2016-10-06 Richard Biener <rguenther@suse.de>
[official-gcc.git] / gcc / tree-vect-patterns.c
blob3dfbc7bb63dece204986826f58cecef5faa87a86
1 /* Analysis Utilities for Loop Vectorization.
2 Copyright (C) 2006-2016 Free Software Foundation, Inc.
3 Contributed by Dorit Nuzman <dorit@il.ibm.com>
5 This file is part of GCC.
7 GCC is free software; you can redistribute it and/or modify it under
8 the terms of the GNU General Public License as published by the Free
9 Software Foundation; either version 3, or (at your option) any later
10 version.
12 GCC is distributed in the hope that it will be useful, but WITHOUT ANY
13 WARRANTY; without even the implied warranty of MERCHANTABILITY or
14 FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
15 for more details.
17 You should have received a copy of the GNU General Public License
18 along with GCC; see the file COPYING3. If not see
19 <http://www.gnu.org/licenses/>. */
21 #include "config.h"
22 #include "system.h"
23 #include "coretypes.h"
24 #include "backend.h"
25 #include "rtl.h"
26 #include "tree.h"
27 #include "gimple.h"
28 #include "ssa.h"
29 #include "expmed.h"
30 #include "optabs-tree.h"
31 #include "insn-config.h"
32 #include "recog.h" /* FIXME: for insn_data */
33 #include "fold-const.h"
34 #include "stor-layout.h"
35 #include "tree-eh.h"
36 #include "gimplify.h"
37 #include "gimple-iterator.h"
38 #include "cfgloop.h"
39 #include "tree-vectorizer.h"
40 #include "dumpfile.h"
41 #include "builtins.h"
42 #include "internal-fn.h"
43 #include "case-cfn-macros.h"
45 /* Pattern recognition functions */
46 static gimple *vect_recog_widen_sum_pattern (vec<gimple *> *, tree *,
47 tree *);
48 static gimple *vect_recog_widen_mult_pattern (vec<gimple *> *, tree *,
49 tree *);
50 static gimple *vect_recog_dot_prod_pattern (vec<gimple *> *, tree *,
51 tree *);
52 static gimple *vect_recog_sad_pattern (vec<gimple *> *, tree *,
53 tree *);
54 static gimple *vect_recog_pow_pattern (vec<gimple *> *, tree *, tree *);
55 static gimple *vect_recog_over_widening_pattern (vec<gimple *> *, tree *,
56 tree *);
57 static gimple *vect_recog_widen_shift_pattern (vec<gimple *> *,
58 tree *, tree *);
59 static gimple *vect_recog_rotate_pattern (vec<gimple *> *, tree *, tree *);
60 static gimple *vect_recog_vector_vector_shift_pattern (vec<gimple *> *,
61 tree *, tree *);
62 static gimple *vect_recog_divmod_pattern (vec<gimple *> *,
63 tree *, tree *);
65 static gimple *vect_recog_mult_pattern (vec<gimple *> *,
66 tree *, tree *);
68 static gimple *vect_recog_mixed_size_cond_pattern (vec<gimple *> *,
69 tree *, tree *);
70 static gimple *vect_recog_bool_pattern (vec<gimple *> *, tree *, tree *);
71 static gimple *vect_recog_mask_conversion_pattern (vec<gimple *> *, tree *, tree *);
73 struct vect_recog_func
75 vect_recog_func_ptr fn;
76 const char *name;
79 /* Note that ordering matters - the first pattern matching on a stmt
80 is taken which means usually the more complex one needs to preceed
81 the less comples onex (widen_sum only after dot_prod or sad for example). */
82 static vect_recog_func vect_vect_recog_func_ptrs[NUM_PATTERNS] = {
83 { vect_recog_widen_mult_pattern, "widen_mult" },
84 { vect_recog_dot_prod_pattern, "dot_prod" },
85 { vect_recog_sad_pattern, "sad" },
86 { vect_recog_widen_sum_pattern, "widen_sum" },
87 { vect_recog_pow_pattern, "pow" },
88 { vect_recog_widen_shift_pattern, "widen_shift" },
89 { vect_recog_over_widening_pattern, "over_widening" },
90 { vect_recog_rotate_pattern, "rotate" },
91 { vect_recog_vector_vector_shift_pattern, "vector_vector_shift" },
92 { vect_recog_divmod_pattern, "divmod" },
93 { vect_recog_mult_pattern, "mult" },
94 { vect_recog_mixed_size_cond_pattern, "mixed_size_cond" },
95 { vect_recog_bool_pattern, "bool" },
96 { vect_recog_mask_conversion_pattern, "mask_conversion" }
99 static inline void
100 append_pattern_def_seq (stmt_vec_info stmt_info, gimple *stmt)
102 gimple_seq_add_stmt_without_update (&STMT_VINFO_PATTERN_DEF_SEQ (stmt_info),
103 stmt);
106 static inline void
107 new_pattern_def_seq (stmt_vec_info stmt_info, gimple *stmt)
109 STMT_VINFO_PATTERN_DEF_SEQ (stmt_info) = NULL;
110 append_pattern_def_seq (stmt_info, stmt);
113 /* Check whether STMT2 is in the same loop or basic block as STMT1.
114 Which of the two applies depends on whether we're currently doing
115 loop-based or basic-block-based vectorization, as determined by
116 the vinfo_for_stmt for STMT1 (which must be defined).
