller than "
1462 "user specified loop bound parameter or minimum "
1463 "profitable iterations (whichever is more conservative).");
1467 if (!LOOP_VINFO_NITERS_KNOWN_P (loop_vinfo)
1468 || LOOP_VINFO_INT_NITERS (loop_vinfo) % vectorization_factor != 0
1469 || LOOP_VINFO_PEELING_FOR_GAPS (loop_vinfo))
1471 if (vect_print_dump_info (REPORT_DETAILS))
1472 fprintf (vect_dump, "epilog loop required.");
1473 if (!vect_can_advance_ivs_p (loop_vinfo))
1475 if (vect_print_dump_info (REPORT_UNVECTORIZED_LOCATIONS))
1477 "not vectorized: can't create epilog loop 1.");
1480 if (!slpeel_can_duplicate_loop_p (loop, single_exit (loop)))
1482 if (vect_print_dump_info (REPORT_UNVECTORIZED_LOCATIONS))
1484 "not vectorized: can't create epilog loop 2.");
1493 /* Function vect_analyze_loop_2.
1495 Apply a set of analyses on LOOP, and create a loop_vec_info struct
1496 for it. The different analyses will record information in the
1497 loop_vec_info struct. */
1499 vect_analyze_loop_2 (loop_vec_info loop_vinfo)
1501 bool ok, slp = false;
1502 int max_vf = MAX_VECTORIZATION_FACTOR;
1505 /* Find all data references in the loop (which correspond to vdefs/vuses)
1506 and analyze their evolution in the loop. Also adjust the minimal
1507 vectorization factor according to the loads and stores.
1509 FORNOW: Handle only simple, array references, which
1510 alignment can be forced, and aligned pointer-references. */
1512 ok = vect_analyze_data_refs (loop_vinfo, NULL, &min_vf);
1515 if (vect_print_dump_info (REPORT_DETAILS))
1516 fprintf (vect_dump, "bad data references.");
1520 /* Classify all cross-iteration scalar data-flow cycles.
1521 Cross-iteration cycles caused by virtual phis are analyzed separately. */
1523 vect_analyze_scalar_cycles (loop_vinfo);
1525 vect_pattern_recog (loop_vinfo);
1527 /* Data-flow analysis to detect stmts that do not need to be vectorized. */
1529 ok = vect_mark_stmts_to_be_vectorized (loop_vinfo);
1532 if (vect_print_dump_info (REPORT_DETAILS))
1533 fprintf (vect_dump, "unexpected pattern.");
1537 /* Analyze data dependences between the data-refs in the loop
1538 and adjust the maximum vectorization factor according to
1540 FORNOW: fail at the first data dependence that we encounter. */
1542 ok = vect_analyze_data_ref_dependences (loop_vinfo, NULL, &max_vf);
1546 if (vect_print_dump_info (REPORT_DETAILS))
1547 fprintf (vect_dump, "bad data dependence.");
1551 ok = vect_determine_vectorization_factor (loop_vinfo);
1554 if (vect_print_dump_info (REPORT_DETAILS))
1555 fprintf (vect_dump, "can't determine vectorization factor.");
1558 if (max_vf < LOOP_VINFO_VECT_FACTOR (loop_vinfo))
1560 if (vect_print_dump_info (REPORT_DETAILS))
1561 fprintf (vect_dump, "bad data dependence.");
1565 /* Analyze the alignment of the data-refs in the loop.
1566 Fail if a data reference is found that cannot be vectorized. */
1568 ok = vect_analyze_data_refs_alignment (loop_vinfo, NULL);
1571 if (vect_print_dump_info (REPORT_DETAILS))
1572 fprintf (vect_dump, "bad data alignment.");
1576 /* Analyze the access patterns of the data-refs in the loop (consecutive,
1577 complex, etc.). FORNOW: Only handle consecutive access pattern. */
1579 ok = vect_analyze_data_ref_accesses (loop_vinfo, NULL);
1582 if (vect_print_dump_info (REPORT_DETAILS))
1583 fprintf (vect_dump, "bad data access.");
1587 /* Prune the list of ddrs to be tested at run-time by versioning for alias.
1588 It is important to call pruning after vect_analyze_data_ref_accesses,
1589 since we use grouping information gathered by interleaving analysis. */
1590 ok = vect_prune_runtime_alias_test_list (loop_vinfo);
1593 if (vect_print_dump_info (REPORT_DETAILS))
1594 fprintf (vect_dump, "too long list of versioning for alias "
1599 /* This pass will decide on using loop versioning and/or loop peeling in
1600 order to enhance the alignment of data references in the loop. */
1602 ok = vect_enhance_data_refs_alignment (loop_vinfo);
1605 if (vect_print_dump_info (REPORT_DETAILS))
1606 fprintf (vect_dump, "bad data alignment.");
1610 /* Check the SLP opportunities in the loop, analyze and build SLP trees. */
1611 ok = vect_analyze_slp (loop_vinfo, NULL);
1614 /* Decide which possible SLP instances to SLP. */
1615 slp = vect_make_slp_decision (loop_vinfo);
1617 /* Find stmts that need to be both vectorized and SLPed. */
1618 vect_detect_hybrid_slp (loop_vinfo);
1623 /* Scan all the operations in the loop and make sure they are
1626 ok = vect_analyze_loop_operations (loop_vinfo, slp);
1629 if (vect_print_dump_info (REPORT_DETAILS))
1630 fprintf (vect_dump, "bad operation or unsupported loop bound.");
1637 /* Function vect_analyze_loop.
1639 Apply a set of analyses on LOOP, and create a loop_vec_info struct
1640 for it. The different analyses will record information in the
1641 loop_vec_info struct. */
1643 vect_analyze_loop (struct loop *loop)
1645 loop_vec_info loop_vinfo;
1646 unsigned int vector_sizes;
1648 /* Autodetect first vector size we try. */
1649 current_vector_size = 0;
1650 vector_sizes = targetm.vectorize.autovectorize_vector_sizes ();
1652 if (vect_print_dump_info (REPORT_DETAILS))
1653 fprintf (vect_dump, "===== analyze_loop_nest =====");
1655 if (loop_outer (loop)
1656 && loop_vec_info_for_loop (loop_outer (loop))
1657 && LOOP_VINFO_VECTORIZABLE_P (loop_vec_info_for_loop (loop_outer (loop))))
1659 if (vect_print_dump_info (REPORT_DETAILS))
1660 fprintf (vect_dump, "outer-loop already vectorized.");
1666 /* Check the CFG characteristics of the loop (nesting, entry/exit). */
1667 loop_vinfo = vect_analyze_loop_form (loop);
1670 if (vect_print_dump_info (REPORT_DETAILS))
1671 fprintf (vect_dump, "bad loop form.");
1675 if (vect_analyze_loop_2 (loop_vinfo))
1677 LOOP_VINFO_VECTORIZABLE_P (loop_vinfo) = 1;
1682 destroy_loop_vec_info (loop_vinfo, true);
1684 vector_sizes &= ~current_vector_size;
1685 if (vector_sizes == 0
1686 || current_vector_size == 0)
1689 /* Try the next biggest vector size. */
1690 current_vector_size = 1 << floor_log2 (vector_sizes);
1691 if (vect_print_dump_info (REPORT_DETAILS))
1692 fprintf (vect_dump, "***** Re-trying analysis with "
1693 "vector size %d\n", current_vector_size);
1698 /* Function reduction_code_for_scalar_code
1701 CODE - tree_code of a reduction operations.
1704 REDUC_CODE - the corresponding tree-code to be used to reduce the
1705 vector of partial results into a single scalar result (which
1706 will also reside in a vector) or ERROR_MARK if the operation is
1707 a supported reduction operation, but does not have such tree-code.
1709 Return FALSE if CODE currently cannot be vectorized as reduction. */
1712 reduction_code_for_scalar_code (enum tree_code code,
1713 enum tree_code *reduc_code)
1718 *reduc_code = REDUC_MAX_EXPR;
1722 *reduc_code = REDUC_MIN_EXPR;
1726 *reduc_code = REDUC_PLUS_EXPR;
1734 *reduc_code = ERROR_MARK;
1743 /* Error reporting helper for vect_is_simple_reduction below. GIMPLE statement
1744 STMT is printed with a message MSG. */
1747 report_vect_op (gimple stmt, const char *msg)
1749 fprintf (vect_dump, "%s", msg);
1750 print_gimple_stmt (vect_dump, stmt, 0, TDF_SLIM);
1754 /* Detect SLP reduction of the form:
1764 PHI is the reduction phi node (#a1 = phi <a5, a0> above)
1765 FIRST_STMT is the first reduction stmt in the chain
1766 (a2 = operation (a1)).
1768 Return TRUE if a reduction chain was detected. */
1771 vect_is_slp_reduction (loop_vec_info loop_info, gimple phi, gimple first_stmt)
1773 struct loop *loop = (gimple_bb (phi))->loop_father;
1774 struct loop *vect_loop = LOOP_VINFO_LOOP (loop_info);
1775 enum tree_code code;
1776 gimple current_stmt = NULL, loop_use_stmt = NULL, first, next_stmt;
1777 stmt_vec_info use_stmt_info, current_stmt_info;
1779 imm_use_iterator imm_iter;
1780 use_operand_p use_p;
1781 int nloop_uses, size = 0, n_out_of_loop_uses;
1784 if (loop != vect_loop)
1787 lhs = PHI_RESULT (phi);
1788 code = gimple_assign_rhs_code (first_stmt);
1792 n_out_of_loop_uses = 0;
1793 FOR_EACH_IMM_USE_FAST (use_p, imm_iter, lhs)
1795 gimple use_stmt = USE_STMT (use_p);
1796 if (is_gimple_debug (use_stmt))
1799 use_stmt = USE_STMT (use_p);
1801 /* Check if we got back to the reduction phi. */
1802 if (use_stmt == phi)
1804 loop_use_stmt = use_stmt;
1809 if (flow_bb_inside_loop_p (loop, gimple_bb (use_stmt)))
1811 if (vinfo_for_stmt (use_stmt)
1812 && !STMT_VINFO_IN_PATTERN_P (vinfo_for_stmt (use_stmt)))
1814 loop_use_stmt = use_stmt;
1819 n_out_of_loop_uses++;
1821 /* There are can be either a single use in the loop or two uses in
1823 if (nloop_uses > 1 || (n_out_of_loop_uses && nloop_uses))
1830 /* We reached a statement with no loop uses. */
1831 if (nloop_uses == 0)
1834 /* This is a loop exit phi, and we haven't reached the reduction phi. */
1835 if (gimple_code (loop_use_stmt) == GIMPLE_PHI)
1838 if (!is_gimple_assign (loop_use_stmt)
1839 || code != gimple_assign_rhs_code (loop_use_stmt)
1840 || !flow_bb_inside_loop_p (loop, gimple_bb (loop_use_stmt)))
1843 /* Insert USE_STMT into reduction chain. */
1844 use_stmt_info = vinfo_for_stmt (loop_use_stmt);
1847 current_stmt_info = vinfo_for_stmt (current_stmt);
1848 GROUP_NEXT_ELEMENT (current_stmt_info) = loop_use_stmt;
1849 GROUP_FIRST_ELEMENT (use_stmt_info)
1850 = GROUP_FIRST_ELEMENT (current_stmt_info);
1853 GROUP_FIRST_ELEMENT (use_stmt_info) = loop_use_stmt;
1855 lhs = gimple_assign_lhs (loop_use_stmt);
1856 current_stmt = loop_use_stmt;
1860 if (!found || loop_use_stmt != phi || size < 2)
1863 /* Swap the operands, if needed, to make the reduction operand be the second
1865 lhs = PHI_RESULT (phi);
1866 next_stmt = GROUP_FIRST_ELEMENT (vinfo_for_stmt (current_stmt));
1869 if (gimple_assign_rhs2 (next_stmt) == lhs)
1871 tree op = gimple_assign_rhs1 (next_stmt);
1872 gimple def_stmt = NULL;
1874 if (TREE_CODE (op) == SSA_NAME)
1875 def_stmt = SSA_NAME_DEF_STMT (op);
1877 /* Check that the other def is either defined in the loop
1878 ("vect_internal_def"), or it's an induction (defined by a
1879 loop-header phi-node). */
1881 && gimple_bb (def_stmt)
1882 && flow_bb_inside_loop_p (loop, gimple_bb (def_stmt))
1883 && (is_gimple_assign (def_stmt)
1884 || is_gimple_call (def_stmt)
1885 || STMT_VINFO_DEF_TYPE (vinfo_for_stmt (def_stmt))
1886 == vect_induction_def
1887 || (gimple_code (def_stmt) == GIMPLE_PHI
1888 && STMT_VINFO_DEF_TYPE (vinfo_for_stmt (def_stmt))
1889 == vect_internal_def
1890 && !is_loop_header_bb_p (gimple_bb (def_stmt)))))
1892 lhs = gimple_assign_lhs (next_stmt);
1893 next_stmt = GROUP_NEXT_ELEMENT (vinfo_for_stmt (next_stmt));
1901 tree op = gimple_assign_rhs2 (next_stmt);
1902 gimple def_stmt = NULL;
1904 if (TREE_CODE (op) == SSA_NAME)
1905 def_stmt = SSA_NAME_DEF_STMT (op);
1907 /* Check that the other def is either defined in the loop
1908 ("vect_internal_def"), or it's an induction (defined by a
1909 loop-header phi-node). */
1911 && gimple_bb (def_stmt)
1912 && flow_bb_inside_loop_p (loop, gimple_bb (def_stmt))
1913 && (is_gimple_assign (def_stmt)
1914 || is_gimple_call (def_stmt)
1915 || STMT_VINFO_DEF_TYPE (vinfo_for_stmt (def_stmt))
1916 == vect_induction_def
1917 || (gimple_code (def_stmt) == GIMPLE_PHI
1918 && STMT_VINFO_DEF_TYPE (vinfo_for_stmt (def_stmt))
1919 == vect_internal_def
1920 && !is_loop_header_bb_p (gimple_bb (def_stmt)))))
1922 if (vect_print_dump_info (REPORT_DETAILS))
1924 fprintf (vect_dump, "swapping oprnds: ");
1925 print_gimple_stmt (vect_dump, next_stmt, 0, TDF_SLIM);
1928 swap_tree_operands (next_stmt,
1929 gimple_assign_rhs1_ptr (next_stmt),
1930 gimple_assign_rhs2_ptr (next_stmt));
1931 mark_symbols_for_renaming (next_stmt);
1937 lhs = gimple_assign_lhs (next_stmt);
1938 next_stmt = GROUP_NEXT_ELEMENT (vinfo_for_stmt (next_stmt));
1941 /* Save the chain for further analysis in SLP detection. */
1942 first = GROUP_FIRST_ELEMENT (vinfo_for_stmt (current_stmt));
1943 VEC_safe_push (gimple, heap, LOOP_VINFO_REDUCTION_CHAINS (loop_info), first);
1944 GROUP_SIZE (vinfo_for_stmt (first)) = size;
1950 /* Function vect_is_simple_reduction_1
1952 (1) Detect a cross-iteration def-use cycle that represents a simple
1953 reduction computation. We look for the following pattern:
1958 a2 = operation (a3, a1)
1961 1. operation is commutative and associative and it is safe to
1962 change the order of the computation (if CHECK_REDUCTION is true)
1963 2. no uses for a2 in the loop (a2 is used out of the loop)
1964 3. no uses of a1 in the loop besides the reduction operation
1965 4. no uses of a1 outside the loop.
1967 Conditions 1,4 are tested here.
1968 Conditions 2,3 are tested in vect_mark_stmts_to_be_vectorized.
1970 (2) Detect a cross-iteration def-use cycle in nested loops, i.e.,
1971 nested cycles, if CHECK_REDUCTION is false.
1973 (3) Detect cycles of phi nodes in outer-loop vectorization, i.e., double
1977 inner loop (def of a3)
1980 If MODIFY is true it tries also to rework the code in-place to enable
1981 detection of more reduction patterns. For the time being we rewrite
1982 "res -= RHS" into "rhs += -RHS" when it seems worthwhile.