118 If this returns true, vinfo_for_stmt for STMT2 is guaranteed
119 to be defined as well. */
121 static bool
122 vect_same_loop_or_bb_p (gimple *stmt1, gimple *stmt2)
124 stmt_vec_info stmt_vinfo = vinfo_for_stmt (stmt1);
125 return vect_stmt_in_region_p (stmt_vinfo->vinfo, stmt2);
128 /* If the LHS of DEF_STMT has a single use, and that statement is
129 in the same loop or basic block, return it. */
131 static gimple *
132 vect_single_imm_use (gimple *def_stmt)
134 tree lhs = gimple_assign_lhs (def_stmt);
135 use_operand_p use_p;
136 gimple *use_stmt;
138 if (!single_imm_use (lhs, &use_p, &use_stmt))
139 return NULL;
141 if (!vect_same_loop_or_bb_p (def_stmt, use_stmt))
142 return NULL;
144 return use_stmt;
147 /* Check whether NAME, an ssa-name used in USE_STMT,
148 is a result of a type promotion, such that:
149 DEF_STMT: NAME = NOP (name0)
150 If CHECK_SIGN is TRUE, check that either both types are signed or both are
151 unsigned. */
153 static bool
154 type_conversion_p (tree name, gimple *use_stmt, bool check_sign,
155 tree *orig_type, gimple **def_stmt, bool *promotion)
157 gimple *dummy_gimple;
158 stmt_vec_info stmt_vinfo;
159 tree type = TREE_TYPE (name);
160 tree oprnd0;
161 enum vect_def_type dt;
163 stmt_vinfo = vinfo_for_stmt (use_stmt);
164 if (!vect_is_simple_use (name, stmt_vinfo->vinfo, def_stmt, &dt))
165 return false;
167 if (dt != vect_internal_def
168 && dt != vect_external_def && dt != vect_constant_def)
169 return false;
171 if (!*def_stmt)
172 return false;
174 if (dt == vect_internal_def)
176 stmt_vec_info def_vinfo = vinfo_for_stmt (*def_stmt);
177 if (STMT_VINFO_IN_PATTERN_P (def_vinfo))
178 return false;
181 if (!is_gimple_assign (*def_stmt))
182 return false;
184 if (!CONVERT_EXPR_CODE_P (gimple_assign_rhs_code (*def_stmt)))
185 return false;
187 oprnd0 = gimple_assign_rhs1 (*def_stmt);
189 *orig_type = TREE_TYPE (oprnd0);
190 if (!INTEGRAL_TYPE_P (type) || !INTEGRAL_TYPE_P (*orig_type)
191 || ((TYPE_UNSIGNED (type) != TYPE_UNSIGNED (*orig_type)) && check_sign))
192 return false;
194 if (TYPE_PRECISION (type) >= (TYPE_PRECISION (*orig_type) * 2))
195 *promotion = true;
196 else
197 *promotion = false;
199 if (!vect_is_simple_use (oprnd0, stmt_vinfo->vinfo, &dummy_gimple, &dt))
200 return false;
202 return true;
205 /* Helper to return a new temporary for pattern of TYPE for STMT. If STMT
206 is NULL, the caller must set SSA_NAME_DEF_STMT for the returned SSA var. */
208 static tree
209 vect_recog_temp_ssa_var (tree type, gimple *stmt)
211 return make_temp_ssa_name (type, stmt, "patt");
214 /* Function vect_recog_dot_prod_pattern
216 Try to find the following pattern:
218 type x_t, y_t;
219 TYPE1 prod;
220 TYPE2 sum = init;
221 loop:
222 sum_0 = phi <init, sum_1>
223 S1 x_t = ...
224 S2 y_t = ...
225 S3 x_T = (TYPE1) x_t;
226 S4 y_T = (TYPE1) y_t;
227 S5 prod = x_T * y_T;
228 [S6 prod = (TYPE2) prod; #optional]
229 S7 sum_1 = prod + sum_0;
231 where 'TYPE1' is exactly double the size of type 'type', and 'TYPE2' is the
232 same size of 'TYPE1' or bigger. This is a special case of a reduction
233 computation.
235 Input:
237 * STMTS: Contains a stmt from which the pattern search begins. In the
238 example, when this function is called with S7, the pattern {S3,S4,S5,S6,S7}
239 will be detected.
241 Output:
243 * TYPE_IN: The type of the input arguments to the pattern.
245 * TYPE_OUT: The type of the output of this pattern.
247 * Return value: A new stmt that will be used to replace the sequence of
248 stmts that constitute the pattern. In this case it will be:
249 WIDEN_DOT_PRODUCT <x_t, y_t, sum_0>
251 Note: The dot-prod idiom is a widening reduction pattern that is
252 vectorized without preserving all the intermediate results. It
253 produces only N/2 (widened) results (by summing up pairs of
254 intermediate results) rather than all N results. Therefore, we
255 cannot allow this pattern when we want to get all the results and in
256 the correct order (as is the case when this computation is in an
257 inner-loop nested in an outer-loop that us being vectorized). */
259 static gimple *
260 vect_recog_dot_prod_pattern (vec<gimple *> *stmts, tree *type_in,
261 tree *type_out)
263 gimple *stmt, *last_stmt = (*stmts)[0];
264 tree oprnd0, oprnd1;
265 tree oprnd00, oprnd01;
266 stmt_vec_info stmt_vinfo = vinfo_for_stmt (last_stmt);
267 tree type, half_type;
268 gimple *pattern_stmt;
269 tree prod_type;
270 loop_vec_info loop_info = STMT_VINFO_LOOP_VINFO (stmt_vinfo);
271 struct loop *loop;
272 tree var;
273 bool promotion;
275 if (!loop_info)
276 return NULL;
278 loop = LOOP_VINFO_LOOP (loop_info);
280 /* We don't allow changing the order of the computation in the inner-loop
281 when doing outer-loop vectorization. */
282 if (loop && nested_in_vect_loop_p (loop, last_stmt))
283 return NULL;
285 if (!is_gimple_assign (last_stmt))
286 return NULL;
288 type = gimple_expr_type (last_stmt);
290 /* Look for the following pattern
291 DX = (TYPE1) X;
292 DY = (TYPE1) Y;
293 DPROD = DX * DY;
294 DDPROD = (TYPE2) DPROD;
295 sum_1 = DDPROD + sum_0;
296 In which
297 - DX is double the size of X
298 - DY is double the size of Y
299 - DX, DY, DPROD all have the same type
300 - sum is the same size of DPROD or bigger
301 - sum has been recognized as a reduction variable.