1986 vect_is_simple_reduction_1 (loop_vec_info loop_info, gimple phi,
1987 bool check_reduction, bool *double_reduc,
1990 struct loop *loop = (gimple_bb (phi))->loop_father;
1991 struct loop *vect_loop = LOOP_VINFO_LOOP (loop_info);
1992 edge latch_e = loop_latch_edge (loop);
1993 tree loop_arg = PHI_ARG_DEF_FROM_EDGE (phi, latch_e);
1994 gimple def_stmt, def1 = NULL, def2 = NULL;
1995 enum tree_code orig_code, code;
1996 tree op1, op2, op3 = NULL_TREE, op4 = NULL_TREE;
2000 imm_use_iterator imm_iter;
2001 use_operand_p use_p;
2004 *double_reduc = false;
2006 /* If CHECK_REDUCTION is true, we assume inner-most loop vectorization,
2007 otherwise, we assume outer loop vectorization. */
2008 gcc_assert ((check_reduction && loop == vect_loop)
2009 || (!check_reduction && flow_loop_nested_p (vect_loop, loop)));
2011 name = PHI_RESULT (phi);
2012 /* ??? If there are no uses of the PHI result the inner loop reduction
2013 won't be detected as possibly double-reduction by vectorizable_reduction
2014 because that tries to walk the PHI arg from the preheader edge which
2015 can be constant. See PR60382. */
2016 if (has_zero_uses (name))
2019 FOR_EACH_IMM_USE_FAST (use_p, imm_iter, name)
2021 gimple use_stmt = USE_STMT (use_p);
2022 if (is_gimple_debug (use_stmt))
2025 if (!flow_bb_inside_loop_p (loop, gimple_bb (use_stmt)))
2027 if (vect_print_dump_info (REPORT_DETAILS))
2028 fprintf (vect_dump, "intermediate value used outside loop.");
2033 if (vinfo_for_stmt (use_stmt)
2034 && !is_pattern_stmt_p (vinfo_for_stmt (use_stmt)))
2038 if (vect_print_dump_info (REPORT_DETAILS))
2039 fprintf (vect_dump, "reduction used in loop.");
2044 if (TREE_CODE (loop_arg) != SSA_NAME)
2046 if (vect_print_dump_info (REPORT_DETAILS))
2048 fprintf (vect_dump, "reduction: not ssa_name: ");
2049 print_generic_expr (vect_dump, loop_arg, TDF_SLIM);
2054 def_stmt = SSA_NAME_DEF_STMT (loop_arg);
2057 if (vect_print_dump_info (REPORT_DETAILS))
2058 fprintf (vect_dump, "reduction: no def_stmt.");
2062 if (!is_gimple_assign (def_stmt) && gimple_code (def_stmt) != GIMPLE_PHI)
2064 if (vect_print_dump_info (REPORT_DETAILS))
2065 print_gimple_stmt (vect_dump, def_stmt, 0, TDF_SLIM);
2069 if (is_gimple_assign (def_stmt))
2071 name = gimple_assign_lhs (def_stmt);
2076 name = PHI_RESULT (def_stmt);
2081 FOR_EACH_IMM_USE_FAST (use_p, imm_iter, name)
2083 gimple use_stmt = USE_STMT (use_p);
2084 if (is_gimple_debug (use_stmt))
2086 if (flow_bb_inside_loop_p (loop, gimple_bb (use_stmt))
2087 && vinfo_for_stmt (use_stmt)
2088 && !is_pattern_stmt_p (vinfo_for_stmt (use_stmt)))
2092 if (vect_print_dump_info (REPORT_DETAILS))
2093 fprintf (vect_dump, "reduction used in loop.");
2098 /* If DEF_STMT is a phi node itself, we expect it to have a single argument
2099 defined in the inner loop. */
2102 op1 = PHI_ARG_DEF (def_stmt, 0);
2104 if (gimple_phi_num_args (def_stmt) != 1
2105 || TREE_CODE (op1) != SSA_NAME)
2107 if (vect_print_dump_info (REPORT_DETAILS))
2108 fprintf (vect_dump, "unsupported phi node definition.");
2113 def1 = SSA_NAME_DEF_STMT (op1);
2114 if (flow_bb_inside_loop_p (loop, gimple_bb (def_stmt))
2116 && flow_bb_inside_loop_p (loop->inner, gimple_bb (def1))
2117 && is_gimple_assign (def1))
2119 if (vect_print_dump_info (REPORT_DETAILS))
2120 report_vect_op (def_stmt, "detected double reduction: ");
2122 *double_reduc = true;
2129 code = orig_code = gimple_assign_rhs_code (def_stmt);
2131 /* We can handle "res -= x[i]", which is non-associative by
2132 simply rewriting this into "res += -x[i]". Avoid changing
2133 gimple instruction for the first simple tests and only do this
2134 if we're allowed to change code at all. */
2135 if (code == MINUS_EXPR
2137 && (op1 = gimple_assign_rhs1 (def_stmt))
2138 && TREE_CODE (op1) == SSA_NAME
2139 && SSA_NAME_DEF_STMT (op1) == phi)
2143 && (!commutative_tree_code (code) || !associative_tree_code (code)))
2145 if (vect_print_dump_info (REPORT_DETAILS))
2146 report_vect_op (def_stmt, "reduction: not commutative/associative: ");
2150 if (get_gimple_rhs_class (code) != GIMPLE_BINARY_RHS)
2152 if (code != COND_EXPR)
2154 if (vect_print_dump_info (REPORT_DETAILS))
2155 report_vect_op (def_stmt, "reduction: not binary operation: ");
2160 op3 = gimple_assign_rhs1 (def_stmt);
2161 if (COMPARISON_CLASS_P (op3))
2163 op4 = TREE_OPERAND (op3, 1);
2164 op3 = TREE_OPERAND (op3, 0);
2167 op1 = gimple_assign_rhs2 (def_stmt);
2168 op2 = gimple_assign_rhs3 (def_stmt);
2170 if (TREE_CODE (op1) != SSA_NAME && TREE_CODE (op2) != SSA_NAME)
2172 if (vect_print_dump_info (REPORT_DETAILS))
2173 report_vect_op (def_stmt, "reduction: uses not ssa_names: ");
2180 op1 = gimple_assign_rhs1 (def_stmt);
2181 op2 = gimple_assign_rhs2 (def_stmt);
2183 if (TREE_CODE (op1) != SSA_NAME && TREE_CODE (op2) != SSA_NAME)
2185 if (vect_print_dump_info (REPORT_DETAILS))
2186 report_vect_op (def_stmt, "reduction: uses not ssa_names: ");
2192 type = TREE_TYPE (gimple_assign_lhs (def_stmt));
2193 if ((TREE_CODE (op1) == SSA_NAME
2194 && !types_compatible_p (type,TREE_TYPE (op1)))
2195 || (TREE_CODE (op2) == SSA_NAME
2196 && !types_compatible_p (type, TREE_TYPE (op2)))
2197 || (op3 && TREE_CODE (op3) == SSA_NAME
2198 && !types_compatible_p (type, TREE_TYPE (op3)))
2199 || (op4 && TREE_CODE (op4) == SSA_NAME
2200 && !types_compatible_p (type, TREE_TYPE (op4))))
2202 if (vect_print_dump_info (REPORT_DETAILS))
2204 fprintf (vect_dump, "reduction: multiple types: operation type: ");
2205 print_generic_expr (vect_dump, type, TDF_SLIM);
2206 fprintf (vect_dump, ", operands types: ");
2207 print_generic_expr (vect_dump, TREE_TYPE (op1), TDF_SLIM);
2208 fprintf (vect_dump, ",");
2209 print_generic_expr (vect_dump, TREE_TYPE (op2), TDF_SLIM);
2212 fprintf (vect_dump, ",");
2213 print_generic_expr (vect_dump, TREE_TYPE (op3), TDF_SLIM);
2218 fprintf (vect_dump, ",");
2219 print_generic_expr (vect_dump, TREE_TYPE (op4), TDF_SLIM);
2226 /* Check that it's ok to change the order of the computation.
2227 Generally, when vectorizing a reduction we change the order of the
2228 computation. This may change the behavior of the program in some
2229 cases, so we need to check that this is ok. One exception is when
2230 vectorizing an outer-loop: the inner-loop is executed sequentially,
2231 and therefore vectorizing reductions in the inner-loop during
2232 outer-loop vectorization is safe. */
2234 /* CHECKME: check for !flag_finite_math_only too? */
2235 if (SCALAR_FLOAT_TYPE_P (type) && !flag_associative_math
2238 /* Changing the order of operations changes the semantics. */
2239 if (vect_print_dump_info (REPORT_DETAILS))
2240 report_vect_op (def_stmt, "reduction: unsafe fp math optimization: ");
2243 else if (INTEGRAL_TYPE_P (type) && TYPE_OVERFLOW_TRAPS (type)
2246 /* Changing the order of operations changes the semantics. */
2247 if (vect_print_dump_info (REPORT_DETAILS))
2248 report_vect_op (def_stmt, "reduction: unsafe int math optimization: ");
2251 else if (SAT_FIXED_POINT_TYPE_P (type) && check_reduction)
2253 /* Changing the order of operations changes the semantics. */
2254 if (vect_print_dump_info (REPORT_DETAILS))
2255 report_vect_op (def_stmt,
2256 "reduction: unsafe fixed-point math optimization: ");
2260 /* If we detected "res -= x[i]" earlier, rewrite it into
2261 "res += -x[i]" now. If this turns out to be useless reassoc
2262 will clean it up again. */
2263 if (orig_code == MINUS_EXPR)
2265 tree rhs = gimple_assign_rhs2 (def_stmt);
2266 tree var = TREE_CODE (rhs) == SSA_NAME
2267 ? SSA_NAME_VAR (rhs)
2268 : create_tmp_reg (TREE_TYPE (rhs), NULL);
2269 tree negrhs = make_ssa_name (var, NULL);
2270 gimple negate_stmt = gimple_build_assign_with_ops (NEGATE_EXPR, negrhs,
2272 gimple_stmt_iterator gsi = gsi_for_stmt (def_stmt);
2273 set_vinfo_for_stmt (negate_stmt, new_stmt_vec_info (negate_stmt,
2275 gsi_insert_before (&gsi, negate_stmt, GSI_NEW_STMT);
2276 gimple_assign_set_rhs2 (def_stmt, negrhs);
2277 gimple_assign_set_rhs_code (def_stmt, PLUS_EXPR);
2278 update_stmt (def_stmt);
2281 /* Reduction is safe. We're dealing with one of the following:
2282 1) integer arithmetic and no trapv
2283 2) floating point arithmetic, and special flags permit this optimization
2284 3) nested cycle (i.e., outer loop vectorization). */
2285 if (TREE_CODE (op1) == SSA_NAME)
2286 def1 = SSA_NAME_DEF_STMT (op1);
2288 if (TREE_CODE (op2) == SSA_NAME)
2289 def2 = SSA_NAME_DEF_STMT (op2);
2291 if (code != COND_EXPR
2292 && ((!def1 || gimple_nop_p (def1)) && (!def2 || gimple_nop_p (def2))))
2294 if (vect_print_dump_info (REPORT_DETAILS))
2295 report_vect_op (def_stmt, "reduction: no defs for operands: ");
2299 /* Check that one def is the reduction def, defined by PHI,
2300 the other def is either defined in the loop ("vect_internal_def"),
2301 or it's an induction (defined by a loop-header phi-node). */
2303 if (def2 && def2 == phi
2304 && (code == COND_EXPR
2305 || !def1 || gimple_nop_p (def1)
2306 || (def1 && flow_bb_inside_loop_p (loop, gimple_bb (def1))
2307 && (is_gimple_assign (def1)
2308 || is_gimple_call (def1)
2309 || STMT_VINFO_DEF_TYPE (vinfo_for_stmt (def1))
2310 == vect_induction_def
2311 || (gimple_code (def1) == GIMPLE_PHI
2312 && STMT_VINFO_DEF_TYPE (vinfo_for_stmt (def1))
2313 == vect_internal_def
2314 && !is_loop_header_bb_p (gimple_bb (def1)))))))
2316 if (vect_print_dump_info (REPORT_DETAILS))
2317 report_vect_op (def_stmt, "detected reduction: ");
2321 if (def1 && def1 == phi
2322 && (code == COND_EXPR
2323 || !def2 || gimple_nop_p (def2)
2324 || (def2 && flow_bb_inside_loop_p (loop, gimple_bb (def2))
2325 && (is_gimple_assign (def2)
2326 || is_gimple_call (def2)
2327 || STMT_VINFO_DEF_TYPE (vinfo_for_stmt (def2))
2328 == vect_induction_def
2329 || (gimple_code (def2) == GIMPLE_PHI
2330 && STMT_VINFO_DEF_TYPE (vinfo_for_stmt (def2))
2331 == vect_internal_def
2332 && !is_loop_header_bb_p (gimple_bb (def2)))))))
2334 if (check_reduction)
2336 /* Swap operands (just for simplicity - so that the rest of the code
2337 can assume that the reduction variable is always the last (second)
2339 if (vect_print_dump_info (REPORT_DETAILS))
2340 report_vect_op (def_stmt,
2341 "detected reduction: need to swap operands: ");
2343 swap_tree_operands (def_stmt, gimple_assign_rhs1_ptr (def_stmt),
2344 gimple_assign_rhs2_ptr (def_stmt));
2348 if (vect_print_dump_info (REPORT_DETAILS))
2349 report_vect_op (def_stmt, "detected reduction: ");
2355 /* Try to find SLP reduction chain. */
2356 if (check_reduction && vect_is_slp_reduction (loop_info, phi, def_stmt))
2358 if (vect_print_dump_info (REPORT_DETAILS))
2359 report_vect_op (def_stmt, "reduction: detected reduction chain: ");
2364 if (vect_print_dump_info (REPORT_DETAILS))
2365 report_vect_op (def_stmt, "reduction: unknown pattern: ");
2370 /* Wrapper around vect_is_simple_reduction_1, that won't modify code
2371 in-place. Arguments as there. */
2374 vect_is_simple_reduction (loop_vec_info loop_info, gimple phi,
2375 bool check_reduction, bool *double_reduc)
2377 return vect_is_simple_reduction_1 (loop_info, phi, check_reduction,
2378 double_reduc, false);
2381 /* Wrapper around vect_is_simple_reduction_1, which will modify code
2382 in-place if it enables detection of more reductions. Arguments
2386 vect_force_simple_reduction (loop_vec_info loop_info, gimple phi,
2387 bool check_reduction, bool *double_reduc)
2389 return vect_is_simple_reduction_1 (loop_info, phi, check_reduction,
2390 double_reduc, true);
2393 /* Calculate the cost of one scalar iteration of the loop. */
2395 vect_get_single_scalar_iteration_cost (loop_vec_info loop_vinfo)
2397 struct loop *loop = LOOP_VINFO_LOOP (loop_vinfo);
2398 basic_block *bbs = LOOP_VINFO_BBS (loop_vinfo);
2399 int nbbs = loop->num_nodes, factor, scalar_single_iter_cost = 0;
2400 int innerloop_iters, i, stmt_cost;
2402 /* Count statements in scalar loop. Using this as scalar cost for a single
2405 TODO: Add outer loop support.
2407 TODO: Consider assigning different costs to different scalar
2411 innerloop_iters = 1;
2413 innerloop_iters = 50; /* FIXME */
2415 for (i = 0; i < nbbs; i++)
2417 gimple_stmt_iterator si;
2418 basic_block bb = bbs[i];
2420 if (bb->loop_father == loop->inner)
2421 factor = innerloop_iters;
2425 for (si = gsi_start_bb (bb); !gsi_end_p (si); gsi_next (&si))
2427 gimple stmt = gsi_stmt (si);
2428 stmt_vec_info stmt_info = vinfo_for_stmt (stmt);
2430 if (!is_gimple_assign (stmt) && !is_gimple_call (stmt))
2433 /* Skip stmts that are not vectorized inside the loop. */
2435 && !STMT_VINFO_RELEVANT_P (stmt_info)
2436 && (!STMT_VINFO_LIVE_P (stmt_info)
2437 || !VECTORIZABLE_CYCLE_DEF (STMT_VINFO_DEF_TYPE (stmt_info)))
2438 && !STMT_VINFO_IN_PATTERN_P (stmt_info))
2441 if (STMT_VINFO_DATA_REF (vinfo_for_stmt (stmt)))
2443 if (DR_IS_READ (STMT_VINFO_DATA_REF (vinfo_for_stmt (stmt))))
2444 stmt_cost = vect_get_cost (scalar_load);
2446 stmt_cost = vect_get_cost (scalar_store);
2449 stmt_cost = vect_get_cost (scalar_stmt);
2451 scalar_single_iter_cost += stmt_cost * factor;
2454 return scalar_single_iter_cost;
2457 /* Calculate cost of peeling the loop PEEL_ITERS_PROLOGUE times. */
2459 vect_get_known_peeling_cost (loop_vec_info loop_vinfo, int peel_iters_prologue,
2460 int *peel_iters_epilogue,
2461 int scalar_single_iter_cost)
2463 int peel_guard_costs = 0;
2464 int vf = LOOP_VINFO_VECT_FACTOR (loop_vinfo);
2466 if (!LOOP_VINFO_NITERS_KNOWN_P (loop_vinfo))
2468 *peel_iters_epilogue = vf/2;
2469 if (vect_print_dump_info (REPORT_COST))
2470 fprintf (vect_dump, "cost model: "
2471 "epilogue peel iters set to vf/2 because "
2472 "loop iterations are unknown .");
2474 /* If peeled iterations are known but number of scalar loop
2475 iterations are unknown, count a taken branch per peeled loop. */
2476 peel_guard_costs = 2 * vect_get_cost (cond_branch_taken);
2480 int niters = LOOP_VINFO_INT_NITERS (loop_vinfo);
2481 peel_iters_prologue = niters < peel_iters_prologue ?
2482 niters : peel_iters_prologue;
2483 *peel_iters_epilogue = (niters - peel_iters_prologue) % vf;
2484 /* If we need to peel for gaps, but no peeling is required, we have to
2485 peel VF iterations. */
2486 if (LOOP_VINFO_PEELING_FOR_GAPS (loop_vinfo) && !*peel_iters_epilogue)
2487 *peel_iters_epilogue = vf;
2490 return (peel_iters_prologue * scalar_single_iter_cost)
2491 + (*peel_iters_epilogue * scalar_single_iter_cost)
2495 /* Function vect_estimate_min_profitable_iters
2497 Return the number of iterations required for the vector version of the
2498 loop to be profitable relative to the cost of the scalar version of the
2501 TODO: Take profile info into account before making vectorization
2502 decisions, if available. */
2505 vect_estimate_min_profitable_iters (loop_vec_info loop_vinfo)
2508 int min_profitable_iters;
2509 int peel_iters_prologue;
2510 int peel_iters_epilogue;
2511 int vec_inside_cost = 0;
2512 int vec_outside_cost = 0;
2513 int scalar_single_iter_cost = 0;
2514 int scalar_outside_cost = 0;
2515 int vf = LOOP_VINFO_VECT_FACTOR (loop_vinfo);
2516 struct loop *loop = LOOP_VINFO_LOOP (loop_vinfo);
2517 basic_block *bbs = LOOP_VINFO_BBS (loop_vinfo);
2518 int nbbs = loop->num_nodes;
2519 int npeel = LOOP_PEELING_FOR_ALIGNMENT (loop_vinfo);
2520 int peel_guard_costs = 0;
2521 int innerloop_iters = 0, factor;
2522 VEC (slp_instance, heap) *slp_instances;
2523 slp_instance instance;
2525 /* Cost model disabled. */
2526 if (!flag_vect_cost_model)
2528 if (vect_print_dump_info (REPORT_COST))
2529 fprintf (vect_dump, "cost model disabled.");
2533 /* Requires loop versioning tests to handle misalignment. */
2534 if (LOOP_REQUIRES_VERSIONING_FOR_ALIGNMENT (loop_vinfo))
2536 /* FIXME: Make cost depend on complexity of individual check. */
2538 VEC_length (gimple, LOOP_VINFO_MAY_MISALIGN_STMTS (loop_vinfo));
2539 if (vect_print_dump_info (REPORT_COST))
2540 fprintf (vect_dump, "cost model: Adding cost of checks for loop "
2541 "versioning to treat misalignment.\n");
2544 /* Requires loop versioning with alias checks. */
2545 if (LOOP_REQUIRES_VERSIONING_FOR_ALIAS (loop_vinfo))
2547 /* FIXME: Make cost depend on complexity of individual check. */
2549 VEC_length (ddr_p, LOOP_VINFO_MAY_ALIAS_DDRS (loop_vinfo));
2550 if (vect_print_dump_info (REPORT_COST))
2551 fprintf (vect_dump, "cost model: Adding cost of checks for loop "
2552 "versioning aliasing.\n");
2555 if (LOOP_REQUIRES_VERSIONING_FOR_ALIGNMENT (loop_vinfo)
2556 || LOOP_REQUIRES_VERSIONING_FOR_ALIAS (loop_vinfo))
2557 vec_outside_cost += vect_get_cost (cond_branch_taken);
2559 /* Count statements in scalar loop. Using this as scalar cost for a single
2562 TODO: Add outer loop support.
2564 TODO: Consider assigning different costs to different scalar
2569 innerloop_iters = 50; /* FIXME */
2571 for (i = 0; i < nbbs; i++)
2573 gimple_stmt_iterator si;
2574 basic_block bb = bbs[i];
2576 if (bb->loop_father == loop->inner)
2577 factor = innerloop_iters;
2581 for (si = gsi_start_bb (bb); !gsi_end_p (si); gsi_next (&si))
2583 gimple stmt = gsi_stmt (si);
2584 stmt_vec_info stmt_info = vinfo_for_stmt (stmt);
2586 if (STMT_VINFO_IN_PATTERN_P (stmt_info))
2588 stmt = STMT_VINFO_RELATED_STMT (stmt_info);
2589 stmt_info = vinfo_for_stmt (stmt);
2592 /* Skip stmts that are not vectorized inside the loop. */
2593 if (!STMT_VINFO_RELEVANT_P (stmt_info)
2594 && (!STMT_VINFO_LIVE_P (stmt_info)
2595 || !VECTORIZABLE_CYCLE_DEF (STMT_VINFO_DEF_TYPE (stmt_info))))
2598 vec_inside_cost += STMT_VINFO_INSIDE_OF_LOOP_COST (stmt_info) * factor;
2599 /* FIXME: for stmts in the inner-loop in outer-loop vectorization,
2600 some of the "outside" costs are generated inside the outer-loop. */
2601 vec_outside_cost += STMT_VINFO_OUTSIDE_OF_LOOP_COST (stmt_info);
2602 if (is_pattern_stmt_p (stmt_info)
2603 && STMT_VINFO_PATTERN_DEF_SEQ (stmt_info))
2605 gimple_stmt_iterator gsi;
2607 for (gsi = gsi_start (STMT_VINFO_PATTERN_DEF_SEQ (stmt_info));
2608 !gsi_end_p (gsi); gsi_next (&gsi))
2610 gimple pattern_def_stmt = gsi_stmt (gsi);
2611 stmt_vec_info pattern_def_stmt_info
2612 = vinfo_for_stmt (pattern_def_stmt);
2613 if (STMT_VINFO_RELEVANT_P (pattern_def_stmt_info)
2614 || STMT_VINFO_LIVE_P (pattern_def_stmt_info))
2617 += STMT_VINFO_INSIDE_OF_LOOP_COST
2618 (pattern_def_stmt_info) * factor;
2620 += STMT_VINFO_OUTSIDE_OF_LOOP_COST
2621 (pattern_def_stmt_info);
2628 scalar_single_iter_cost = vect_get_single_scalar_iteration_cost (loop_vinfo);
2630 /* Add additional cost for the peeled instructions in prologue and epilogue
2633 FORNOW: If we don't know the value of peel_iters for prologue or epilogue
2634 at compile-time - we assume it's vf/2 (the worst would be vf-1).
2636 TODO: Build an expression that represents peel_iters for prologue and
2637 epilogue to be used in a run-time test. */
2641 peel_iters_prologue = vf/2;
2642 if (vect_print_dump_info (REPORT_COST))
2643 fprintf (vect_dump, "cost model: "
2644 "prologue peel iters set to vf/2.");
2646 /* If peeling for alignment is unknown, loop bound of main loop becomes
2648 peel_iters_epilogue = vf/2;
2649 if (vect_print_dump_info (REPORT_COST))
2650 fprintf (vect_dump, "cost model: "
2651 "epilogue peel iters set to vf/2 because "
2652 "peeling for alignment is unknown .");
2654 /* If peeled iterations are unknown, count a taken branch and a not taken
2655 branch per peeled loop. Even if scalar loop iterations are known,
2656 vector iterations are not known since peeled prologue iterations are
2657 not known. Hence guards remain the same. */
2658 peel_guard_costs += 2 * (vect_get_cost (cond_branch_taken)
2659 + vect_get_cost (cond_branch_not_taken));
2660 vec_outside_cost += (peel_iters_prologue * scalar_single_iter_cost)
2661 + (peel_iters_epilogue * scalar_single_iter_cost)
2666 peel_iters_prologue = npeel;
2667 vec_outside_cost += vect_get_known_peeling_cost (loop_vinfo,
2668 peel_iters_prologue, &peel_iters_epilogue,
2669 scalar_single_iter_cost);
2672 /* FORNOW: The scalar outside cost is incremented in one of the
2675 1. The vectorizer checks for alignment and aliasing and generates
2676 a condition that allows dynamic vectorization. A cost model
2677 check is ANDED with the versioning condition. Hence scalar code
2678 path now has the added cost of the versioning check.
2680 if (cost > th & versioning_check)
2683 Hence run-time scalar is incremented by not-taken branch cost.
2685 2. The vectorizer then checks if a prologue is required. If the
2686 cost model check was not done before during versioning, it has to
2687 be done before the prologue check.
2690 prologue = scalar_iters
2695 if (prologue == num_iters)
2698 Hence the run-time scalar cost is incremented by a taken branch,
2699 plus a not-taken branch, plus a taken branch cost.
2701 3. The vectorizer then checks if an epilogue is required. If the
2702 cost model check was not done before during prologue check, it
2703 has to be done with the epilogue check.
2709 if (prologue == num_iters)
2712 if ((cost <= th) | (scalar_iters-prologue-epilogue == 0))
2715 Hence the run-time scalar cost should be incremented by 2 taken
2718 TODO: The back end may reorder the BBS's differently and reverse
2719 conditions/branch directions. Change the estimates below to
2720 something more reasonable. */
2722 /* If the number of iterations is known and we do not do versioning, we can
2723 decide whether to vectorize at compile time. Hence the scalar version
2724 do not carry cost model guard costs. */
2725 if (!LOOP_VINFO_NITERS_KNOWN_P (loop_vinfo)
2726 || LOOP_REQUIRES_VERSIONING_FOR_ALIGNMENT (loop_vinfo)
2727 || LOOP_REQUIRES_VERSIONING_FOR_ALIAS (loop_vinfo))
2729 /* Cost model check occurs at versioning. */
2730 if (LOOP_REQUIRES_VERSIONING_FOR_ALIGNMENT (loop_vinfo)
2731 || LOOP_REQUIRES_VERSIONING_FOR_ALIAS (loop_vinfo))
2732 scalar_outside_cost += vect_get_cost (cond_branch_not_taken);
2735 /* Cost model check occurs at prologue generation. */
2736 if (LOOP_PEELING_FOR_ALIGNMENT (loop_vinfo) < 0)
2737 scalar_outside_cost += 2 * vect_get_cost (cond_branch_taken)
2738 + vect_get_cost (cond_branch_not_taken);
2739 /* Cost model check occurs at epilogue generation. */
2741 scalar_outside_cost += 2 * vect_get_cost (cond_branch_taken);
2745 /* Add SLP costs. */
2746 slp_instances = LOOP_VINFO_SLP_INSTANCES (loop_vinfo);
2747 FOR_EACH_VEC_ELT (slp_instance, slp_instances, i, instance)
2749 vec_outside_cost += SLP_INSTANCE_OUTSIDE_OF_LOOP_COST (instance);
2750 vec_inside_cost += SLP_INSTANCE_INSIDE_OF_LOOP_COST (instance);
2753 /* Calculate number of iterations required to make the vector version
2754 profitable, relative to the loop bodies only. The following condition
2756 SIC * niters + SOC > VIC * ((niters-PL_ITERS-EP_ITERS)/VF) + VOC
2758 SIC = scalar iteration cost, VIC = vector iteration cost,
2759 VOC = vector outside cost, VF = vectorization factor,
2760 PL_ITERS = prologue iterations, EP_ITERS= epilogue iterations
2761 SOC = scalar outside cost for run time cost model check. */
2763 if ((scalar_single_iter_cost * vf) > vec_inside_cost)
2765 if (vec_outside_cost <= 0)
2766 min_profitable_iters = 1;
2769 min_profitable_iters = ((vec_outside_cost - scalar_outside_cost) * vf
2770 - vec_inside_cost * peel_iters_prologue
2771 - vec_inside_cost * peel_iters_epilogue)
2772 / ((scalar_single_iter_cost * vf)
2775 if ((scalar_single_iter_cost * vf * min_profitable_iters)
2776 <= ((vec_inside_cost * min_profitable_iters)
2777 + ((vec_outside_cost - scalar_outside_cost) * vf)))
2778 min_profitable_iters++;
2781 /* vector version will never be profitable. */
2784 if (vect_print_dump_info (REPORT_COST))
2785 fprintf (vect_dump, "cost model: the vector iteration cost = %d "
2786 "divided by the scalar iteration cost = %d "
2787 "is greater or equal to the vectorization factor = %d.",
2788 vec_inside_cost, scalar_single_iter_cost, vf);
2792 if (vect_print_dump_info (REPORT_COST))
2794 fprintf (vect_dump, "Cost model analysis: \n");
2795 fprintf (vect_dump, " Vector inside of loop cost: %d\n",
2797 fprintf (vect_dump, " Vector outside of loop cost: %d\n",
2799 fprintf (vect_dump, " Scalar iteration cost: %d\n",
2800 scalar_single_iter_cost);
2801 fprintf (vect_dump, " Scalar outside cost: %d\n", scalar_outside_cost);
2802 fprintf (vect_dump, " prologue iterations: %d\n",
2803 peel_iters_prologue);
2804 fprintf (vect_dump, " epilogue iterations: %d\n",
2805 peel_iters_epilogue);
2806 fprintf (vect_dump, " Calculated minimum iters for profitability: %d\n",
2807 min_profitable_iters);
2810 min_profitable_iters =
2811 min_profitable_iters < vf ? vf : min_profitable_iters;
2813 /* Because the condition we create is:
2814 if (niters <= min_profitable_iters)
2815 then skip the vectorized loop. */
2816 min_profitable_iters--;
2818 if (vect_print_dump_info (REPORT_COST))
2819 fprintf (vect_dump, " Profitability threshold = %d\n",
2820 min_profitable_iters);
2822 return min_profitable_iters;
2826 /* TODO: Close dependency between vect_model_*_cost and vectorizable_*
2827 functions. Design better to avoid maintenance issues. */
2829 /* Function vect_model_reduction_cost.
2831 Models cost for a reduction operation, including the vector ops
2832 generated within the strip-mine loop, the initial definition before
2833 the loop, and the epilogue code that must be generated. */
2836 vect_model_reduction_cost (stmt_vec_info stmt_info, enum tree_code reduc_code,
2840 enum tree_code code;
2843 gimple stmt, orig_stmt;
2845 enum machine_mode mode;
2846 loop_vec_info loop_vinfo = STMT_VINFO_LOOP_VINFO (stmt_info);
2847 struct loop *loop = LOOP_VINFO_LOOP (loop_vinfo);
2850 /* Cost of reduction op inside loop. */
2851 STMT_VINFO_INSIDE_OF_LOOP_COST (stmt_info)
2852 += ncopies * vect_get_cost (vector_stmt);
2854 stmt = STMT_VINFO_STMT (stmt_info);
2856 switch (get_gimple_rhs_class (gimple_assign_rhs_code (stmt)))
2858 case GIMPLE_SINGLE_RHS:
2859 gcc_assert (TREE_OPERAND_LENGTH (gimple_assign_rhs1 (stmt)) == ternary_op);
2860 reduction_op = TREE_OPERAND (gimple_assign_rhs1 (stmt), 2);
2862 case GIMPLE_UNARY_RHS:
2863 reduction_op = gimple_assign_rhs1 (stmt);
2865 case GIMPLE_BINARY_RHS:
2866 reduction_op = gimple_assign_rhs2 (stmt);
2868 case GIMPLE_TERNARY_RHS:
2869 reduction_op = gimple_assign_rhs3 (stmt);
2875 vectype = get_vectype_for_scalar_type (TREE_TYPE (reduction_op));
2878 if (vect_print_dump_info (REPORT_COST))
2880 fprintf (vect_dump, "unsupported data-type ");
2881 print_generic_expr (vect_dump, TREE_TYPE (reduction_op), TDF_SLIM);
2886 mode = TYPE_MODE (vectype);
2887 orig_stmt = STMT_VINFO_RELATED_STMT (stmt_info);
2890 orig_stmt = STMT_VINFO_STMT (stmt_info);
2892 code = gimple_assign_rhs_code (orig_stmt);
2894 /* Add in cost for initial definition. */
2895 outer_cost += vect_get_cost (scalar_to_vec);
2897 /* Determine cost of epilogue code.
2899 We have a reduction operator that will reduce the vector in one statement.
2900 Also requires scalar extract. */
2902 if (!nested_in_vect_loop_p (loop, orig_stmt))
2904 if (reduc_code != ERROR_MARK)
2905 outer_cost += vect_get_cost (vector_stmt)
2906 + vect_get_cost (vec_to_scalar);
2909 int vec_size_in_bits = tree_low_cst (TYPE_SIZE (vectype), 1);
2911 TYPE_SIZE (TREE_TYPE (gimple_assign_lhs (orig_stmt)));
2912 int element_bitsize = tree_low_cst (bitsize, 1);
2913 int nelements = vec_size_in_bits / element_bitsize;
2915 optab = optab_for_tree_code (code, vectype, optab_default);
2917 /* We have a whole vector shift available. */
2918 if (VECTOR_MODE_P (mode)
2919 && optab_handler (optab, mode) != CODE_FOR_nothing
2920 && optab_handler (vec_shr_optab, mode) != CODE_FOR_nothing)
2921 /* Final reduction via vector shifts and the reduction operator. Also
2922 requires scalar extract. */
2923 outer_cost += ((exact_log2(nelements) * 2)
2924 * vect_get_cost (vector_stmt)
2925 + vect_get_cost (vec_to_scalar));
2927 /* Use extracts and reduction op for final reduction. For N elements,
2928 we have N extracts and N-1 reduction ops. */
2929 outer_cost += ((nelements + nelements - 1)
2930 * vect_get_cost (vector_stmt));
2934 STMT_VINFO_OUTSIDE_OF_LOOP_COST (stmt_info) = outer_cost;
2936 if (vect_print_dump_info (REPORT_COST))
2937 fprintf (vect_dump, "vect_model_reduction_cost: inside_cost = %d, "
2938 "outside_cost = %d .", STMT_VINFO_INSIDE_OF_LOOP_COST (stmt_info),
2939 STMT_VINFO_OUTSIDE_OF_LOOP_COST (stmt_info));
2945 /* Function vect_model_induction_cost.
2947 Models cost for induction operations. */
2950 vect_model_induction_cost (stmt_vec_info stmt_info, int ncopies)
2952 /* loop cost for vec_loop. */
2953 STMT_VINFO_INSIDE_OF_LOOP_COST (stmt_info)
2954 = ncopies * vect_get_cost (vector_stmt);
2955 /* prologue cost for vec_init and vec_step. */
2956 STMT_VINFO_OUTSIDE_OF_LOOP_COST (stmt_info)
2957 = 2 * vect_get_cost (scalar_to_vec);
2959 if (vect_print_dump_info (REPORT_COST))
2960 fprintf (vect_dump, "vect_model_induction_cost: inside_cost = %d, "