303 This is equivalent to:
304 DPROD = X w* Y; #widen mult
305 sum_1 = DPROD w+ sum_0; #widen summation
307 DPROD = X w* Y; #widen mult
308 sum_1 = DPROD + sum_0; #summation
311 /* Starting from LAST_STMT, follow the defs of its uses in search
312 of the above pattern. */
314 if (gimple_assign_rhs_code (last_stmt) != PLUS_EXPR)
315 return NULL;
317 if (STMT_VINFO_IN_PATTERN_P (stmt_vinfo))
319 /* Has been detected as widening-summation? */
321 stmt = STMT_VINFO_RELATED_STMT (stmt_vinfo);
322 type = gimple_expr_type (stmt);
323 if (gimple_assign_rhs_code (stmt) != WIDEN_SUM_EXPR)
324 return NULL;
325 oprnd0 = gimple_assign_rhs1 (stmt);
326 oprnd1 = gimple_assign_rhs2 (stmt);
327 half_type = TREE_TYPE (oprnd0);
329 else
331 gimple *def_stmt;
333 if (STMT_VINFO_DEF_TYPE (stmt_vinfo) != vect_reduction_def
334 && ! STMT_VINFO_GROUP_FIRST_ELEMENT (stmt_vinfo))
335 return NULL;
336 oprnd0 = gimple_assign_rhs1 (last_stmt);
337 oprnd1 = gimple_assign_rhs2 (last_stmt);
338 if (!types_compatible_p (TREE_TYPE (oprnd0), type)
339 || !types_compatible_p (TREE_TYPE (oprnd1), type))
340 return NULL;
341 stmt = last_stmt;
343 if (type_conversion_p (oprnd0, stmt, true, &half_type, &def_stmt,
344 &promotion)
345 && promotion)
347 stmt = def_stmt;
348 oprnd0 = gimple_assign_rhs1 (stmt);
350 else
351 half_type = type;
354 /* So far so good. Since last_stmt was detected as a (summation) reduction,
355 we know that oprnd1 is the reduction variable (defined by a loop-header
356 phi), and oprnd0 is an ssa-name defined by a stmt in the loop body.
357 Left to check that oprnd0 is defined by a (widen_)mult_expr */
358 if (TREE_CODE (oprnd0) != SSA_NAME)
359 return NULL;
361 prod_type = half_type;
362 stmt = SSA_NAME_DEF_STMT (oprnd0);
364 /* It could not be the dot_prod pattern if the stmt is outside the loop. */
365 if (!gimple_bb (stmt) || !flow_bb_inside_loop_p (loop, gimple_bb (stmt)))
366 return NULL;
368 /* FORNOW. Can continue analyzing the def-use chain when this stmt in a phi
369 inside the loop (in case we are analyzing an outer-loop). */
370 if (!is_gimple_assign (stmt))
371 return NULL;
372 stmt_vinfo = vinfo_for_stmt (stmt);
373 gcc_assert (stmt_vinfo);
374 if (STMT_VINFO_DEF_TYPE (stmt_vinfo) != vect_internal_def)
375 return NULL;
376 if (gimple_assign_rhs_code (stmt) != MULT_EXPR)
377 return NULL;
378 if (STMT_VINFO_IN_PATTERN_P (stmt_vinfo))
380 /* Has been detected as a widening multiplication? */
382 stmt = STMT_VINFO_RELATED_STMT (stmt_vinfo);
383 if (gimple_assign_rhs_code (stmt) != WIDEN_MULT_EXPR)
384 return NULL;
385 stmt_vinfo = vinfo_for_stmt (stmt);
386 gcc_assert (stmt_vinfo);
387 gcc_assert (STMT_VINFO_DEF_TYPE (stmt_vinfo) == vect_internal_def);
388 oprnd00 = gimple_assign_rhs1 (stmt);
389 oprnd01 = gimple_assign_rhs2 (stmt);
390 STMT_VINFO_PATTERN_DEF_SEQ (vinfo_for_stmt (last_stmt))
391 = STMT_VINFO_PATTERN_DEF_SEQ (stmt_vinfo);
393 else
395 tree half_type0, half_type1;
396 gimple *def_stmt;
397 tree oprnd0, oprnd1;
399 oprnd0 = gimple_assign_rhs1 (stmt);
400 oprnd1 = gimple_assign_rhs2 (stmt);
401 if (!types_compatible_p (TREE_TYPE (oprnd0), prod_type)
402 || !types_compatible_p (TREE_TYPE (oprnd1), prod_type))
403 return NULL;
404 if (!type_conversion_p (oprnd0, stmt, true, &half_type0, &def_stmt,
405 &promotion)
406 || !promotion)
407 return NULL;
408 oprnd00 = gimple_assign_rhs1 (def_stmt);
409 if (!type_conversion_p (oprnd1, stmt, true, &half_type1, &def_stmt,
410 &promotion)
411 || !promotion)
412 return NULL;
413 oprnd01 = gimple_assign_rhs1 (def_stmt);
414 if (!types_compatible_p (half_type0, half_type1))
415 return NULL;
416 if (TYPE_PRECISION (prod_type) != TYPE_PRECISION (half_type0) * 2)
417 return NULL;
420 half_type = TREE_TYPE (oprnd00);
421 *type_in = half_type;
422 *type_out = type;
424 /* Pattern detected. Create a stmt to be used to replace the pattern: */
425 var = vect_recog_temp_ssa_var (type, NULL);
426 pattern_stmt = gimple_build_assign (var, DOT_PROD_EXPR,
427 oprnd00, oprnd01, oprnd1);
429 if (dump_enabled_p ())
431 dump_printf_loc (MSG_NOTE, vect_location,
432 "vect_recog_dot_prod_pattern: detected: ");
433 dump_gimple_stmt (MSG_NOTE, TDF_SLIM, pattern_stmt, 0);
436 return pattern_stmt;
440 /* Function vect_recog_sad_pattern
442 Try to find the following Sum of Absolute Difference (SAD) pattern:
444 type x_t, y_t;
445 signed TYPE1 diff, abs_diff;
446 TYPE2 sum = init;
447 loop:
448 sum_0 = phi <init, sum_1>
449 S1 x_t = ...
450 S2 y_t = ...
451 S3 x_T = (TYPE1) x_t;
452 S4 y_T = (TYPE1) y_t;
453 S5 diff = x_T - y_T;
454 S6 abs_diff = ABS_EXPR <diff>;
455 [S7 abs_diff = (TYPE2) abs_diff; #optional]
456 S8 sum_1 = abs_diff + sum_0;
458 where 'TYPE1' is at least double the size of type 'type', and 'TYPE2' is the
459 same size of 'TYPE1' or bigger. This is a special case of a reduction
460 computation.
462 Input:
464 * STMTS: Contains a stmt from which the pattern search begins. In the
465 example, when this function is called with S8, the pattern
466 {S3,S4,S5,S6,S7,S8} will be detected.
468 Output:
470 * TYPE_IN: The type of the input arguments to the pattern.
472 * TYPE_OUT: The type of the output of this pattern.
474 * Return value: A new stmt that will be used to replace the sequence of
475 stmts that constitute the pattern. In this case it will be:
476 SAD_EXPR <x_t, y_t, sum_0>
479 static gimple *
480 vect_recog_sad_pattern (vec<gimple *> *stmts, tree *type_in,
481 tree *type_out)
483 gimple *last_stmt = (*stmts)[0];
484 tree sad_oprnd0, sad_oprnd1;
485 stmt_vec_info stmt_vinfo = vinfo_for_stmt (last_stmt);
486 tree half_type;
487 loop_vec_info loop_info = STMT_VINFO_LOOP_VINFO (stmt_vinfo);
488 struct loop *loop;
489 bool promotion;
491 if (!loop_info)
492 return NULL;
494 loop = LOOP_VINFO_LOOP (loop_info);
496 /* We don't allow changing the order of the computation in the inner-loop
497 when doing outer-loop vectorization. */
498 if (loop && nested_in_vect_loop_p (loop, last_stmt))
499 return NULL;
501 if (!is_gimple_assign (last_stmt))
502 return NULL;
504 tree sum_type = gimple_expr_type (last_stmt);
506 /* Look for the following pattern
507 DX = (TYPE1) X;
508 DY = (TYPE1) Y;
509 DDIFF = DX - DY;
510 DAD = ABS_EXPR <DDIFF>;
511 DDPROD = (TYPE2) DPROD;
512 sum_1 = DAD + sum_0;
513 In which
514 - DX is at least double the size of X
515 - DY is at least double the size of Y
516 - DX, DY, DDIFF, DAD all have the same type
517 - sum is the same size of DAD or bigger
518 - sum has been recognized as a reduction variable.
520 This is equivalent to:
521 DDIFF = X w- Y; #widen sub
522 DAD = ABS_EXPR <DDIFF>;
523 sum_1 = DAD w+ sum_0; #widen summation
525 DDIFF = X w- Y; #widen sub
526 DAD = ABS_EXPR <DDIFF>;
527 sum_1 = DAD + sum_0; #summation
530 /* Starting from LAST_STMT, follow the defs of its uses in search
531 of the above pattern. */
533 if (gimple_assign_rhs_code (last_stmt) != PLUS_EXPR)
534 return NULL;
536 tree plus_oprnd0, plus_oprnd1;
538 if (STMT_VINFO_IN_PATTERN_P (stmt_vinfo))
540 /* Has been detected as widening-summation? */
542 gimple *stmt = STMT_VINFO_RELATED_STMT (stmt_vinfo);
543 sum_type = gimple_expr_type (stmt);
544 if (gimple_assign_rhs_code (stmt) != WIDEN_SUM_EXPR)
545 return NULL;
546 plus_oprnd0 = gimple_assign_rhs1 (stmt);
547 plus_oprnd1 = gimple_assign_rhs2 (stmt);
548 half_type = TREE_TYPE (plus_oprnd0);
550 else
552 gimple *def_stmt;
554 if (STMT_VINFO_DEF_TYPE (stmt_vinfo) != vect_reduction_def
555 && ! STMT_VINFO_GROUP_FIRST_ELEMENT (stmt_vinfo))
556 return NULL;
557 plus_oprnd0 = gimple_assign_rhs1 (last_stmt);
558 plus_oprnd1 = gimple_assign_rhs2 (last_stmt);
559 if (!types_compatible_p (TREE_TYPE (plus_oprnd0), sum_type)
560 || !types_compatible_p (TREE_TYPE (plus_oprnd1), sum_type))
561 return NULL;
563 /* The type conversion could be promotion, demotion,
564 or just signed -> unsigned. */
565 if (type_conversion_p (plus_oprnd0, last_stmt, false,
566 &half_type, &def_stmt, &promotion))
567 plus_oprnd0 = gimple_assign_rhs1 (def_stmt);
568 else
569 half_type = sum_type;
572 /* So far so good. Since last_stmt was detected as a (summation) reduction,
573 we know that plus_oprnd1 is the reduction variable (defined by a loop-header
574 phi), and plus_oprnd0 is an ssa-name defined by a stmt in the loop body.