2961 "outside_cost = %d .", STMT_VINFO_INSIDE_OF_LOOP_COST (stmt_info),
2962 STMT_VINFO_OUTSIDE_OF_LOOP_COST (stmt_info));
2966 /* Function get_initial_def_for_induction
2969 STMT - a stmt that performs an induction operation in the loop.
2970 IV_PHI - the initial value of the induction variable
2973 Return a vector variable, initialized with the first VF values of
2974 the induction variable. E.g., for an iv with IV_PHI='X' and
2975 evolution S, for a vector of 4 units, we want to return:
2976 [X, X + S, X + 2*S, X + 3*S]. */
2979 get_initial_def_for_induction (gimple iv_phi)
2981 stmt_vec_info stmt_vinfo = vinfo_for_stmt (iv_phi);
2982 loop_vec_info loop_vinfo = STMT_VINFO_LOOP_VINFO (stmt_vinfo);
2983 struct loop *loop = LOOP_VINFO_LOOP (loop_vinfo);
2987 edge pe = loop_preheader_edge (loop);
2988 struct loop *iv_loop;
2990 tree vec, vec_init, vec_step, t;
2994 gimple init_stmt, induction_phi, new_stmt;
2995 tree induc_def, vec_def, vec_dest;
2996 tree init_expr, step_expr;
2997 int vf = LOOP_VINFO_VECT_FACTOR (loop_vinfo);
3002 stmt_vec_info phi_info = vinfo_for_stmt (iv_phi);
3003 bool nested_in_vect_loop = false;
3004 gimple_seq stmts = NULL;
3005 imm_use_iterator imm_iter;
3006 use_operand_p use_p;
3010 gimple_stmt_iterator si;
3011 basic_block bb = gimple_bb (iv_phi);
3015 /* Is phi in an inner-loop, while vectorizing an enclosing outer-loop? */
3016 if (nested_in_vect_loop_p (loop, iv_phi))
3018 nested_in_vect_loop = true;
3019 iv_loop = loop->inner;
3023 gcc_assert (iv_loop == (gimple_bb (iv_phi))->loop_father);
3025 latch_e = loop_latch_edge (iv_loop);
3026 loop_arg = PHI_ARG_DEF_FROM_EDGE (iv_phi, latch_e);
3028 access_fn = analyze_scalar_evolution (iv_loop, PHI_RESULT (iv_phi));
3029 gcc_assert (access_fn);
3030 STRIP_NOPS (access_fn);
3031 ok = vect_is_simple_iv_evolution (iv_loop->num, access_fn,
3032 &init_expr, &step_expr);
3034 pe = loop_preheader_edge (iv_loop);
3036 scalar_type = TREE_TYPE (init_expr);
3037 vectype = get_vectype_for_scalar_type (scalar_type);
3038 resvectype = get_vectype_for_scalar_type (TREE_TYPE (PHI_RESULT (iv_phi)));
3039 gcc_assert (vectype);
3040 nunits = TYPE_VECTOR_SUBPARTS (vectype);
3041 ncopies = vf / nunits;
3043 gcc_assert (phi_info);
3044 gcc_assert (ncopies >= 1);
3046 /* Find the first insertion point in the BB. */
3047 si = gsi_after_labels (bb);
3049 /* Create the vector that holds the initial_value of the induction. */
3050 if (nested_in_vect_loop)
3052 /* iv_loop is nested in the loop to be vectorized. init_expr had already
3053 been created during vectorization of previous stmts. We obtain it
3054 from the STMT_VINFO_VEC_STMT of the defining stmt. */
3055 tree iv_def = PHI_ARG_DEF_FROM_EDGE (iv_phi,
3056 loop_preheader_edge (iv_loop));
3057 vec_init = vect_get_vec_def_for_operand (iv_def, iv_phi, NULL);
3061 /* iv_loop is the loop to be vectorized. Create:
3062 vec_init = [X, X+S, X+2*S, X+3*S] (S = step_expr, X = init_expr) */
3063 new_var = vect_get_new_vect_var (scalar_type, vect_scalar_var, "var_");
3064 add_referenced_var (new_var);
3066 new_name = force_gimple_operand (init_expr, &stmts, false, new_var);
3069 new_bb = gsi_insert_seq_on_edge_immediate (pe, stmts);
3070 gcc_assert (!new_bb);
3074 t = tree_cons (NULL_TREE, new_name, t);
3075 for (i = 1; i < nunits; i++)
3077 /* Create: new_name_i = new_name + step_expr */
3078 enum tree_code code = POINTER_TYPE_P (scalar_type)
3079 ? POINTER_PLUS_EXPR : PLUS_EXPR;
3080 init_stmt = gimple_build_assign_with_ops (code, new_var,
3081 new_name, step_expr);
3082 new_name = make_ssa_name (new_var, init_stmt);
3083 gimple_assign_set_lhs (init_stmt, new_name);
3085 new_bb = gsi_insert_on_edge_immediate (pe, init_stmt);
3086 gcc_assert (!new_bb);
3088 if (vect_print_dump_info (REPORT_DETAILS))
3090 fprintf (vect_dump, "created new init_stmt: ");
3091 print_gimple_stmt (vect_dump, init_stmt, 0, TDF_SLIM);
3093 t = tree_cons (NULL_TREE, new_name, t);
3095 /* Create a vector from [new_name_0, new_name_1, ..., new_name_nunits-1] */
3096 vec = build_constructor_from_list (vectype, nreverse (t));
3097 vec_init = vect_init_vector (iv_phi, vec, vectype, NULL);
3101 /* Create the vector that holds the step of the induction. */
3102 if (nested_in_vect_loop)
3103 /* iv_loop is nested in the loop to be vectorized. Generate:
3104 vec_step = [S, S, S, S] */
3105 new_name = step_expr;
3108 /* iv_loop is the loop to be vectorized. Generate:
3109 vec_step = [VF*S, VF*S, VF*S, VF*S] */
3110 expr = build_int_cst (TREE_TYPE (step_expr), vf);
3111 new_name = fold_build2 (MULT_EXPR, TREE_TYPE (step_expr),
3115 t = unshare_expr (new_name);
3116 gcc_assert (CONSTANT_CLASS_P (new_name));
3117 stepvectype = get_vectype_for_scalar_type (TREE_TYPE (new_name));
3118 gcc_assert (stepvectype);
3119 vec = build_vector_from_val (stepvectype, t);
3120 vec_step = vect_init_vector (iv_phi, vec, stepvectype, NULL);
3123 /* Create the following def-use cycle:
3128 vec_iv = PHI <vec_init, vec_loop>
3132 vec_loop = vec_iv + vec_step; */
3134 /* Create the induction-phi that defines the induction-operand. */
3135 vec_dest = vect_get_new_vect_var (vectype, vect_simple_var, "vec_iv_");
3136 add_referenced_var (vec_dest);
3137 induction_phi = create_phi_node (vec_dest, iv_loop->header);
3138 set_vinfo_for_stmt (induction_phi,
3139 new_stmt_vec_info (induction_phi, loop_vinfo, NULL));
3140 induc_def = PHI_RESULT (induction_phi);
3142 /* Create the iv update inside the loop */
3143 new_stmt = gimple_build_assign_with_ops (PLUS_EXPR, vec_dest,
3144 induc_def, vec_step);
3145 vec_def = make_ssa_name (vec_dest, new_stmt);
3146 gimple_assign_set_lhs (new_stmt, vec_def);
3147 gsi_insert_before (&si, new_stmt, GSI_SAME_STMT);
3148 set_vinfo_for_stmt (new_stmt, new_stmt_vec_info (new_stmt, loop_vinfo,
3151 /* Set the arguments of the phi node: */
3152 add_phi_arg (induction_phi, vec_init, pe, UNKNOWN_LOCATION);
3153 add_phi_arg (induction_phi, vec_def, loop_latch_edge (iv_loop),
3157 /* In case that vectorization factor (VF) is bigger than the number
3158 of elements that we can fit in a vectype (nunits), we have to generate
3159 more than one vector stmt - i.e - we need to "unroll" the
3160 vector stmt by a factor VF/nunits. For more details see documentation
3161 in vectorizable_operation. */
3165 stmt_vec_info prev_stmt_vinfo;
3166 /* FORNOW. This restriction should be relaxed. */
3167 gcc_assert (!nested_in_vect_loop);
3169 /* Create the vector that holds the step of the induction. */
3170 expr = build_int_cst (TREE_TYPE (step_expr), nunits);
3171 new_name = fold_build2 (MULT_EXPR, TREE_TYPE (step_expr),
3173 t = unshare_expr (new_name);
3174 gcc_assert (CONSTANT_CLASS_P (new_name));
3175 vec = build_vector_from_val (stepvectype, t);
3176 vec_step = vect_init_vector (iv_phi, vec, stepvectype, NULL);
3178 vec_def = induc_def;
3179 prev_stmt_vinfo = vinfo_for_stmt (induction_phi);
3180 for (i = 1; i < ncopies; i++)
3182 /* vec_i = vec_prev + vec_step */
3183 new_stmt = gimple_build_assign_with_ops (PLUS_EXPR, vec_dest,
3185 vec_def = make_ssa_name (vec_dest, new_stmt);
3186 gimple_assign_set_lhs (new_stmt, vec_def);
3188 gsi_insert_before (&si, new_stmt, GSI_SAME_STMT);
3189 if (!useless_type_conversion_p (resvectype, vectype))
3191 new_stmt = gimple_build_assign_with_ops
3193 vect_get_new_vect_var (resvectype, vect_simple_var,
3195 build1 (VIEW_CONVERT_EXPR, resvectype,
3196 gimple_assign_lhs (new_stmt)), NULL_TREE);
3197 gimple_assign_set_lhs (new_stmt,
3199 (gimple_assign_lhs (new_stmt), new_stmt));
3200 gsi_insert_before (&si, new_stmt, GSI_SAME_STMT);
3202 set_vinfo_for_stmt (new_stmt,
3203 new_stmt_vec_info (new_stmt, loop_vinfo, NULL));
3204 STMT_VINFO_RELATED_STMT (prev_stmt_vinfo) = new_stmt;
3205 prev_stmt_vinfo = vinfo_for_stmt (new_stmt);
3209 if (nested_in_vect_loop)
3211 /* Find the loop-closed exit-phi of the induction, and record
3212 the final vector of induction results: */
3214 FOR_EACH_IMM_USE_FAST (use_p, imm_iter, loop_arg)
3216 if (!flow_bb_inside_loop_p (iv_loop, gimple_bb (USE_STMT (use_p))))
3218 exit_phi = USE_STMT (use_p);
3224 stmt_vec_info stmt_vinfo = vinfo_for_stmt (exit_phi);
3225 /* FORNOW. Currently not supporting the case that an inner-loop induction
3226 is not used in the outer-loop (i.e. only outside the outer-loop). */
3227 gcc_assert (STMT_VINFO_RELEVANT_P (stmt_vinfo)
3228 && !STMT_VINFO_LIVE_P (stmt_vinfo));
3230 STMT_VINFO_VEC_STMT (stmt_vinfo) = new_stmt;
3231 if (vect_print_dump_info (REPORT_DETAILS))
3233 fprintf (vect_dump, "vector of inductions after inner-loop:");
3234 print_gimple_stmt (vect_dump, new_stmt, 0, TDF_SLIM);
3240 if (vect_print_dump_info (REPORT_DETAILS))
3242 fprintf (vect_dump, "transform induction: created def-use cycle: ");
3243 print_gimple_stmt (vect_dump, induction_phi, 0, TDF_SLIM);
3244 fprintf (vect_dump, "\n");
3245 print_gimple_stmt (vect_dump, SSA_NAME_DEF_STMT (vec_def), 0, TDF_SLIM);
3248 STMT_VINFO_VEC_STMT (phi_info) = induction_phi;
3249 if (!useless_type_conversion_p (resvectype, vectype))
3251 new_stmt = gimple_build_assign_with_ops
3253 vect_get_new_vect_var (resvectype, vect_simple_var, "vec_iv_"),
3254 build1 (VIEW_CONVERT_EXPR, resvectype, induc_def), NULL_TREE);
3255 induc_def = make_ssa_name (gimple_assign_lhs (new_stmt), new_stmt);
3256 gimple_assign_set_lhs (new_stmt, induc_def);
3257 si = gsi_start_bb (bb);
3258 gsi_insert_before (&si, new_stmt, GSI_SAME_STMT);
3259 set_vinfo_for_stmt (new_stmt,
3260 new_stmt_vec_info (new_stmt, loop_vinfo, NULL));
3261 STMT_VINFO_RELATED_STMT (vinfo_for_stmt (new_stmt))
3262 = STMT_VINFO_RELATED_STMT (vinfo_for_stmt (induction_phi));
3269 /* Function get_initial_def_for_reduction
3272 STMT - a stmt that performs a reduction operation in the loop.
3273 INIT_VAL - the initial value of the reduction variable
3276 ADJUSTMENT_DEF - a tree that holds a value to be added to the final result
3277 of the reduction (used for adjusting the epilog - see below).
3278 Return a vector variable, initialized according to the operation that STMT
3279 performs. This vector will be used as the initial value of the
3280 vector of partial results.
3282 Option1 (adjust in epilog): Initialize the vector as follows:
3283 add/bit or/xor: [0,0,...,0,0]
3284 mult/bit and: [1,1,...,1,1]
3285 min/max/cond_expr: [init_val,init_val,..,init_val,init_val]
3286 and when necessary (e.g. add/mult case) let the caller know
3287 that it needs to adjust the result by init_val.
3289 Option2: Initialize the vector as follows:
3290 add/bit or/xor: [init_val,0,0,...,0]
3291 mult/bit and: [init_val,1,1,...,1]
3292 min/max/cond_expr: [init_val,init_val,...,init_val]
3293 and no adjustments are needed.
3295 For example, for the following code:
3301 STMT is 's = s + a[i]', and the reduction variable is 's'.
3302 For a vector of 4 units, we want to return either [0,0,0,init_val],
3303 or [0,0,0,0] and let the caller know that it needs to adjust
3304 the result at the end by 'init_val'.
3306 FORNOW, we are using the 'adjust in epilog' scheme, because this way the
3307 initialization vector is simpler (same element in all entries), if
3308 ADJUSTMENT_DEF is not NULL, and Option2 otherwise.
3310 A cost model should help decide between these two schemes. */
3313 get_initial_def_for_reduction (gimple stmt, tree init_val,
3314 tree *adjustment_def)
3316 stmt_vec_info stmt_vinfo = vinfo_for_stmt (stmt);
3317 loop_vec_info loop_vinfo = STMT_VINFO_LOOP_VINFO (stmt_vinfo);
3318 struct loop *loop = LOOP_VINFO_LOOP (loop_vinfo);
3319 tree scalar_type = TREE_TYPE (init_val);
3320 tree vectype = get_vectype_for_scalar_type (scalar_type);
3322 enum tree_code code = gimple_assign_rhs_code (stmt);
3327 bool nested_in_vect_loop = false;
3329 REAL_VALUE_TYPE real_init_val = dconst0;
3330 int int_init_val = 0;
3331 gimple def_stmt = NULL;
3333 gcc_assert (vectype);
3334 nunits = TYPE_VECTOR_SUBPARTS (vectype);
3336 gcc_assert (POINTER_TYPE_P (scalar_type) || INTEGRAL_TYPE_P (scalar_type)
3337 || SCALAR_FLOAT_TYPE_P (scalar_type));
3339 if (nested_in_vect_loop_p (loop, stmt))
3340 nested_in_vect_loop = true;
3342 gcc_assert (loop == (gimple_bb (stmt))->loop_father);
3344 /* In case of double reduction we only create a vector variable to be put
3345 in the reduction phi node. The actual statement creation is done in
3346 vect_create_epilog_for_reduction. */
3347 if (adjustment_def && nested_in_vect_loop
3348 && TREE_CODE (init_val) == SSA_NAME
3349 && (def_stmt = SSA_NAME_DEF_STMT (init_val))
3350 && gimple_code (def_stmt) == GIMPLE_PHI
3351 && flow_bb_inside_loop_p (loop, gimple_bb (def_stmt))
3352 && vinfo_for_stmt (def_stmt)
3353 && STMT_VINFO_DEF_TYPE (vinfo_for_stmt (def_stmt))
3354 == vect_double_reduction_def)
3356 *adjustment_def = NULL;
3357 return vect_create_destination_var (init_val, vectype);
3360 if (TREE_CONSTANT (init_val))
3362 if (SCALAR_FLOAT_TYPE_P (scalar_type))
3363 init_value = build_real (scalar_type, TREE_REAL_CST (init_val));
3365 init_value = build_int_cst (scalar_type, TREE_INT_CST_LOW (init_val));
3368 init_value = init_val;
3372 case WIDEN_SUM_EXPR:
3380 /* ADJUSMENT_DEF is NULL when called from
3381 vect_create_epilog_for_reduction to vectorize double reduction. */
3384 if (nested_in_vect_loop)
3385 *adjustment_def = vect_get_vec_def_for_operand (init_val, stmt,
3388 *adjustment_def = init_val;
3391 if (code == MULT_EXPR)
3393 real_init_val = dconst1;
3397 if (code == BIT_AND_EXPR)
3400 if (SCALAR_FLOAT_TYPE_P (scalar_type))
3401 def_for_init = build_real (scalar_type, real_init_val);
3403 def_for_init = build_int_cst (scalar_type, int_init_val);
3405 /* Create a vector of '0' or '1' except the first element. */
3406 for (i = nunits - 2; i >= 0; --i)
3407 t = tree_cons (NULL_TREE, def_for_init, t);
3409 /* Option1: the first element is '0' or '1' as well. */
3412 t = tree_cons (NULL_TREE, def_for_init, t);
3413 init_def = build_vector (vectype, t);
3417 /* Option2: the first element is INIT_VAL. */
3418 t = tree_cons (NULL_TREE, init_value, t);
3419 if (TREE_CONSTANT (init_val))
3420 init_def = build_vector (vectype, t);
3422 init_def = build_constructor_from_list (vectype, t);
3431 *adjustment_def = NULL_TREE;
3432 init_def = vect_get_vec_def_for_operand (init_val, stmt, NULL);
3436 init_def = build_vector_from_val (vectype, init_value);
3447 /* Function vect_create_epilog_for_reduction
3449 Create code at the loop-epilog to finalize the result of a reduction
3452 VECT_DEFS is list of vector of partial results, i.e., the lhs's of vector
3453 reduction statements.
3454 STMT is the scalar reduction stmt that is being vectorized.
3455 NCOPIES is > 1 in case the vectorization factor (VF) is bigger than the
3456 number of elements that we can fit in a vectype (nunits). In this case
3457 we have to generate more than one vector stmt - i.e - we need to "unroll"
3458 the vector stmt by a factor VF/nunits. For more details see documentation
3459 in vectorizable_operation.
3460 REDUC_CODE is the tree-code for the epilog reduction.
3461 REDUCTION_PHIS is a list of the phi-nodes that carry the reduction
3463 REDUC_INDEX is the index of the operand in the right hand side of the
3464 statement that is defined by REDUCTION_PHI.
3465 DOUBLE_REDUC is TRUE if double reduction phi nodes should be handled.
3466 SLP_NODE is an SLP node containing a group of reduction statements. The
3467 first one in this group is STMT.
3470 1. Creates the reduction def-use cycles: sets the arguments for
3472 The loop-entry argument is the vectorized initial-value of the reduction.
3473 The loop-latch argument is taken from VECT_DEFS - the vector of partial
3475 2. "Reduces" each vector of partial results VECT_DEFS into a single result,
3476 by applying the operation specified by REDUC_CODE if available, or by
3477 other means (whole-vector shifts or a scalar loop).
3478 The function also creates a new phi node at the loop exit to preserve
3479 loop-closed form, as illustrated below.
3481 The flow at the entry to this function:
3484 vec_def = phi <null, null> # REDUCTION_PHI
3485 VECT_DEF = vector_stmt # vectorized form of STMT
3486 s_loop = scalar_stmt # (scalar) STMT
3488 s_out0 = phi <s_loop> # (scalar) EXIT_PHI
3492 The above is transformed by this function into:
3495 vec_def = phi <vec_init, VECT_DEF> # REDUCTION_PHI
3496 VECT_DEF = vector_stmt # vectorized form of STMT
3497 s_loop = scalar_stmt # (scalar) STMT
3499 s_out0 = phi <s_loop> # (scalar) EXIT_PHI
3500 v_out1 = phi <VECT_DEF> # NEW_EXIT_PHI
3501 v_out2 = reduce <v_out1>
3502 s_out3 = extract_field <v_out2, 0>
3503 s_out4 = adjust_result <s_out3>
3509 vect_create_epilog_for_reduction (VEC (tree, heap) *vect_defs, gimple stmt,
3510 int ncopies, enum tree_code reduc_code,
3511 VEC (gimple, heap) *reduction_phis,
3512 int reduc_index, bool double_reduc,
3515 stmt_vec_info stmt_info = vinfo_for_stmt (stmt);
3516 stmt_vec_info prev_phi_info;
3518 enum machine_mode mode;
3519 loop_vec_info loop_vinfo = STMT_VINFO_LOOP_VINFO (stmt_info);
3520 struct loop *loop = LOOP_VINFO_LOOP (loop_vinfo), *outer_loop = NULL;
3521 basic_block exit_bb;
3524 gimple new_phi = NULL, phi;
3525 gimple_stmt_iterator exit_gsi;
3527 tree new_temp = NULL_TREE, new_dest, new_name, new_scalar_dest;
3528 gimple epilog_stmt = NULL;
3529 enum tree_code code = gimple_assign_rhs_code (stmt);
3531 tree bitsize, bitpos;
3532 tree adjustment_def = NULL;
3533 tree vec_initial_def = NULL;
3534 tree reduction_op, expr, def;
3535 tree orig_name, scalar_result;
3536 imm_use_iterator imm_iter, phi_imm_iter;
3537 use_operand_p use_p, phi_use_p;
3538 bool extract_scalar_result = false;
3539 gimple use_stmt, orig_stmt, reduction_phi = NULL;
3540 bool nested_in_vect_loop = false;
3541 VEC (gimple, heap) *new_phis = NULL;
3542 VEC (gimple, heap) *inner_phis = NULL;
3543 enum vect_def_type dt = vect_unknown_def_type;
3545 VEC (tree, heap) *scalar_results = NULL;
3546 unsigned int group_size = 1, k, ratio;
3547 VEC (tree, heap) *vec_initial_defs = NULL;
3548 VEC (gimple, heap) *phis;
3549 bool slp_reduc = false;
3550 tree new_phi_result;
3551 gimple inner_phi = NULL;
3554 group_size = VEC_length (gimple, SLP_TREE_SCALAR_STMTS (slp_node));
3556 if (nested_in_vect_loop_p (loop, stmt))
3560 nested_in_vect_loop = true;
3561 gcc_assert (!slp_node);
3564 switch (get_gimple_rhs_class (gimple_assign_rhs_code (stmt)))
3566 case GIMPLE_SINGLE_RHS:
3567 gcc_assert (TREE_OPERAND_LENGTH (gimple_assign_rhs1 (stmt))
3569 reduction_op = TREE_OPERAND (gimple_assign_rhs1 (stmt), reduc_index);
3571 case GIMPLE_UNARY_RHS:
3572 reduction_op = gimple_assign_rhs1 (stmt);
3574 case GIMPLE_BINARY_RHS:
3575 reduction_op = reduc_index ?
3576 gimple_assign_rhs2 (stmt) : gimple_assign_rhs1 (stmt);
3578 case GIMPLE_TERNARY_RHS:
3579 reduction_op = gimple_op (stmt, reduc_index + 1);
3585 vectype = get_vectype_for_scalar_type (TREE_TYPE (reduction_op));
3586 gcc_assert (vectype);
3587 mode = TYPE_MODE (vectype);
3589 /* 1. Create the reduction def-use cycle:
3590 Set the arguments of REDUCTION_PHIS, i.e., transform
3593 vec_def = phi <null, null> # REDUCTION_PHI
3594 VECT_DEF = vector_stmt # vectorized form of STMT
3600 vec_def = phi <vec_init, VECT_DEF> # REDUCTION_PHI
3601 VECT_DEF = vector_stmt # vectorized form of STMT
3604 (in case of SLP, do it for all the phis). */
3606 /* Get the loop-entry arguments. */
3608 vect_get_vec_defs (reduction_op, NULL_TREE, stmt, &vec_initial_defs,
3609 NULL, slp_node, reduc_index);
3612 vec_initial_defs = VEC_alloc (tree, heap, 1);
3613 /* For the case of reduction, vect_get_vec_def_for_operand returns
3614 the scalar def before the loop, that defines the initial value
3615 of the reduction variable. */
3616 vec_initial_def = vect_get_vec_def_for_operand (reduction_op, stmt,
3618 VEC_quick_push (tree, vec_initial_defs, vec_initial_def);
3621 /* Set phi nodes arguments. */
3622 FOR_EACH_VEC_ELT (gimple, reduction_phis, i, phi)
3624 tree vec_init_def = VEC_index (tree, vec_initial_defs, i);
3625 tree def = VEC_index (tree, vect_defs, i);
3626 for (j = 0; j < ncopies; j++)
3628 /* Set the loop-entry arg of the reduction-phi. */
3629 add_phi_arg (phi, vec_init_def, loop_preheader_edge (loop),
3632 /* Set the loop-latch arg for the reduction-phi. */
3634 def = vect_get_vec_def_for_stmt_copy (vect_unknown_def_type, def);
3636 add_phi_arg (phi, def, loop_latch_edge (loop), UNKNOWN_LOCATION);
3638 if (vect_print_dump_info (REPORT_DETAILS))
3640 fprintf (vect_dump, "transform reduction: created def-use"
3642 print_gimple_stmt (vect_dump, phi, 0, TDF_SLIM);
3643 fprintf (vect_dump, "\n");
3644 print_gimple_stmt (vect_dump, SSA_NAME_DEF_STMT (def), 0,
3648 phi = STMT_VINFO_RELATED_STMT (vinfo_for_stmt (phi));
3652 VEC_free (tree, heap, vec_initial_defs);
3654 /* 2. Create epilog code.
3655 The reduction epilog code operates across the elements of the vector
3656 of partial results computed by the vectorized loop.
3657 The reduction epilog code consists of:
3659 step 1: compute the scalar result in a vector (v_out2)
3660 step 2: extract the scalar result (s_out3) from the vector (v_out2)
3661 step 3: adjust the scalar result (s_out3) if needed.
3663 Step 1 can be accomplished using one the following three schemes:
3664 (scheme 1) using reduc_code, if available.
3665 (scheme 2) using whole-vector shifts, if available.
3666 (scheme 3) using a scalar loop. In this case steps 1+2 above are
3669 The overall epilog code looks like this:
3671 s_out0 = phi <s_loop> # original EXIT_PHI
3672 v_out1 = phi <VECT_DEF> # NEW_EXIT_PHI
3673 v_out2 = reduce <v_out1> # step 1
3674 s_out3 = extract_field <v_out2, 0> # step 2
3675 s_out4 = adjust_result <s_out3> # step 3
3677 (step 3 is optional, and steps 1 and 2 may be combined).
3678 Lastly, the uses of s_out0 are replaced by s_out4. */
3681 /* 2.1 Create new loop-exit-phis to preserve loop-closed form:
3682 v_out1 = phi <VECT_DEF>
3683 Store them in NEW_PHIS. */
3685 exit_bb = single_exit (loop)->dest;
3686 prev_phi_info = NULL;
3687 new_phis = VEC_alloc (gimple, heap, VEC_length (tree, vect_defs));
3688 FOR_EACH_VEC_ELT (tree, vect_defs, i, def)
3690 for (j = 0; j < ncopies; j++)
3692 phi = create_phi_node (SSA_NAME_VAR (def), exit_bb);
3693 set_vinfo_for_stmt (phi, new_stmt_vec_info (phi, loop_vinfo, NULL));
3695 VEC_quick_push (gimple, new_phis, phi);
3698 def = vect_get_vec_def_for_stmt_copy (dt, def);
3699 STMT_VINFO_RELATED_STMT (prev_phi_info) = phi;
3702 SET_PHI_ARG_DEF (phi, single_exit (loop)->dest_idx, def);
3703 prev_phi_info = vinfo_for_stmt (phi);
3707 /* The epilogue is created for the outer-loop, i.e., for the loop being
3708 vectorized. Create exit phis for the outer loop. */
3712 exit_bb = single_exit (loop)->dest;
3713 inner_phis = VEC_alloc (gimple, heap, VEC_length (tree, vect_defs));
3714 FOR_EACH_VEC_ELT (gimple, new_phis, i, phi)
3716 gimple outer_phi = create_phi_node (SSA_NAME_VAR (PHI_RESULT (phi)),
3718 SET_PHI_ARG_DEF (outer_phi, single_exit (loop)->dest_idx,
3720 set_vinfo_for_stmt (outer_phi, new_stmt_vec_info (outer_phi,
3722 VEC_quick_push (gimple, inner_phis, phi);
3723 VEC_replace (gimple, new_phis, i, outer_phi);
3724 prev_phi_info = vinfo_for_stmt (outer_phi);
3725 while (STMT_VINFO_RELATED_STMT (vinfo_for_stmt (phi)))
3727 phi = STMT_VINFO_RELATED_STMT (vinfo_for_stmt (phi));
3728 outer_phi = create_phi_node (SSA_NAME_VAR (PHI_RESULT (phi)),
3730 SET_PHI_ARG_DEF (outer_phi, single_exit (loop)->dest_idx,
3732 set_vinfo_for_stmt (outer_phi, new_stmt_vec_info (outer_phi,
3734 STMT_VINFO_RELATED_STMT (prev_phi_info) = outer_phi;
3735 prev_phi_info = vinfo_for_stmt (outer_phi);
3740 exit_gsi = gsi_after_labels (exit_bb);
3742 /* 2.2 Get the relevant tree-code to use in the epilog for schemes 2,3
3743 (i.e. when reduc_code is not available) and in the final adjustment
3744 code (if needed). Also get the original scalar reduction variable as
3745 defined in the loop. In case STMT is a "pattern-stmt" (i.e. - it
3746 represents a reduction pattern), the tree-code and scalar-def are
3747 taken from the original stmt that the pattern-stmt (STMT) replaces.
3748 Otherwise (it is a regular reduction) - the tree-code and scalar-def
3749 are taken from STMT. */
3751 orig_stmt = STMT_VINFO_RELATED_STMT (stmt_info);
3754 /* Regular reduction */
3759 /* Reduction pattern */
3760 stmt_vec_info stmt_vinfo = vinfo_for_stmt (orig_stmt);
3761 gcc_assert (STMT_VINFO_IN_PATTERN_P (stmt_vinfo));
3762 gcc_assert (STMT_VINFO_RELATED_STMT (stmt_vinfo) == stmt);
3765 code = gimple_assign_rhs_code (orig_stmt);
3766 /* For MINUS_EXPR the initial vector is [init_val,0,...,0], therefore,
3767 partial results are added and not subtracted. */
3768 if (code == MINUS_EXPR)
3771 scalar_dest = gimple_assign_lhs (orig_stmt);
3772 scalar_type = TREE_TYPE (scalar_dest);
3773 scalar_results = VEC_alloc (tree, heap, group_size);
3774 new_scalar_dest = vect_create_destination_var (scalar_dest, NULL);
3775 bitsize = TYPE_SIZE (scalar_type);
3777 /* In case this is a reduction in an inner-loop while vectorizing an outer
3778 loop - we don't need to extract a single scalar result at the end of the
3779 inner-loop (unless it is double reduction, i.e., the use of reduction is
3780 outside the outer-loop). The final vector of partial results will be used
3781 in the vectorized outer-loop, or reduced to a scalar result at the end of
3783 if (nested_in_vect_loop && !double_reduc)
3784 goto vect_finalize_reduction;
3786 /* SLP reduction without reduction chain, e.g.,
3790 b2 = operation (b1) */
3791 slp_reduc = (slp_node && !GROUP_FIRST_ELEMENT (vinfo_for_stmt (stmt)));
3793 /* In case of reduction chain, e.g.,
3796 a3 = operation (a2),
3798 we may end up with more than one vector result. Here we reduce them to
3800 if (GROUP_FIRST_ELEMENT (vinfo_for_stmt (stmt)))
3802 tree first_vect = PHI_RESULT (VEC_index (gimple, new_phis, 0));
3804 gimple new_vec_stmt = NULL;
3806 vec_dest = vect_create_destination_var (scalar_dest, vectype);
3807 for (k = 1; k < VEC_length (gimple, new_phis); k++)
3809 gimple next_phi = VEC_index (gimple, new_phis, k);
3810 tree second_vect = PHI_RESULT (next_phi);
3812 tmp = build2 (code, vectype, first_vect, second_vect);
3813 new_vec_stmt = gimple_build_assign (vec_dest, tmp);
3814 first_vect = make_ssa_name (vec_dest, new_vec_stmt);
3815 gimple_assign_set_lhs (new_vec_stmt, first_vect);
3816 gsi_insert_before (&exit_gsi, new_vec_stmt, GSI_SAME_STMT);
3819 new_phi_result = first_vect;
3822 VEC_truncate (gimple, new_phis, 0);
3823 VEC_safe_push (gimple, heap, new_phis, new_vec_stmt);
3827 new_phi_result = PHI_RESULT (VEC_index (gimple, new_phis, 0));
3829 /* 2.3 Create the reduction code, using one of the three schemes described
3830 above. In SLP we simply need to extract all the elements from the
3831 vector (without reducing them), so we use scalar shifts. */
3832 if (reduc_code != ERROR_MARK && !slp_reduc)
3836 /*** Case 1: Create:
3837 v_out2 = reduc_expr <v_out1> */
3839 if (vect_print_dump_info (REPORT_DETAILS))
3840 fprintf (vect_dump, "Reduce using direct vector reduction.");
3842 vec_dest = vect_create_destination_var (scalar_dest, vectype);
3843 tmp = build1 (reduc_code, vectype, new_phi_result);
3844 epilog_stmt = gimple_build_assign (vec_dest, tmp);
3845 new_temp = make_ssa_name (vec_dest, epilog_stmt);
3846 gimple_assign_set_lhs (epilog_stmt, new_temp);
3847 gsi_insert_before (&exit_gsi, epilog_stmt, GSI_SAME_STMT);
3849 extract_scalar_result = true;
3853 enum tree_code shift_code = ERROR_MARK;
3854 bool have_whole_vector_shift = true;
3856 int element_bitsize = tree_low_cst (bitsize, 1);
3857 int vec_size_in_bits = tree_low_cst (TYPE_SIZE (vectype), 1);
3860 if (optab_handler (vec_shr_optab, mode) != CODE_FOR_nothing)
3861 shift_code = VEC_RSHIFT_EXPR;
3863 have_whole_vector_shift = false;
3865 /* Regardless of whether we have a whole vector shift, if we're
3866 emulating the operation via tree-vect-generic, we don't want
3867 to use it. Only the first round of the reduction is likely
3868 to still be profitable via emulation. */
3869 /* ??? It might be better to emit a reduction tree code here, so that
3870 tree-vect-generic can expand the first round via bit tricks. */
3871 if (!VECTOR_MODE_P (mode))
3872 have_whole_vector_shift = false;
3875 optab optab = optab_for_tree_code (code, vectype, optab_default);
3876 if (optab_handler (optab, mode) == CODE_FOR_nothing)
3877 have_whole_vector_shift = false;
3880 if (have_whole_vector_shift && !slp_reduc)
3882 /*** Case 2: Create:
3883 for (offset = VS/2; offset >= element_size; offset/=2)
3885 Create: va' = vec_shift <va, offset>
3886 Create: va = vop <va, va'>
3889 if (vect_print_dump_info (REPORT_DETAILS))
3890 fprintf (vect_dump, "Reduce using vector shifts");
3892 vec_dest = vect_create_destination_var (scalar_dest, vectype);
3893 new_temp = new_phi_result;
3894 for (bit_offset = vec_size_in_bits/2;
3895 bit_offset >= element_bitsize;
3898 tree bitpos = size_int (bit_offset);
3900 epilog_stmt = gimple_build_assign_with_ops (shift_code,
3901 vec_dest, new_temp, bitpos);
3902 new_name = make_ssa_name (vec_dest, epilog_stmt);
3903 gimple_assign_set_lhs (epilog_stmt, new_name);
3904 gsi_insert_before (&exit_gsi, epilog_stmt, GSI_SAME_STMT);
3906 epilog_stmt = gimple_build_assign_with_ops (code, vec_dest,
3907 new_name, new_temp);
3908 new_temp = make_ssa_name (vec_dest, epilog_stmt);
3909 gimple_assign_set_lhs (epilog_stmt, new_temp);
3910 gsi_insert_before (&exit_gsi, epilog_stmt, GSI_SAME_STMT);
3913 extract_scalar_result = true;
3919 /*** Case 3: Create:
3920 s = extract_field <v_out2, 0>
3921 for (offset = element_size;
3922 offset < vector_size;
3923 offset += element_size;)
3925 Create: s' = extract_field <v_out2, offset>
3926 Create: s = op <s, s'> // For non SLP cases
3929 if (vect_print_dump_info (REPORT_DETAILS))
3930 fprintf (vect_dump, "Reduce using scalar code. ");
3932 vec_size_in_bits = tree_low_cst (TYPE_SIZE (vectype), 1);
3933 FOR_EACH_VEC_ELT (gimple, new_phis, i, new_phi)
3935 if (gimple_code (new_phi) == GIMPLE_PHI)
3936 vec_temp = PHI_RESULT (new_phi);
3938 vec_temp = gimple_assign_lhs (new_phi);
3939 rhs = build3 (BIT_FIELD_REF, scalar_type, vec_temp, bitsize,
3941 epilog_stmt = gimple_build_assign (new_scalar_dest, rhs);
3942 new_temp = make_ssa_name (new_scalar_dest, epilog_stmt);
3943 gimple_assign_set_lhs (epilog_stmt, new_temp);
3944 gsi_insert_before (&exit_gsi, epilog_stmt, GSI_SAME_STMT);
3946 /* In SLP we don't need to apply reduction operation, so we just
3947 collect s' values in SCALAR_RESULTS. */
3949 VEC_safe_push (tree, heap, scalar_results, new_temp);
3951 for (bit_offset = element_bitsize;
3952 bit_offset < vec_size_in_bits;
3953 bit_offset += element_bitsize)
3955 tree bitpos = bitsize_int (bit_offset);
3956 tree rhs = build3 (BIT_FIELD_REF, scalar_type, vec_temp,
3959 epilog_stmt = gimple_build_assign (new_scalar_dest, rhs);
3960 new_name = make_ssa_name (new_scalar_dest, epilog_stmt);
3961 gimple_assign_set_lhs (epilog_stmt, new_name);
3962 gsi_insert_before (&exit_gsi, epilog_stmt, GSI_SAME_STMT);
3966 /* In SLP we don't need to apply reduction operation, so
3967 we just collect s' values in SCALAR_RESULTS. */
3968 new_temp = new_name;
3969 VEC_safe_push (tree, heap, scalar_results, new_name);
3973 epilog_stmt = gimple_build_assign_with_ops (code,
3974 new_scalar_dest, new_name, new_temp);
3975 new_temp = make_ssa_name (new_scalar_dest, epilog_stmt);
3976 gimple_assign_set_lhs (epilog_stmt, new_temp);
3977 gsi_insert_before (&exit_gsi, epilog_stmt, GSI_SAME_STMT);
3982 /* The only case where we need to reduce scalar results in SLP, is
3983 unrolling. If the size of SCALAR_RESULTS is greater than
3984 GROUP_SIZE, we reduce them combining elements modulo
3988 tree res, first_res, new_res;
3991 /* Reduce multiple scalar results in case of SLP unrolling. */
3992 for (j = group_size; VEC_iterate (tree, scalar_results, j, res);
3995 first_res = VEC_index (tree, scalar_results, j % group_size);
3996 new_stmt = gimple_build_assign_with_ops (code,
3997 new_scalar_dest, first_res, res);
3998 new_res = make_ssa_name (new_scalar_dest, new_stmt);
3999 gimple_assign_set_lhs (new_stmt, new_res);
4000 gsi_insert_before (&exit_gsi, new_stmt, GSI_SAME_STMT);
4001 VEC_replace (tree, scalar_results, j % group_size, new_res);
4005 /* Not SLP - we have one scalar to keep in SCALAR_RESULTS. */
4006 VEC_safe_push (tree, heap, scalar_results, new_temp);
4008 extract_scalar_result = false;
4012 /* 2.4 Extract the final scalar result. Create:
4013 s_out3 = extract_field <v_out2, bitpos> */
4015 if (extract_scalar_result)
4019 if (vect_print_dump_info (REPORT_DETAILS))
4020 fprintf (vect_dump, "extract scalar result");
4022 if (BYTES_BIG_ENDIAN)
4023 bitpos = size_binop (MULT_EXPR,
4024 bitsize_int (TYPE_VECTOR_SUBPARTS (vectype) - 1),
4025 TYPE_SIZE (scalar_type));
4027 bitpos = bitsize_zero_node;
4029 rhs = build3 (BIT_FIELD_REF, scalar_type, new_temp, bitsize, bitpos);
4030 epilog_stmt = gimple_build_assign (new_scalar_dest, rhs);
4031 new_temp = make_ssa_name (new_scalar_dest, epilog_stmt);
4032 gimple_assign_set_lhs (epilog_stmt, new_temp);
4033 gsi_insert_before (&exit_gsi, epilog_stmt, GSI_SAME_STMT);
4034 VEC_safe_push (tree, heap, scalar_results, new_temp);
4037 vect_finalize_reduction:
4042 /* 2.5 Adjust the final result by the initial value of the reduction
4043 variable. (When such adjustment is not needed, then
4044 'adjustment_def' is zero). For example, if code is PLUS we create:
4045 new_temp = loop_exit_def + adjustment_def */
4049 gcc_assert (!slp_reduc);
4050 if (nested_in_vect_loop)
4052 new_phi = VEC_index (gimple, new_phis, 0);
4053 gcc_assert (TREE_CODE (TREE_TYPE (adjustment_def)) == VECTOR_TYPE);
4054 expr = build2 (code, vectype, PHI_RESULT (new_phi), adjustment_def);
4055 new_dest = vect_create_destination_var (scalar_dest, vectype);
4059 new_temp = VEC_index (tree, scalar_results, 0);
4060 gcc_assert (TREE_CODE (TREE_TYPE (adjustment_def)) != VECTOR_TYPE);
4061 expr = build2 (code, scalar_type, new_temp, adjustment_def);
4062 new_dest = vect_create_destination_var (scalar_dest, scalar_type);
4065 epilog_stmt = gimple_build_assign (new_dest, expr);
4066 new_temp = make_ssa_name (new_dest, epilog_stmt);
4067 gimple_assign_set_lhs (epilog_stmt, new_temp);
4068 SSA_NAME_DEF_STMT (new_temp) = epilog_stmt;
4069 gsi_insert_before (&exit_gsi, epilog_stmt, GSI_SAME_STMT);
4070 if (nested_in_vect_loop)
4072 set_vinfo_for_stmt (epilog_stmt,
4073 new_stmt_vec_info (epilog_stmt, loop_vinfo,
4075 STMT_VINFO_RELATED_STMT (vinfo_for_stmt (epilog_stmt)) =
4076 STMT_VINFO_RELATED_STMT (vinfo_for_stmt (new_phi));
4079 VEC_quick_push (tree, scalar_results, new_temp);
4081 VEC_replace (tree, scalar_results, 0, new_temp);
4084 VEC_replace (tree, scalar_results, 0, new_temp);
4086 VEC_replace (gimple, new_phis, 0, epilog_stmt);
4089 /* 2.6 Handle the loop-exit phis. Replace the uses of scalar loop-exit
4090 phis with new adjusted scalar results, i.e., replace use <s_out0>
4095 s_out0 = phi <s_loop> # (scalar) EXIT_PHI
4096 v_out1 = phi <VECT_DEF> # NEW_EXIT_PHI
4097 v_out2 = reduce <v_out1>
4098 s_out3 = extract_field <v_out2, 0>
4099 s_out4 = adjust_result <s_out3>
4106 s_out0 = phi <s_loop> # (scalar) EXIT_PHI
4107 v_out1 = phi <VECT_DEF> # NEW_EXIT_PHI
4108 v_out2 = reduce <v_out1>
4109 s_out3 = extract_field <v_out2, 0>
4110 s_out4 = adjust_result <s_out3>
4115 /* In SLP reduction chain we reduce vector results into one vector if
4116 necessary, hence we set here GROUP_SIZE to 1. SCALAR_DEST is the LHS of
4117 the last stmt in the reduction chain, since we are looking for the loop
4119 if (GROUP_FIRST_ELEMENT (vinfo_for_stmt (stmt)))
4121 scalar_dest = gimple_assign_lhs (VEC_index (gimple,
4122 SLP_TREE_SCALAR_STMTS (slp_node),
4127 /* In SLP we may have several statements in NEW_PHIS and REDUCTION_PHIS (in
4128 case that GROUP_SIZE is greater than vectorization factor). Therefore, we
4129 need to match SCALAR_RESULTS with corresponding statements. The first
4130 (GROUP_SIZE / number of new vector stmts) scalar results correspond to
4131 the first vector stmt, etc.