575 Then check that plus_oprnd0 is defined by an abs_expr. */
577 if (TREE_CODE (plus_oprnd0) != SSA_NAME)
578 return NULL;
580 tree abs_type = half_type;
581 gimple *abs_stmt = SSA_NAME_DEF_STMT (plus_oprnd0);
583 /* It could not be the sad pattern if the abs_stmt is outside the loop. */
584 if (!gimple_bb (abs_stmt) || !flow_bb_inside_loop_p (loop, gimple_bb (abs_stmt)))
585 return NULL;
587 /* FORNOW. Can continue analyzing the def-use chain when this stmt in a phi
588 inside the loop (in case we are analyzing an outer-loop). */
589 if (!is_gimple_assign (abs_stmt))
590 return NULL;
592 stmt_vec_info abs_stmt_vinfo = vinfo_for_stmt (abs_stmt);
593 gcc_assert (abs_stmt_vinfo);
594 if (STMT_VINFO_DEF_TYPE (abs_stmt_vinfo) != vect_internal_def)
595 return NULL;
596 if (gimple_assign_rhs_code (abs_stmt) != ABS_EXPR)
597 return NULL;
599 tree abs_oprnd = gimple_assign_rhs1 (abs_stmt);
600 if (!types_compatible_p (TREE_TYPE (abs_oprnd), abs_type))
601 return NULL;
602 if (TYPE_UNSIGNED (abs_type))
603 return NULL;
605 /* We then detect if the operand of abs_expr is defined by a minus_expr. */
607 if (TREE_CODE (abs_oprnd) != SSA_NAME)
608 return NULL;
610 gimple *diff_stmt = SSA_NAME_DEF_STMT (abs_oprnd);
612 /* It could not be the sad pattern if the diff_stmt is outside the loop. */
613 if (!gimple_bb (diff_stmt)
614 || !flow_bb_inside_loop_p (loop, gimple_bb (diff_stmt)))
615 return NULL;
617 /* FORNOW. Can continue analyzing the def-use chain when this stmt in a phi
618 inside the loop (in case we are analyzing an outer-loop). */
619 if (!is_gimple_assign (diff_stmt))
620 return NULL;
622 stmt_vec_info diff_stmt_vinfo = vinfo_for_stmt (diff_stmt);
623 gcc_assert (diff_stmt_vinfo);
624 if (STMT_VINFO_DEF_TYPE (diff_stmt_vinfo) != vect_internal_def)
625 return NULL;
626 if (gimple_assign_rhs_code (diff_stmt) != MINUS_EXPR)
627 return NULL;
629 tree half_type0, half_type1;
630 gimple *def_stmt;
632 tree minus_oprnd0 = gimple_assign_rhs1 (diff_stmt);
633 tree minus_oprnd1 = gimple_assign_rhs2 (diff_stmt);
635 if (!types_compatible_p (TREE_TYPE (minus_oprnd0), abs_type)
636 || !types_compatible_p (TREE_TYPE (minus_oprnd1), abs_type))
637 return NULL;
638 if (!type_conversion_p (minus_oprnd0, diff_stmt, false,
639 &half_type0, &def_stmt, &promotion)
640 || !promotion)
641 return NULL;
642 sad_oprnd0 = gimple_assign_rhs1 (def_stmt);
644 if (!type_conversion_p (minus_oprnd1, diff_stmt, false,
645 &half_type1, &def_stmt, &promotion)
646 || !promotion)
647 return NULL;
648 sad_oprnd1 = gimple_assign_rhs1 (def_stmt);
650 if (!types_compatible_p (half_type0, half_type1))
651 return NULL;
652 if (TYPE_PRECISION (abs_type) < TYPE_PRECISION (half_type0) * 2
653 || TYPE_PRECISION (sum_type) < TYPE_PRECISION (half_type0) * 2)
654 return NULL;
656 *type_in = TREE_TYPE (sad_oprnd0);
657 *type_out = sum_type;
659 /* Pattern detected. Create a stmt to be used to replace the pattern: */
660 tree var = vect_recog_temp_ssa_var (sum_type, NULL);
661 gimple *pattern_stmt = gimple_build_assign (var, SAD_EXPR, sad_oprnd0,
662 sad_oprnd1, plus_oprnd1);
664 if (dump_enabled_p ())
666 dump_printf_loc (MSG_NOTE, vect_location,
667 "vect_recog_sad_pattern: detected: ");
668 dump_gimple_stmt (MSG_NOTE, TDF_SLIM, pattern_stmt, 0);
671 return pattern_stmt;
675 /* Handle widening operation by a constant. At the moment we support MULT_EXPR
676 and LSHIFT_EXPR.
678 For MULT_EXPR we check that CONST_OPRND fits HALF_TYPE, and for LSHIFT_EXPR
679 we check that CONST_OPRND is less or equal to the size of HALF_TYPE.
681 Otherwise, if the type of the result (TYPE) is at least 4 times bigger than
682 HALF_TYPE, and there is an intermediate type (2 times smaller than TYPE)
683 that satisfies the above restrictions, we can perform a widening opeartion
684 from the intermediate type to TYPE and replace a_T = (TYPE) a_t;
685 with a_it = (interm_type) a_t; Store such operation in *WSTMT. */
687 static bool
688 vect_handle_widen_op_by_const (gimple *stmt, enum tree_code code,
689 tree const_oprnd, tree *oprnd,
690 gimple **wstmt, tree type,
691 tree *half_type, gimple *def_stmt)
693 tree new_type, new_oprnd;
695 if (code != MULT_EXPR && code != LSHIFT_EXPR)
696 return false;
698 if (((code == MULT_EXPR && int_fits_type_p (const_oprnd, *half_type))
699 || (code == LSHIFT_EXPR
700 && compare_tree_int (const_oprnd, TYPE_PRECISION (*half_type))
701 != 1))
702 && TYPE_PRECISION (type) == (TYPE_PRECISION (*half_type) * 2))
704 /* CONST_OPRND is a constant of HALF_TYPE. */
705 *oprnd = gimple_assign_rhs1 (def_stmt);
706 return true;
709 if (TYPE_PRECISION (type) < (TYPE_PRECISION (*half_type) * 4))
710 return false;
712 if (!vect_same_loop_or_bb_p (stmt, def_stmt))
713 return false;
715 /* TYPE is 4 times bigger than HALF_TYPE, try widening operation for
716 a type 2 times bigger than HALF_TYPE. */
717 new_type = build_nonstandard_integer_type (TYPE_PRECISION (type) / 2,
718 TYPE_UNSIGNED (type));
719 if ((code == MULT_EXPR && !int_fits_type_p (const_oprnd, new_type))
720 || (code == LSHIFT_EXPR
721 && compare_tree_int (const_oprnd, TYPE_PRECISION (new_type)) == 1))
722 return false;
724 /* Use NEW_TYPE for widening operation and create a_T = (NEW_TYPE) a_t; */
725 *oprnd = gimple_assign_rhs1 (def_stmt);
726 new_oprnd = make_ssa_name (new_type);
727 *wstmt = gimple_build_assign (new_oprnd, NOP_EXPR, *oprnd);
728 *oprnd = new_oprnd;
730 *half_type = new_type;
731 return true;
735 /* Function vect_recog_widen_mult_pattern
737 Try to find the following pattern:
739 type1 a_t;
740 type2 b_t;
741 TYPE a_T, b_T, prod_T;
743 S1 a_t = ;
744 S2 b_t = ;
745 S3 a_T = (TYPE) a_t;
746 S4 b_T = (TYPE) b_t;
747 S5 prod_T = a_T * b_T;
749 where type 'TYPE' is at least double the size of type 'type1' and 'type2'.