4132 (RATIO is equal to (GROUP_SIZE / number of new vector stmts)). */
4133 if (group_size > VEC_length (gimple, new_phis))
4135 ratio = group_size / VEC_length (gimple, new_phis);
4136 gcc_assert (!(group_size % VEC_length (gimple, new_phis)));
4141 for (k = 0; k < group_size; k++)
4145 epilog_stmt = VEC_index (gimple, new_phis, k / ratio);
4146 reduction_phi = VEC_index (gimple, reduction_phis, k / ratio);
4148 inner_phi = VEC_index (gimple, inner_phis, k / ratio);
4153 gimple current_stmt = VEC_index (gimple,
4154 SLP_TREE_SCALAR_STMTS (slp_node), k);
4156 orig_stmt = STMT_VINFO_RELATED_STMT (vinfo_for_stmt (current_stmt));
4157 /* SLP statements can't participate in patterns. */
4158 gcc_assert (!orig_stmt);
4159 scalar_dest = gimple_assign_lhs (current_stmt);
4162 phis = VEC_alloc (gimple, heap, 3);
4163 /* Find the loop-closed-use at the loop exit of the original scalar
4164 result. (The reduction result is expected to have two immediate uses -
4165 one at the latch block, and one at the loop exit). */
4166 FOR_EACH_IMM_USE_FAST (use_p, imm_iter, scalar_dest)
4167 if (!flow_bb_inside_loop_p (loop, gimple_bb (USE_STMT (use_p))))
4168 VEC_safe_push (gimple, heap, phis, USE_STMT (use_p));
4170 /* We expect to have found an exit_phi because of loop-closed-ssa
4172 gcc_assert (!VEC_empty (gimple, phis));
4174 FOR_EACH_VEC_ELT (gimple, phis, i, exit_phi)
4178 stmt_vec_info exit_phi_vinfo = vinfo_for_stmt (exit_phi);
4181 /* FORNOW. Currently not supporting the case that an inner-loop
4182 reduction is not used in the outer-loop (but only outside the
4183 outer-loop), unless it is double reduction. */
4184 gcc_assert ((STMT_VINFO_RELEVANT_P (exit_phi_vinfo)
4185 && !STMT_VINFO_LIVE_P (exit_phi_vinfo))
4188 STMT_VINFO_VEC_STMT (exit_phi_vinfo) = epilog_stmt;
4190 || STMT_VINFO_DEF_TYPE (exit_phi_vinfo)
4191 != vect_double_reduction_def)
4194 /* Handle double reduction:
4196 stmt1: s1 = phi <s0, s2> - double reduction phi (outer loop)
4197 stmt2: s3 = phi <s1, s4> - (regular) reduc phi (inner loop)
4198 stmt3: s4 = use (s3) - (regular) reduc stmt (inner loop)
4199 stmt4: s2 = phi <s4> - double reduction stmt (outer loop)
4201 At that point the regular reduction (stmt2 and stmt3) is
4202 already vectorized, as well as the exit phi node, stmt4.
4203 Here we vectorize the phi node of double reduction, stmt1, and
4204 update all relevant statements. */
4206 /* Go through all the uses of s2 to find double reduction phi
4207 node, i.e., stmt1 above. */
4208 orig_name = PHI_RESULT (exit_phi);
4209 FOR_EACH_IMM_USE_STMT (use_stmt, imm_iter, orig_name)
4211 stmt_vec_info use_stmt_vinfo = vinfo_for_stmt (use_stmt);
4212 stmt_vec_info new_phi_vinfo;
4213 tree vect_phi_init, preheader_arg, vect_phi_res, init_def;
4214 basic_block bb = gimple_bb (use_stmt);
4217 /* Check that USE_STMT is really double reduction phi
4219 if (gimple_code (use_stmt) != GIMPLE_PHI
4220 || gimple_phi_num_args (use_stmt) != 2
4222 || STMT_VINFO_DEF_TYPE (use_stmt_vinfo)
4223 != vect_double_reduction_def
4224 || bb->loop_father != outer_loop)
4227 /* Create vector phi node for double reduction:
4228 vs1 = phi <vs0, vs2>
4229 vs1 was created previously in this function by a call to
4230 vect_get_vec_def_for_operand and is stored in
4232 vs2 is defined by INNER_PHI, the vectorized EXIT_PHI;
4233 vs0 is created here. */
4235 /* Create vector phi node. */
4236 vect_phi = create_phi_node (vec_initial_def, bb);
4237 new_phi_vinfo = new_stmt_vec_info (vect_phi,
4238 loop_vec_info_for_loop (outer_loop), NULL);
4239 set_vinfo_for_stmt (vect_phi, new_phi_vinfo);
4241 /* Create vs0 - initial def of the double reduction phi. */
4242 preheader_arg = PHI_ARG_DEF_FROM_EDGE (use_stmt,
4243 loop_preheader_edge (outer_loop));
4244 init_def = get_initial_def_for_reduction (stmt,
4245 preheader_arg, NULL);
4246 vect_phi_init = vect_init_vector (use_stmt, init_def,
4249 /* Update phi node arguments with vs0 and vs2. */
4250 add_phi_arg (vect_phi, vect_phi_init,
4251 loop_preheader_edge (outer_loop),
4253 add_phi_arg (vect_phi, PHI_RESULT (inner_phi),
4254 loop_latch_edge (outer_loop), UNKNOWN_LOCATION);
4255 if (vect_print_dump_info (REPORT_DETAILS))
4257 fprintf (vect_dump, "created double reduction phi "
4259 print_gimple_stmt (vect_dump, vect_phi, 0, TDF_SLIM);
4262 vect_phi_res = PHI_RESULT (vect_phi);
4264 /* Replace the use, i.e., set the correct vs1 in the regular
4265 reduction phi node. FORNOW, NCOPIES is always 1, so the
4266 loop is redundant. */
4267 use = reduction_phi;
4268 for (j = 0; j < ncopies; j++)
4270 edge pr_edge = loop_preheader_edge (loop);
4271 SET_PHI_ARG_DEF (use, pr_edge->dest_idx, vect_phi_res);
4272 use = STMT_VINFO_RELATED_STMT (vinfo_for_stmt (use));
4278 VEC_free (gimple, heap, phis);
4279 if (nested_in_vect_loop)
4287 phis = VEC_alloc (gimple, heap, 3);
4288 /* Find the loop-closed-use at the loop exit of the original scalar
4289 result. (The reduction result is expected to have two immediate uses,
4290 one at the latch block, and one at the loop exit). For double
4291 reductions we are looking for exit phis of the outer loop. */
4292 FOR_EACH_IMM_USE_FAST (use_p, imm_iter, scalar_dest)
4294 if (!flow_bb_inside_loop_p (loop, gimple_bb (USE_STMT (use_p))))
4295 VEC_safe_push (gimple, heap, phis, USE_STMT (use_p));
4298 if (double_reduc && gimple_code (USE_STMT (use_p)) == GIMPLE_PHI)
4300 tree phi_res = PHI_RESULT (USE_STMT (use_p));
4302 FOR_EACH_IMM_USE_FAST (phi_use_p, phi_imm_iter, phi_res)
4304 if (!flow_bb_inside_loop_p (loop,
4305 gimple_bb (USE_STMT (phi_use_p))))
4306 VEC_safe_push (gimple, heap, phis,
4307 USE_STMT (phi_use_p));
4313 FOR_EACH_VEC_ELT (gimple, phis, i, exit_phi)
4315 /* Replace the uses: */
4316 orig_name = PHI_RESULT (exit_phi);
4317 scalar_result = VEC_index (tree, scalar_results, k);
4318 FOR_EACH_IMM_USE_STMT (use_stmt, imm_iter, orig_name)
4319 FOR_EACH_IMM_USE_ON_STMT (use_p, imm_iter)
4320 SET_USE (use_p, scalar_result);
4323 VEC_free (gimple, heap, phis);
4326 VEC_free (tree, heap, scalar_results);
4327 VEC_free (gimple, heap, new_phis);
4331 /* Function vectorizable_reduction.
4333 Check if STMT performs a reduction operation that can be vectorized.
4334 If VEC_STMT is also passed, vectorize the STMT: create a vectorized
4335 stmt to replace it, put it in VEC_STMT, and insert it at GSI.
4336 Return FALSE if not a vectorizable STMT, TRUE otherwise.
4338 This function also handles reduction idioms (patterns) that have been
4339 recognized in advance during vect_pattern_recog. In this case, STMT may be
4341 X = pattern_expr (arg0, arg1, ..., X)
4342 and it's STMT_VINFO_RELATED_STMT points to the last stmt in the original
4343 sequence that had been detected and replaced by the pattern-stmt (STMT).
4345 In some cases of reduction patterns, the type of the reduction variable X is
4346 different than the type of the other arguments of STMT.
4347 In such cases, the vectype that is used when transforming STMT into a vector
4348 stmt is different than the vectype that is used to determine the
4349 vectorization factor, because it consists of a different number of elements
4350 than the actual number of elements that are being operated upon in parallel.
4352 For example, consider an accumulation of shorts into an int accumulator.
4353 On some targets it's possible to vectorize this pattern operating on 8
4354 shorts at a time (hence, the vectype for purposes of determining the
4355 vectorization factor should be V8HI); on the other hand, the vectype that
4356 is used to create the vector form is actually V4SI (the type of the result).
4358 Upon entry to this function, STMT_VINFO_VECTYPE records the vectype that
4359 indicates what is the actual level of parallelism (V8HI in the example), so
4360 that the right vectorization factor would be derived. This vectype
4361 corresponds to the type of arguments to the reduction stmt, and should *NOT*
4362 be used to create the vectorized stmt. The right vectype for the vectorized
4363 stmt is obtained from the type of the result X:
4364 get_vectype_for_scalar_type (TREE_TYPE (X))
4366 This means that, contrary to "regular" reductions (or "regular" stmts in
4367 general), the following equation:
4368 STMT_VINFO_VECTYPE == get_vectype_for_scalar_type (TREE_TYPE (X))
4369 does *NOT* necessarily hold for reduction patterns. */
4372 vectorizable_reduction (gimple stmt, gimple_stmt_iterator *gsi,
4373 gimple *vec_stmt, slp_tree slp_node)
4377 tree loop_vec_def0 = NULL_TREE, loop_vec_def1 = NULL_TREE;
4378 stmt_vec_info stmt_info = vinfo_for_stmt (stmt);
4379 tree vectype_out = STMT_VINFO_VECTYPE (stmt_info);
4380 tree vectype_in = NULL_TREE;
4381 loop_vec_info loop_vinfo = STMT_VINFO_LOOP_VINFO (stmt_info);
4382 struct loop *loop = LOOP_VINFO_LOOP (loop_vinfo);
4383 enum tree_code code, orig_code, epilog_reduc_code;
4384 enum machine_mode vec_mode;
4386 optab optab, reduc_optab;
4387 tree new_temp = NULL_TREE;
4390 enum vect_def_type dt;
4391 gimple new_phi = NULL;
4395 stmt_vec_info orig_stmt_info;
4396 tree expr = NULL_TREE;
4400 stmt_vec_info prev_stmt_info, prev_phi_info;
4401 bool single_defuse_cycle = false;
4402 tree reduc_def = NULL_TREE;
4403 gimple new_stmt = NULL;
4406 bool nested_cycle = false, found_nested_cycle_def = false;
4407 gimple reduc_def_stmt = NULL;
4408 /* The default is that the reduction variable is the last in statement. */
4409 int reduc_index = 2;
4410 bool double_reduc = false, dummy;
4412 struct loop * def_stmt_loop, *outer_loop = NULL;
4414 gimple def_arg_stmt;
4415 VEC (tree, heap) *vec_oprnds0 = NULL, *vec_oprnds1 = NULL, *vect_defs = NULL;
4416 VEC (gimple, heap) *phis = NULL;
4418 tree def0, def1, tem, op0, op1 = NULL_TREE;
4420 /* In case of reduction chain we switch to the first stmt in the chain, but
4421 we don't update STMT_INFO, since only the last stmt is marked as reduction
4422 and has reduction properties. */
4423 if (GROUP_FIRST_ELEMENT (vinfo_for_stmt (stmt)))
4424 stmt = GROUP_FIRST_ELEMENT (stmt_info);
4426 if (nested_in_vect_loop_p (loop, stmt))
4430 nested_cycle = true;
4433 /* 1. Is vectorizable reduction? */
4434 /* Not supportable if the reduction variable is used in the loop, unless
4435 it's a reduction chain. */
4436 if (STMT_VINFO_RELEVANT (stmt_info) > vect_used_in_outer
4437 && !GROUP_FIRST_ELEMENT (stmt_info))
4440 /* Reductions that are not used even in an enclosing outer-loop,
4441 are expected to be "live" (used out of the loop). */
4442 if (STMT_VINFO_RELEVANT (stmt_info) == vect_unused_in_scope
4443 && !STMT_VINFO_LIVE_P (stmt_info))
4446 /* Make sure it was already recognized as a reduction computation. */
4447 if (STMT_VINFO_DEF_TYPE (stmt_info) != vect_reduction_def
4448 && STMT_VINFO_DEF_TYPE (stmt_info) != vect_nested_cycle)
4451 /* 2. Has this been recognized as a reduction pattern?
4453 Check if STMT represents a pattern that has been recognized
4454 in earlier analysis stages. For stmts that represent a pattern,
4455 the STMT_VINFO_RELATED_STMT field records the last stmt in
4456 the original sequence that constitutes the pattern. */
4458 orig_stmt = STMT_VINFO_RELATED_STMT (stmt_info);
4461 orig_stmt_info = vinfo_for_stmt (orig_stmt);
4462 gcc_assert (STMT_VINFO_IN_PATTERN_P (orig_stmt_info));
4463 gcc_assert (!STMT_VINFO_IN_PATTERN_P (stmt_info));
4466 /* 3. Check the operands of the operation. The first operands are defined
4467 inside the loop body. The last operand is the reduction variable,
4468 which is defined by the loop-header-phi. */
4470 gcc_assert (is_gimple_assign (stmt));
4473 switch (get_gimple_rhs_class (gimple_assign_rhs_code (stmt)))
4475 case GIMPLE_SINGLE_RHS:
4476 op_type = TREE_OPERAND_LENGTH (gimple_assign_rhs1 (stmt));
4477 if (op_type == ternary_op)
4479 tree rhs = gimple_assign_rhs1 (stmt);
4480 ops[0] = TREE_OPERAND (rhs, 0);
4481 ops[1] = TREE_OPERAND (rhs, 1);
4482 ops[2] = TREE_OPERAND (rhs, 2);
4483 code = TREE_CODE (rhs);
4489 case GIMPLE_BINARY_RHS:
4490 code = gimple_assign_rhs_code (stmt);
4491 op_type = TREE_CODE_LENGTH (code);
4492 gcc_assert (op_type == binary_op);
4493 ops[0] = gimple_assign_rhs1 (stmt);
4494 ops[1] = gimple_assign_rhs2 (stmt);
4497 case GIMPLE_TERNARY_RHS:
4498 code = gimple_assign_rhs_code (stmt);
4499 op_type = TREE_CODE_LENGTH (code);
4500 gcc_assert (op_type == ternary_op);
4501 ops[0] = gimple_assign_rhs1 (stmt);
4502 ops[1] = gimple_assign_rhs2 (stmt);
4503 ops[2] = gimple_assign_rhs3 (stmt);
4506 case GIMPLE_UNARY_RHS:
4513 if (code == COND_EXPR && slp_node)
4516 scalar_dest = gimple_assign_lhs (stmt);
4517 scalar_type = TREE_TYPE (scalar_dest);
4518 if (!POINTER_TYPE_P (scalar_type) && !INTEGRAL_TYPE_P (scalar_type)
4519 && !SCALAR_FLOAT_TYPE_P (scalar_type))
4522 /* Do not try to vectorize bit-precision reductions. */
4523 if ((TYPE_PRECISION (scalar_type)
4524 != GET_MODE_PRECISION (TYPE_MODE (scalar_type))))
4527 /* All uses but the last are expected to be defined in the loop.
4528 The last use is the reduction variable. In case of nested cycle this
4529 assumption is not true: we use reduc_index to record the index of the
4530 reduction variable. */
4531 for (i = 0; i < op_type - 1; i++)
4533 /* The condition of COND_EXPR is checked in vectorizable_condition(). */
4534 if (i == 0 && code == COND_EXPR)
4537 is_simple_use = vect_is_simple_use_1 (ops[i], stmt, loop_vinfo, NULL,
4538 &def_stmt, &def, &dt, &tem);
4541 gcc_assert (is_simple_use);
4543 if (dt != vect_internal_def
4544 && dt != vect_external_def
4545 && dt != vect_constant_def
4546 && dt != vect_induction_def
4547 && !(dt == vect_nested_cycle && nested_cycle))
4550 if (dt == vect_nested_cycle)
4552 found_nested_cycle_def = true;
4553 reduc_def_stmt = def_stmt;
4558 is_simple_use = vect_is_simple_use_1 (ops[i], stmt, loop_vinfo, NULL,
4559 &def_stmt, &def, &dt, &tem);
4562 gcc_assert (is_simple_use);
4563 if (!(dt == vect_reduction_def
4564 || dt == vect_nested_cycle
4565 || ((dt == vect_internal_def || dt == vect_external_def
4566 || dt == vect_constant_def || dt == vect_induction_def)
4567 && nested_cycle && found_nested_cycle_def)))
4569 /* For pattern recognized stmts, orig_stmt might be a reduction,
4570 but some helper statements for the pattern might not, or
4571 might be COND_EXPRs with reduction uses in the condition. */
4572 gcc_assert (orig_stmt);
4575 if (!found_nested_cycle_def)
4576 reduc_def_stmt = def_stmt;
4578 gcc_assert (gimple_code (reduc_def_stmt) == GIMPLE_PHI);
4580 gcc_assert (orig_stmt == vect_is_simple_reduction (loop_vinfo,
4586 gimple tmp = vect_is_simple_reduction (loop_vinfo, reduc_def_stmt,
4587 !nested_cycle, &dummy);
4588 /* We changed STMT to be the first stmt in reduction chain, hence we
4589 check that in this case the first element in the chain is STMT. */
4590 gcc_assert (stmt == tmp
4591 || GROUP_FIRST_ELEMENT (vinfo_for_stmt (tmp)) == stmt);
4594 if (STMT_VINFO_LIVE_P (vinfo_for_stmt (reduc_def_stmt)))
4597 if (slp_node || PURE_SLP_STMT (stmt_info))
4600 ncopies = (LOOP_VINFO_VECT_FACTOR (loop_vinfo)
4601 / TYPE_VECTOR_SUBPARTS (vectype_in));
4603 gcc_assert (ncopies >= 1);
4605 vec_mode = TYPE_MODE (vectype_in);
4607 if (code == COND_EXPR)
4609 if (!vectorizable_condition (stmt, gsi, NULL, ops[reduc_index], 0, NULL))
4611 if (vect_print_dump_info (REPORT_DETAILS))
4612 fprintf (vect_dump, "unsupported condition in reduction");
4619 /* 4. Supportable by target? */
4621 /* 4.1. check support for the operation in the loop */
4622 optab = optab_for_tree_code (code, vectype_in, optab_default);
4625 if (vect_print_dump_info (REPORT_DETAILS))
4626 fprintf (vect_dump, "no optab.");
4631 if (optab_handler (optab, vec_mode) == CODE_FOR_nothing)
4633 if (vect_print_dump_info (REPORT_DETAILS))
4634 fprintf (vect_dump, "op not supported by target.");
4636 if (GET_MODE_SIZE (vec_mode) != UNITS_PER_WORD
4637 || LOOP_VINFO_VECT_FACTOR (loop_vinfo)
4638 < vect_min_worthwhile_factor (code))
4641 if (vect_print_dump_info (REPORT_DETAILS))
4642 fprintf (vect_dump, "proceeding using word mode.");
4645 /* Worthwhile without SIMD support? */
4646 if (!VECTOR_MODE_P (TYPE_MODE (vectype_in))
4647 && LOOP_VINFO_VECT_FACTOR (loop_vinfo)
4648 < vect_min_worthwhile_factor (code))
4650 if (vect_print_dump_info (REPORT_DETAILS))
4651 fprintf (vect_dump, "not worthwhile without SIMD support.");
4657 /* 4.2. Check support for the epilog operation.
4659 If STMT represents a reduction pattern, then the type of the
4660 reduction variable may be different than the type of the rest
4661 of the arguments. For example, consider the case of accumulation
4662 of shorts into an int accumulator; The original code:
4663 S1: int_a = (int) short_a;
4664 orig_stmt-> S2: int_acc = plus <int_a ,int_acc>;
4667 STMT: int_acc = widen_sum <short_a, int_acc>
4670 1. The tree-code that is used to create the vector operation in the
4671 epilog code (that reduces the partial results) is not the
4672 tree-code of STMT, but is rather the tree-code of the original
4673 stmt from the pattern that STMT is replacing. I.e, in the example
4674 above we want to use 'widen_sum' in the loop, but 'plus' in the
4676 2. The type (mode) we use to check available target support
4677 for the vector operation to be created in the *epilog*, is
4678 determined by the type of the reduction variable (in the example
4679 above we'd check this: optab_handler (plus_optab, vect_int_mode])).
4680 However the type (mode) we use to check available target support
4681 for the vector operation to be created *inside the loop*, is
4682 determined by the type of the other arguments to STMT (in the
4683 example we'd check this: optab_handler (widen_sum_optab,
4686 This is contrary to "regular" reductions, in which the types of all
4687 the arguments are the same as the type of the reduction variable.
4688 For "regular" reductions we can therefore use the same vector type
4689 (and also the same tree-code) when generating the epilog code and
4690 when generating the code inside the loop. */
4694 /* This is a reduction pattern: get the vectype from the type of the
4695 reduction variable, and get the tree-code from orig_stmt. */
4696 orig_code = gimple_assign_rhs_code (orig_stmt);
4697 gcc_assert (vectype_out);
4698 vec_mode = TYPE_MODE (vectype_out);
4702 /* Regular reduction: use the same vectype and tree-code as used for
4703 the vector code inside the loop can be used for the epilog code. */
4709 def_bb = gimple_bb (reduc_def_stmt);
4710 def_stmt_loop = def_bb->loop_father;
4711 def_arg = PHI_ARG_DEF_FROM_EDGE (reduc_def_stmt,
4712 loop_preheader_edge (def_stmt_loop));
4713 if (TREE_CODE (def_arg) == SSA_NAME
4714 && (def_arg_stmt = SSA_NAME_DEF_STMT (def_arg))
4715 && gimple_code (def_arg_stmt) == GIMPLE_PHI
4716 && flow_bb_inside_loop_p (outer_loop, gimple_bb (def_arg_stmt))
4717 && vinfo_for_stmt (def_arg_stmt)
4718 && STMT_VINFO_DEF_TYPE (vinfo_for_stmt (def_arg_stmt))
4719 == vect_double_reduction_def)
4720 double_reduc = true;
4723 epilog_reduc_code = ERROR_MARK;
4724 if (reduction_code_for_scalar_code (orig_code, &epilog_reduc_code))
4726 reduc_optab = optab_for_tree_code (epilog_reduc_code, vectype_out,
4730 if (vect_print_dump_info (REPORT_DETAILS))
4731 fprintf (vect_dump, "no optab for reduction.");
4733 epilog_reduc_code = ERROR_MARK;
4737 && optab_handler (reduc_optab, vec_mode) == CODE_FOR_nothing)
4739 if (vect_print_dump_info (REPORT_DETAILS))
4740 fprintf (vect_dump, "reduc op not supported by target.");
4742 epilog_reduc_code = ERROR_MARK;
4747 if (!nested_cycle || double_reduc)
4749 if (vect_print_dump_info (REPORT_DETAILS))
4750 fprintf (vect_dump, "no reduc code for scalar code.");
4756 if (double_reduc && ncopies > 1)
4758 if (vect_print_dump_info (REPORT_DETAILS))
4759 fprintf (vect_dump, "multiple types in double reduction");
4764 /* In case of widenning multiplication by a constant, we update the type
4765 of the constant to be the type of the other operand. We check that the
4766 constant fits the type in the pattern recognition pass. */
4767 if (code == DOT_PROD_EXPR
4768 && !types_compatible_p (TREE_TYPE (ops[0]), TREE_TYPE (ops[1])))
4770 if (TREE_CODE (ops[0]) == INTEGER_CST)
4771 ops[0] = fold_convert (TREE_TYPE (ops[1]), ops[0]);
4772 else if (TREE_CODE (ops[1]) == INTEGER_CST)
4773 ops[1] = fold_convert (TREE_TYPE (ops[0]), ops[1]);
4776 if (vect_print_dump_info (REPORT_DETAILS))
4777 fprintf (vect_dump, "invalid types in dot-prod");
4783 if (!vec_stmt) /* transformation not required. */
4785 if (!vect_model_reduction_cost (stmt_info, epilog_reduc_code, ncopies))
4787 STMT_VINFO_TYPE (stmt_info) = reduc_vec_info_type;
4793 if (vect_print_dump_info (REPORT_DETAILS))
4794 fprintf (vect_dump, "transform reduction.");
4796 /* FORNOW: Multiple types are not supported for condition. */
4797 if (code == COND_EXPR)
4798 gcc_assert (ncopies == 1);
4800 /* Create the destination vector */
4801 vec_dest = vect_create_destination_var (scalar_dest, vectype_out);
4803 /* In case the vectorization factor (VF) is bigger than the number
4804 of elements that we can fit in a vectype (nunits), we have to generate
4805 more than one vector stmt - i.e - we need to "unroll" the
4806 vector stmt by a factor VF/nunits. For more details see documentation
4807 in vectorizable_operation. */
4809 /* If the reduction is used in an outer loop we need to generate
4810 VF intermediate results, like so (e.g. for ncopies=2):
4815 (i.e. we generate VF results in 2 registers).
4816 In this case we have a separate def-use cycle for each copy, and therefore
4817 for each copy we get the vector def for the reduction variable from the
4818 respective phi node created for this copy.
4820 Otherwise (the reduction is unused in the loop nest), we can combine
4821 together intermediate results, like so (e.g. for ncopies=2):
4825 (i.e. we generate VF/2 results in a single register).
4826 In this case for each copy we get the vector def for the reduction variable
4827 from the vectorized reduction operation generated in the previous iteration.
4830 if (STMT_VINFO_RELEVANT (stmt_info) == vect_unused_in_scope)
4832 single_defuse_cycle = true;
4836 epilog_copies = ncopies;
4838 prev_stmt_info = NULL;
4839 prev_phi_info = NULL;
4842 vec_num = SLP_TREE_NUMBER_OF_VEC_STMTS (slp_node);
4843 gcc_assert (TYPE_VECTOR_SUBPARTS (vectype_out)
4844 == TYPE_VECTOR_SUBPARTS (vectype_in));
4849 vec_oprnds0 = VEC_alloc (tree, heap, 1);
4850 if (op_type == ternary_op)
4851 vec_oprnds1 = VEC_alloc (tree, heap, 1);
4854 phis = VEC_alloc (gimple, heap, vec_num);
4855 vect_defs = VEC_alloc (tree, heap, vec_num);
4857 VEC_quick_push (tree, vect_defs, NULL_TREE);
4859 for (j = 0; j < ncopies; j++)
4861 if (j == 0 || !single_defuse_cycle)
4863 for (i = 0; i < vec_num; i++)
4865 /* Create the reduction-phi that defines the reduction
4867 new_phi = create_phi_node (vec_dest, loop->header);
4868 set_vinfo_for_stmt (new_phi,
4869 new_stmt_vec_info (new_phi, loop_vinfo,
4871 if (j == 0 || slp_node)
4872 VEC_quick_push (gimple, phis, new_phi);
4876 if (code == COND_EXPR)
4878 gcc_assert (!slp_node);
4879 vectorizable_condition (stmt, gsi, vec_stmt,
4880 PHI_RESULT (VEC_index (gimple, phis, 0)),
4882 /* Multiple types are not supported for condition. */
4889 op0 = ops[!reduc_index];
4890 if (op_type == ternary_op)
4892 if (reduc_index == 0)
4899 vect_get_vec_defs (op0, op1, stmt, &vec_oprnds0, &vec_oprnds1,
4903 loop_vec_def0 = vect_get_vec_def_for_operand (ops[!reduc_index],
4905 VEC_quick_push (tree, vec_oprnds0, loop_vec_def0);
4906 if (op_type == ternary_op)
4908 loop_vec_def1 = vect_get_vec_def_for_operand (op1, stmt,
4910 VEC_quick_push (tree, vec_oprnds1, loop_vec_def1);
4918 enum vect_def_type dt;
4922 vect_is_simple_use (ops[!reduc_index], stmt, loop_vinfo, NULL,
4923 &dummy_stmt, &dummy, &dt);
4924 loop_vec_def0 = vect_get_vec_def_for_stmt_copy (dt,
4926 VEC_replace (tree, vec_oprnds0, 0, loop_vec_def0);
4927 if (op_type == ternary_op)
4929 vect_is_simple_use (op1, stmt, loop_vinfo, NULL, &dummy_stmt,
4931 loop_vec_def1 = vect_get_vec_def_for_stmt_copy (dt,
4933 VEC_replace (tree, vec_oprnds1, 0, loop_vec_def1);
4937 if (single_defuse_cycle)
4938 reduc_def = gimple_assign_lhs (new_stmt);
4940 STMT_VINFO_RELATED_STMT (prev_phi_info) = new_phi;
4943 FOR_EACH_VEC_ELT (tree, vec_oprnds0, i, def0)
4946 reduc_def = PHI_RESULT (VEC_index (gimple, phis, i));
4949 if (!single_defuse_cycle || j == 0)
4950 reduc_def = PHI_RESULT (new_phi);
4953 def1 = ((op_type == ternary_op)
4954 ? VEC_index (tree, vec_oprnds1, i) : NULL);
4955 if (op_type == binary_op)
4957 if (reduc_index == 0)
4958 expr = build2 (code, vectype_out, reduc_def, def0);
4960 expr = build2 (code, vectype_out, def0, reduc_def);
4964 if (reduc_index == 0)
4965 expr = build3 (code, vectype_out, reduc_def, def0, def1);
4968 if (reduc_index == 1)
4969 expr = build3 (code, vectype_out, def0, reduc_def, def1);
4971 expr = build3 (code, vectype_out, def0, def1, reduc_def);
4975 new_stmt = gimple_build_assign (vec_dest, expr);
4976 new_temp = make_ssa_name (vec_dest, new_stmt);
4977 gimple_assign_set_lhs (new_stmt, new_temp);
4978 vect_finish_stmt_generation (stmt, new_stmt, gsi);
4982 VEC_quick_push (gimple, SLP_TREE_VEC_STMTS (slp_node), new_stmt);
4983 VEC_quick_push (tree, vect_defs, new_temp);
4986 VEC_replace (tree, vect_defs, 0, new_temp);
4993 STMT_VINFO_VEC_STMT (stmt_info) = *vec_stmt = new_stmt;
4995 STMT_VINFO_RELATED_STMT (prev_stmt_info) = new_stmt;
4997 prev_stmt_info = vinfo_for_stmt (new_stmt);
4998 prev_phi_info = vinfo_for_stmt (new_phi);
5001 /* Finalize the reduction-phi (set its arguments) and create the
5002 epilog reduction code. */
5003 if ((!single_defuse_cycle || code == COND_EXPR) && !slp_node)
5005 new_temp = gimple_assign_lhs (*vec_stmt);
5006 VEC_replace (tree, vect_defs, 0, new_temp);
5009 vect_create_epilog_for_reduction (vect_defs, stmt, epilog_copies,
5010 epilog_reduc_code, phis, reduc_index,
5011 double_reduc, slp_node);
5013 VEC_free (gimple, heap, phis);
5014 VEC_free (tree, heap, vec_oprnds0);
5016 VEC_free (tree, heap, vec_oprnds1);
5021 /* Function vect_min_worthwhile_factor.