751 Also detect unsigned cases:
753 unsigned type1 a_t;
754 unsigned type2 b_t;
755 unsigned TYPE u_prod_T;
756 TYPE a_T, b_T, prod_T;
758 S1 a_t = ;
759 S2 b_t = ;
760 S3 a_T = (TYPE) a_t;
761 S4 b_T = (TYPE) b_t;
762 S5 prod_T = a_T * b_T;
763 S6 u_prod_T = (unsigned TYPE) prod_T;
765 and multiplication by constants:
767 type a_t;
768 TYPE a_T, prod_T;
770 S1 a_t = ;
771 S3 a_T = (TYPE) a_t;
772 S5 prod_T = a_T * CONST;
774 A special case of multiplication by constants is when 'TYPE' is 4 times
775 bigger than 'type', but CONST fits an intermediate type 2 times smaller
776 than 'TYPE'. In that case we create an additional pattern stmt for S3
777 to create a variable of the intermediate type, and perform widen-mult
778 on the intermediate type as well:
780 type a_t;
781 interm_type a_it;
782 TYPE a_T, prod_T, prod_T';
784 S1 a_t = ;
785 S3 a_T = (TYPE) a_t;
786 '--> a_it = (interm_type) a_t;
787 S5 prod_T = a_T * CONST;
788 '--> prod_T' = a_it w* CONST;
790 Input/Output:
792 * STMTS: Contains a stmt from which the pattern search begins. In the
793 example, when this function is called with S5, the pattern {S3,S4,S5,(S6)}
794 is detected. In case of unsigned widen-mult, the original stmt (S5) is
795 replaced with S6 in STMTS. In case of multiplication by a constant
796 of an intermediate type (the last case above), STMTS also contains S3
797 (inserted before S5).
799 Output:
801 * TYPE_IN: The type of the input arguments to the pattern.
803 * TYPE_OUT: The type of the output of this pattern.
805 * Return value: A new stmt that will be used to replace the sequence of
806 stmts that constitute the pattern. In this case it will be:
807 WIDEN_MULT <a_t, b_t>
808 If the result of WIDEN_MULT needs to be converted to a larger type, the
809 returned stmt will be this type conversion stmt.
812 static gimple *
813 vect_recog_widen_mult_pattern (vec<gimple *> *stmts,
814 tree *type_in, tree *type_out)
816 gimple *last_stmt = stmts->pop ();
817 gimple *def_stmt0, *def_stmt1;
818 tree oprnd0, oprnd1;
819 tree type, half_type0, half_type1;
820 gimple *new_stmt = NULL, *pattern_stmt = NULL;
821 tree vectype, vecitype;
822 tree var;
823 enum tree_code dummy_code;
824 int dummy_int;
825 vec<tree> dummy_vec;
826 bool op1_ok;
827 bool promotion;
829 if (!is_gimple_assign (last_stmt))
830 return NULL;
832 type = gimple_expr_type (last_stmt);
834 /* Starting from LAST_STMT, follow the defs of its uses in search
835 of the above pattern. */
837 if (gimple_assign_rhs_code (last_stmt) != MULT_EXPR)
838 return NULL;
840 oprnd0 = gimple_assign_rhs1 (last_stmt);
841 oprnd1 = gimple_assign_rhs2 (last_stmt);
842 if (!types_compatible_p (TREE_TYPE (oprnd0), type)
843 || !types_compatible_p (TREE_TYPE (oprnd1), type))
844 return NULL;
846 /* Check argument 0. */
847 if (!type_conversion_p (oprnd0, last_stmt, false, &half_type0, &def_stmt0,
848 &promotion)
849 || !promotion)
850 return NULL;
851 /* Check argument 1. */
852 op1_ok = type_conversion_p (oprnd1, last_stmt, false, &half_type1,
853 &def_stmt1, &promotion);
855 if (op1_ok && promotion)
857 oprnd0 = gimple_assign_rhs1 (def_stmt0);
858 oprnd1 = gimple_assign_rhs1 (def_stmt1);
860 else
862 if (TREE_CODE (oprnd1) == INTEGER_CST
863 && TREE_CODE (half_type0) == INTEGER_TYPE
864 && vect_handle_widen_op_by_const (last_stmt, MULT_EXPR, oprnd1,
865 &oprnd0, &new_stmt, type,
866 &half_type0, def_stmt0))
868 half_type1 = half_type0;
869 oprnd1 = fold_convert (half_type1, oprnd1);
871 else
872 return NULL;
875 /* If the two arguments have different sizes, convert the one with
876 the smaller type into the larger type. */
877 if (TYPE_PRECISION (half_type0) != TYPE_PRECISION (half_type1))
879 /* If we already used up the single-stmt slot give up. */
880 if (new_stmt)
881 return NULL;
883 tree* oprnd = NULL;
884 gimple *def_stmt = NULL;
886 if (TYPE_PRECISION (half_type0) < TYPE_PRECISION (half_type1))
888 def_stmt = def_stmt0;
889 half_type0 = half_type1;
890 oprnd = &oprnd0;
892 else
894 def_stmt = def_stmt1;