5023 For a loop where we could vectorize the operation indicated by CODE,
5024 return the minimum vectorization factor that makes it worthwhile
5025 to use generic vectors. */
5027 vect_min_worthwhile_factor (enum tree_code code)
5048 /* Function vectorizable_induction
5050 Check if PHI performs an induction computation that can be vectorized.
5051 If VEC_STMT is also passed, vectorize the induction PHI: create a vectorized
5052 phi to replace it, put it in VEC_STMT, and add it to the same basic block.
5053 Return FALSE if not a vectorizable STMT, TRUE otherwise. */
5056 vectorizable_induction (gimple phi, gimple_stmt_iterator *gsi ATTRIBUTE_UNUSED,
5059 stmt_vec_info stmt_info = vinfo_for_stmt (phi);
5060 tree vectype = STMT_VINFO_VECTYPE (stmt_info);
5061 loop_vec_info loop_vinfo = STMT_VINFO_LOOP_VINFO (stmt_info);
5062 struct loop *loop = LOOP_VINFO_LOOP (loop_vinfo);
5063 int nunits = TYPE_VECTOR_SUBPARTS (vectype);
5064 int ncopies = LOOP_VINFO_VECT_FACTOR (loop_vinfo) / nunits;
5067 gcc_assert (ncopies >= 1);
5068 /* FORNOW. These restrictions should be relaxed. */
5069 if (nested_in_vect_loop_p (loop, phi))
5071 imm_use_iterator imm_iter;
5072 use_operand_p use_p;
5079 if (vect_print_dump_info (REPORT_DETAILS))
5080 fprintf (vect_dump, "multiple types in nested loop.");
5085 latch_e = loop_latch_edge (loop->inner);
5086 loop_arg = PHI_ARG_DEF_FROM_EDGE (phi, latch_e);
5087 FOR_EACH_IMM_USE_FAST (use_p, imm_iter, loop_arg)
5089 if (!flow_bb_inside_loop_p (loop->inner,
5090 gimple_bb (USE_STMT (use_p))))
5092 exit_phi = USE_STMT (use_p);
5098 stmt_vec_info exit_phi_vinfo = vinfo_for_stmt (exit_phi);
5099 if (!(STMT_VINFO_RELEVANT_P (exit_phi_vinfo)
5100 && !STMT_VINFO_LIVE_P (exit_phi_vinfo)))
5102 if (vect_print_dump_info (REPORT_DETAILS))
5103 fprintf (vect_dump, "inner-loop induction only used outside "
5104 "of the outer vectorized loop.");
5110 if (!STMT_VINFO_RELEVANT_P (stmt_info))
5113 /* FORNOW: SLP not supported. */
5114 if (STMT_SLP_TYPE (stmt_info))
5117 gcc_assert (STMT_VINFO_DEF_TYPE (stmt_info) == vect_induction_def);
5119 if (gimple_code (phi) != GIMPLE_PHI)
5122 if (!vec_stmt) /* transformation not required. */
5124 STMT_VINFO_TYPE (stmt_info) = induc_vec_info_type;
5125 if (vect_print_dump_info (REPORT_DETAILS))
5126 fprintf (vect_dump, "=== vectorizable_induction ===");
5127 vect_model_induction_cost (stmt_info, ncopies);
5133 if (vect_print_dump_info (REPORT_DETAILS))
5134 fprintf (vect_dump, "transform induction phi.");
5136 vec_def = get_initial_def_for_induction (phi);
5137 *vec_stmt = SSA_NAME_DEF_STMT (vec_def);
5141 /* Function vectorizable_live_operation.
5143 STMT computes a value that is used outside the loop. Check if
5144 it can be supported. */
5147 vectorizable_live_operation (gimple stmt,
5148 gimple_stmt_iterator *gsi ATTRIBUTE_UNUSED,
5149 gimple *vec_stmt ATTRIBUTE_UNUSED)
5151 stmt_vec_info stmt_info = vinfo_for_stmt (stmt);
5152 loop_vec_info loop_vinfo = STMT_VINFO_LOOP_VINFO (stmt_info);
5153 struct loop *loop = LOOP_VINFO_LOOP (loop_vinfo);
5159 enum vect_def_type dt;
5160 enum tree_code code;
5161 enum gimple_rhs_class rhs_class;
5163 gcc_assert (STMT_VINFO_LIVE_P (stmt_info));
5165 if (STMT_VINFO_DEF_TYPE (stmt_info) == vect_reduction_def)
5168 if (!is_gimple_assign (stmt))
5171 if (TREE_CODE (gimple_assign_lhs (stmt)) != SSA_NAME)
5174 /* FORNOW. CHECKME. */
5175 if (nested_in_vect_loop_p (loop, stmt))
5178 code = gimple_assign_rhs_code (stmt);
5179 op_type = TREE_CODE_LENGTH (code);
5180 rhs_class = get_gimple_rhs_class (code);
5181 gcc_assert (rhs_class != GIMPLE_UNARY_RHS || op_type == unary_op);
5182 gcc_assert (rhs_class != GIMPLE_BINARY_RHS || op_type == binary_op);
5184 /* FORNOW: support only if all uses are invariant. This means
5185 that the scalar operations can remain in place, unvectorized.
5186 The original last scalar value that they compute will be used. */
5188 for (i = 0; i < op_type; i++)
5190 if (rhs_class == GIMPLE_SINGLE_RHS)
5191 op = TREE_OPERAND (gimple_op (stmt, 1), i);
5193 op = gimple_op (stmt, i + 1);
5195 && !vect_is_simple_use (op, stmt, loop_vinfo, NULL, &def_stmt, &def,
5198 if (vect_print_dump_info (REPORT_DETAILS))
5199 fprintf (vect_dump, "use not simple.");
5203 if (dt != vect_external_def && dt != vect_constant_def)
5207 /* No transformation is required for the cases we currently support. */
5211 /* Kill any debug uses outside LOOP of SSA names defined in STMT. */
5214 vect_loop_kill_debug_uses (struct loop *loop, gimple stmt)
5216 ssa_op_iter op_iter;
5217 imm_use_iterator imm_iter;
5218 def_operand_p def_p;
5221 FOR_EACH_PHI_OR_STMT_DEF (def_p, stmt, op_iter, SSA_OP_DEF)
5223 FOR_EACH_IMM_USE_STMT (ustmt, imm_iter, DEF_FROM_PTR (def_p))
5227 if (!is_gimple_debug (ustmt))
5230 bb = gimple_bb (ustmt);
5232 if (!flow_bb_inside_loop_p (loop, bb))
5234 if (gimple_debug_bind_p (ustmt))
5236 if (vect_print_dump_info (REPORT_DETAILS))
5237 fprintf (vect_dump, "killing debug use");
5239 gimple_debug_bind_reset_value (ustmt);
5240 update_stmt (ustmt);
5249 /* Function vect_transform_loop.
5251 The analysis phase has determined that the loop is vectorizable.
5252 Vectorize the loop - created vectorized stmts to replace the scalar
5253 stmts in the loop, and update the loop exit condition. */
5256 vect_transform_loop (loop_vec_info loop_vinfo)
5258 struct loop *loop = LOOP_VINFO_LOOP (loop_vinfo);
5259 basic_block *bbs = LOOP_VINFO_BBS (loop_vinfo);
5260 int nbbs = loop->num_nodes;
5261 gimple_stmt_iterator si;
5264 int vectorization_factor = LOOP_VINFO_VECT_FACTOR (loop_vinfo);
5266 bool slp_scheduled = false;
5267 unsigned int nunits;
5268 tree cond_expr = NULL_TREE;
5269 gimple_seq cond_expr_stmt_list = NULL;
5270 bool do_peeling_for_loop_bound;
5271 gimple stmt, pattern_stmt;
5272 gimple_seq pattern_def_seq = NULL;
5273 gimple_stmt_iterator pattern_def_si = gsi_start (NULL);
5274 bool transform_pattern_stmt = false;
5276 if (vect_print_dump_info (REPORT_DETAILS))
5277 fprintf (vect_dump, "=== vec_transform_loop ===");
5279 /* Peel the loop if there are data refs with unknown alignment.
5280 Only one data ref with unknown store is allowed. */
5282 if (LOOP_PEELING_FOR_ALIGNMENT (loop_vinfo))
5283 vect_do_peeling_for_alignment (loop_vinfo);
5285 do_peeling_for_loop_bound
5286 = (!LOOP_VINFO_NITERS_KNOWN_P (loop_vinfo)
5287 || (LOOP_VINFO_NITERS_KNOWN_P (loop_vinfo)
5288 && LOOP_VINFO_INT_NITERS (loop_vinfo) % vectorization_factor != 0)
5289 || LOOP_VINFO_PEELING_FOR_GAPS (loop_vinfo));
5291 if (LOOP_REQUIRES_VERSIONING_FOR_ALIGNMENT (loop_vinfo)
5292 || LOOP_REQUIRES_VERSIONING_FOR_ALIAS (loop_vinfo))
5293 vect_loop_versioning (loop_vinfo,
5294 !do_peeling_for_loop_bound,
5295 &cond_expr, &cond_expr_stmt_list);
5297 /* If the loop has a symbolic number of iterations 'n' (i.e. it's not a
5298 compile time constant), or it is a constant that doesn't divide by the
5299 vectorization factor, then an epilog loop needs to be created.
5300 We therefore duplicate the loop: the original loop will be vectorized,
5301 and will compute the first (n/VF) iterations. The second copy of the loop
5302 will remain scalar and will compute the remaining (n%VF) iterations.
5303 (VF is the vectorization factor). */
5305 if (do_peeling_for_loop_bound)
5306 vect_do_peeling_for_loop_bound (loop_vinfo, &ratio,
5307 cond_expr, cond_expr_stmt_list);
5309 ratio = build_int_cst (TREE_TYPE (LOOP_VINFO_NITERS (loop_vinfo)),
5310 LOOP_VINFO_INT_NITERS (loop_vinfo) / vectorization_factor);
5312 /* 1) Make sure the loop header has exactly two entries
5313 2) Make sure we have a preheader basic block. */
5315 gcc_assert (EDGE_COUNT (loop->header->preds) == 2);
5317 split_edge (loop_preheader_edge (loop));
5319 /* FORNOW: the vectorizer supports only loops which body consist
5320 of one basic block (header + empty latch). When the vectorizer will
5321 support more involved loop forms, the order by which the BBs are
5322 traversed need to be reconsidered. */
5324 for (i = 0; i < nbbs; i++)
5326 basic_block bb = bbs[i];
5327 stmt_vec_info stmt_info;
5330 for (si = gsi_start_phis (bb); !gsi_end_p (si); gsi_next (&si))
5332 phi = gsi_stmt (si);
5333 if (vect_print_dump_info (REPORT_DETAILS))
5335 fprintf (vect_dump, "------>vectorizing phi: ");
5336 print_gimple_stmt (vect_dump, phi, 0, TDF_SLIM);
5338 stmt_info = vinfo_for_stmt (phi);
5342 if (MAY_HAVE_DEBUG_STMTS && !STMT_VINFO_LIVE_P (stmt_info))
5343 vect_loop_kill_debug_uses (loop, phi);
5345 if (!STMT_VINFO_RELEVANT_P (stmt_info)
5346 && !STMT_VINFO_LIVE_P (stmt_info))
5349 if ((TYPE_VECTOR_SUBPARTS (STMT_VINFO_VECTYPE (stmt_info))
5350 != (unsigned HOST_WIDE_INT) vectorization_factor)
5351 && vect_print_dump_info (REPORT_DETAILS))
5352 fprintf (vect_dump, "multiple-types.");
5354 if (STMT_VINFO_DEF_TYPE (stmt_info) == vect_induction_def)
5356 if (vect_print_dump_info (REPORT_DETAILS))
5357 fprintf (vect_dump, "transform phi.");
5358 vect_transform_stmt (phi, NULL, NULL, NULL, NULL);
5362 pattern_stmt = NULL;
5363 for (si = gsi_start_bb (bb); !gsi_end_p (si) || transform_pattern_stmt;)
5367 if (transform_pattern_stmt)
5368 stmt = pattern_stmt;
5370 stmt = gsi_stmt (si);
5372 if (vect_print_dump_info (REPORT_DETAILS))
5374 fprintf (vect_dump, "------>vectorizing statement: ");
5375 print_gimple_stmt (vect_dump, stmt, 0, TDF_SLIM);
5378 stmt_info = vinfo_for_stmt (stmt);
5380 /* vector stmts created in the outer-loop during vectorization of
5381 stmts in an inner-loop may not have a stmt_info, and do not
5382 need to be vectorized. */
5389 if (MAY_HAVE_DEBUG_STMTS && !STMT_VINFO_LIVE_P (stmt_info))
5390 vect_loop_kill_debug_uses (loop, stmt);
5392 if (!STMT_VINFO_RELEVANT_P (stmt_info)
5393 && !STMT_VINFO_LIVE_P (stmt_info))
5395 if (STMT_VINFO_IN_PATTERN_P (stmt_info)
5396 && (pattern_stmt = STMT_VINFO_RELATED_STMT (stmt_info))
5397 && (STMT_VINFO_RELEVANT_P (vinfo_for_stmt (pattern_stmt))
5398 || STMT_VINFO_LIVE_P (vinfo_for_stmt (pattern_stmt))))
5400 stmt = pattern_stmt;
5401 stmt_info = vinfo_for_stmt (stmt);
5409 else if (STMT_VINFO_IN_PATTERN_P (stmt_info)
5410 && (pattern_stmt = STMT_VINFO_RELATED_STMT (stmt_info))
5411 && (STMT_VINFO_RELEVANT_P (vinfo_for_stmt (pattern_stmt))
5412 || STMT_VINFO_LIVE_P (vinfo_for_stmt (pattern_stmt))))
5413 transform_pattern_stmt = true;
5415 /* If pattern statement has def stmts, vectorize them too. */
5416 if (is_pattern_stmt_p (stmt_info))
5418 if (pattern_def_seq == NULL)
5420 pattern_def_seq = STMT_VINFO_PATTERN_DEF_SEQ (stmt_info);
5421 pattern_def_si = gsi_start (pattern_def_seq);
5423 else if (!gsi_end_p (pattern_def_si))
5424 gsi_next (&pattern_def_si);
5425 if (pattern_def_seq != NULL)
5427 gimple pattern_def_stmt = NULL;
5428 stmt_vec_info pattern_def_stmt_info = NULL;
5430 while (!gsi_end_p (pattern_def_si))
5432 pattern_def_stmt = gsi_stmt (pattern_def_si);
5433 pattern_def_stmt_info
5434 = vinfo_for_stmt (pattern_def_stmt);
5435 if (STMT_VINFO_RELEVANT_P (pattern_def_stmt_info)
5436 || STMT_VINFO_LIVE_P (pattern_def_stmt_info))
5438 gsi_next (&pattern_def_si);
5441 if (!gsi_end_p (pattern_def_si))
5443 if (vect_print_dump_info (REPORT_DETAILS))
5445 fprintf (vect_dump, "==> vectorizing pattern def"
5447 print_gimple_stmt (vect_dump, pattern_def_stmt, 0,
5451 stmt = pattern_def_stmt;
5452 stmt_info = pattern_def_stmt_info;
5456 pattern_def_si = gsi_start (NULL);
5457 transform_pattern_stmt = false;
5461 transform_pattern_stmt = false;
5464 gcc_assert (STMT_VINFO_VECTYPE (stmt_info));
5465 nunits = (unsigned int) TYPE_VECTOR_SUBPARTS (
5466 STMT_VINFO_VECTYPE (stmt_info));
5467 if (!STMT_SLP_TYPE (stmt_info)
5468 && nunits != (unsigned int) vectorization_factor
5469 && vect_print_dump_info (REPORT_DETAILS))
5470 /* For SLP VF is set according to unrolling factor, and not to
5471 vector size, hence for SLP this print is not valid. */
5472 fprintf (vect_dump, "multiple-types.");
5474 /* SLP. Schedule all the SLP instances when the first SLP stmt is
5476 if (STMT_SLP_TYPE (stmt_info))
5480 slp_scheduled = true;
5482 if (vect_print_dump_info (REPORT_DETAILS))
5483 fprintf (vect_dump, "=== scheduling SLP instances ===");
5485 vect_schedule_slp (loop_vinfo, NULL);
5488 /* Hybrid SLP stmts must be vectorized in addition to SLP. */
5489 if (!vinfo_for_stmt (stmt) || PURE_SLP_STMT (stmt_info))
5491 if (!transform_pattern_stmt && gsi_end_p (pattern_def_si))
5493 pattern_def_seq = NULL;
5500 /* -------- vectorize statement ------------ */
5501 if (vect_print_dump_info (REPORT_DETAILS))
5502 fprintf (vect_dump, "transform statement.");
5504 strided_store = false;
5505 is_store = vect_transform_stmt (stmt, &si, &strided_store, NULL, NULL);
5508 if (STMT_VINFO_STRIDED_ACCESS (stmt_info))
5510 /* Interleaving. If IS_STORE is TRUE, the vectorization of the
5511 interleaving chain was completed - free all the stores in
5514 vect_remove_stores (GROUP_FIRST_ELEMENT (stmt_info));
5519 /* Free the attached stmt_vec_info and remove the stmt. */
5520 free_stmt_vec_info (gsi_stmt (si));
5521 gsi_remove (&si, true);
5526 if (!transform_pattern_stmt && gsi_end_p (pattern_def_si))
5528 pattern_def_seq = NULL;
5534 slpeel_make_loop_iterate_ntimes (loop, ratio);
5536 /* The memory tags and pointers in vectorized statements need to
5537 have their SSA forms updated. FIXME, why can't this be delayed
5538 until all the loops have been transformed? */
5539 update_ssa (TODO_update_ssa);
5541 if (vect_print_dump_info (REPORT_VECTORIZED_LOCATIONS))
5542 fprintf (vect_dump, "LOOP VECTORIZED.");
5543 if (loop->inner && vect_print_dump_info (REPORT_VECTORIZED_LOCATIONS))
5544 fprintf (vect_dump, "OUTER LOOP VECTORIZED.");