895 half_type1 = half_type0;
896 oprnd = &oprnd1;
899 tree old_oprnd = gimple_assign_rhs1 (def_stmt);
900 tree new_oprnd = make_ssa_name (half_type0);
901 new_stmt = gimple_build_assign (new_oprnd, NOP_EXPR, old_oprnd);
902 *oprnd = new_oprnd;
905 /* Handle unsigned case. Look for
906 S6 u_prod_T = (unsigned TYPE) prod_T;
907 Use unsigned TYPE as the type for WIDEN_MULT_EXPR. */
908 if (TYPE_UNSIGNED (type) != TYPE_UNSIGNED (half_type0))
910 gimple *use_stmt;
911 tree use_lhs;
912 tree use_type;
914 if (TYPE_UNSIGNED (type) == TYPE_UNSIGNED (half_type1))
915 return NULL;
917 use_stmt = vect_single_imm_use (last_stmt);
918 if (!use_stmt || !is_gimple_assign (use_stmt)
919 || !CONVERT_EXPR_CODE_P (gimple_assign_rhs_code (use_stmt)))
920 return NULL;
922 use_lhs = gimple_assign_lhs (use_stmt);
923 use_type = TREE_TYPE (use_lhs);
924 if (!INTEGRAL_TYPE_P (use_type)
925 || (TYPE_UNSIGNED (type) == TYPE_UNSIGNED (use_type))
926 || (TYPE_PRECISION (type) != TYPE_PRECISION (use_type)))
927 return NULL;
929 type = use_type;
930 last_stmt = use_stmt;
933 if (!types_compatible_p (half_type0, half_type1))
934 return NULL;
936 /* If TYPE is more than twice larger than HALF_TYPE, we use WIDEN_MULT
937 to get an intermediate result of type ITYPE. In this case we need
938 to build a statement to convert this intermediate result to type TYPE. */
939 tree itype = type;
940 if (TYPE_PRECISION (type) > TYPE_PRECISION (half_type0) * 2)
941 itype = build_nonstandard_integer_type
942 (GET_MODE_BITSIZE (TYPE_MODE (half_type0)) * 2,
943 TYPE_UNSIGNED (type));
945 /* Pattern detected. */
946 if (dump_enabled_p ())
947 dump_printf_loc (MSG_NOTE, vect_location,
948 "vect_recog_widen_mult_pattern: detected:\n");
950 /* Check target support */
951 vectype = get_vectype_for_scalar_type (half_type0);
952 vecitype = get_vectype_for_scalar_type (itype);
953 if (!vectype
954 || !vecitype
955 || !supportable_widening_operation (WIDEN_MULT_EXPR, last_stmt,
956 vecitype, vectype,
957 &dummy_code, &dummy_code,
958 &dummy_int, &dummy_vec))
959 return NULL;
961 *type_in = vectype;
962 *type_out = get_vectype_for_scalar_type (type);
964 /* Pattern supported. Create a stmt to be used to replace the pattern: */
965 var = vect_recog_temp_ssa_var (itype, NULL);
966 pattern_stmt = gimple_build_assign (var, WIDEN_MULT_EXPR, oprnd0, oprnd1);
968 stmt_vec_info stmt_vinfo = vinfo_for_stmt (last_stmt);
969 STMT_VINFO_PATTERN_DEF_SEQ (stmt_vinfo) = NULL;
971 /* If the original two operands have different sizes, we may need to convert
972 the smaller one into the larget type. If this is the case, at this point
973 the new stmt is already built. */
974 if (new_stmt)
976 append_pattern_def_seq (stmt_vinfo, new_stmt);
977 stmt_vec_info new_stmt_info
978 = new_stmt_vec_info (new_stmt, stmt_vinfo->vinfo);
979 set_vinfo_for_stmt (new_stmt, new_stmt_info);
980 STMT_VINFO_VECTYPE (new_stmt_info) = vectype;
983 /* If ITYPE is not TYPE, we need to build a type convertion stmt to convert
984 the result of the widen-mult operation into type TYPE. */
985 if (itype != type)
987 append_pattern_def_seq (stmt_vinfo, pattern_stmt);
988 stmt_vec_info pattern_stmt_info
989 = new_stmt_vec_info (pattern_stmt, stmt_vinfo->vinfo);
990 set_vinfo_for_stmt (pattern_stmt, pattern_stmt_info);
991 STMT_VINFO_VECTYPE (pattern_stmt_info) = vecitype;
992 pattern_stmt = gimple_build_assign (vect_recog_temp_ssa_var (type, NULL),
993 NOP_EXPR,
994 gimple_assign_lhs (pattern_stmt));
997 if (dump_enabled_p ())
998 dump_gimple_stmt_loc (MSG_NOTE, vect_location, TDF_SLIM, pattern_stmt, 0);
1000 stmts->safe_push (last_stmt);
1001 return pattern_stmt;
1005 /* Function vect_recog_pow_pattern
1007 Try to find the following pattern:
1009 x = POW (y, N);
1011 with POW being one of pow, powf, powi, powif and N being
1012 either 2 or 0.5.
1014 Input:
1016 * LAST_STMT: A stmt from which the pattern search begins.
1018 Output:
1020 * TYPE_IN: The type of the input arguments to the pattern.
1022 * TYPE_OUT: The type of the output of this pattern.
1024 * Return value: A new stmt that will be used to replace the sequence of
1025 stmts that constitute the pattern. In this case it will be:
1026 x = x * x
1028 x = sqrt (x)
1031 static gimple *
1032 vect_recog_pow_pattern (vec<gimple *> *stmts, tree *type_in,
1033 tree *type_out)
1035 gimple *last_stmt = (*stmts)[0];
1036 tree base, exp = NULL;
1037 gimple *stmt;
1038 tree var;
1040 if (!is_gimple_call (last_stmt) || gimple_call_lhs (last_stmt) == NULL)
1041 return NULL;
1043 switch (gimple_call_combined_fn (last_stmt))
1045 CASE_CFN_POW:
1046 CASE_CFN_POWI:
1047 base = gimple_call_arg (last_stmt, 0);
1048 exp = gimple_call_arg (last_stmt, 1);
1049 if (TREE_CODE (exp) != REAL_CST
1050 && TREE_CODE (exp) != INTEGER_CST)
1051 return NULL;
1052 break;
1054 default:
1055 return NULL;
1058 /* We now have a pow or powi builtin function call with a constant
1059 exponent. */
1061 *type_out = NULL_TREE;
1063 /* Catch squaring. */
1064 if ((tree_fits_shwi_p (exp)
1065 && tree_to_shwi (exp) == 2)
1066 || (TREE_CODE (exp) == REAL_CST
1067 && real_equal (&TREE_REAL_CST (exp), &dconst2)))
1069 *type_in = TREE_TYPE (base);
1071 var = vect_recog_temp_ssa_var (TREE_TYPE (base), NULL);
1072 stmt = gimple_build_assign (var, MULT_EXPR, base, base);
1073 return stmt;
1076 /* Catch square root. */
1077 if (TREE_CODE (exp) == REAL_CST
1078 && real_equal (&TREE_REAL_CST (exp), &dconsthalf))
1080 *type_in = get_vectype_for_scalar_type (TREE_TYPE (base));
1081 if (*type_in
1082 && direct_internal_fn_supported_p (IFN_SQRT, *type_in,
1083 OPTIMIZE_FOR_SPEED))
1085 gcall *stmt = gimple_build_call_internal (IFN_SQRT, 1, base);
1086 var = vect_recog_temp_ssa_var (TREE_TYPE (base), stmt);
1087 gimple_call_set_lhs (stmt, var);
1088 return stmt;
1092 return NULL;
1096 /* Function vect_recog_widen_sum_pattern
1098 Try to find the following pattern:
1100 type x_t;
1101 TYPE x_T, sum = init;
1102 loop:
1103 sum_0 = phi <init, sum_1>
1104 S1 x_t = *p;
1105 S2 x_T = (TYPE) x_t;
1106 S3 sum_1 = x_T + sum_0;
1108 where type 'TYPE' is at least double the size of type 'type', i.e - we're
1109 summing elements of type 'type' into an accumulator of type 'TYPE'. This is
1110 a special case of a reduction computation.
1112 Input:
1114 * LAST_STMT: A stmt from which the pattern search begins. In the example,
1115 when this function is called with S3, the pattern {S2,S3} will be detected.
1117 Output:
1119 * TYPE_IN: The type of the input arguments to the pattern.
1121 * TYPE_OUT: The type of the output of this pattern.
1123 * Return value: A new stmt that will be used to replace the sequence of
1124 stmts that constitute the pattern. In this case it will be:
1125 WIDEN_SUM <x_t, sum_0>
1127 Note: The widening-sum idiom is a widening reduction pattern that is
1128 vectorized without preserving all the intermediate results. It
1129 produces only N/2 (widened) results (by summing up pairs of
1130 intermediate results) rather than all N results. Therefore, we
1131 cannot allow this pattern when we want to get all the results and in
1132 the correct order (as is the case when this computation is in an
1133 inner-loop nested in an outer-loop that us being vectorized). */
1135 static gimple *
1136 vect_recog_widen_sum_pattern (vec<gimple *> *stmts, tree *type_in,
1137 tree *type_out)
1139 gimple *stmt, *last_stmt = (*stmts)[0];
1140 tree oprnd0, oprnd1;
1141 stmt_vec_info stmt_vinfo = vinfo_for_stmt (last_stmt);
1142 tree type, half_type;
1143 gimple *pattern_stmt;
1144 loop_vec_info loop_info = STMT_VINFO_LOOP_VINFO (stmt_vinfo);
1145 struct loop *loop;
1146 tree var;
1147 bool promotion;
1149 if (!loop_info)
1150 return NULL;
1152 loop = LOOP_VINFO_LOOP (loop_info);
1154 /* We don't allow changing the order of the computation in the inner-loop
1155 when doing outer-loop vectorization. */
1156 if (loop && nested_in_vect_loop_p (loop, last_stmt))
1157 return NULL;
1159 if (!is_gimple_assign (last_stmt))
1160 return NULL;
1162 type = gimple_expr_type (last_stmt);
1164 /* Look for the following pattern
1165 DX = (TYPE) X;
1166 sum_1 = DX + sum_0;
1167 In which DX is at least double the size of X, and sum_1 has been
1168 recognized as a reduction variable.
1171 /* Starting from LAST_STMT, follow the defs of its uses in search
1172 of the above pattern. */
1174 if (gimple_assign_rhs_code (last_stmt) != PLUS_EXPR)
1175 return NULL;
1177 if (STMT_VINFO_DEF_TYPE (stmt_vinfo) != vect_reduction_def
1178 && ! STMT_VINFO_GROUP_FIRST_ELEMENT (stmt_vinfo))
1179 return NULL;
1181 oprnd0 = gimple_assign_rhs1 (last_stmt);
1182 oprnd1 = gimple_assign_rhs2 (last_stmt);
1183 if (!types_compatible_p (TREE_TYPE (oprnd0), type)
1184 || !types_compatible_p (TREE_TYPE (oprnd1), type))
1185 return NULL;
1187 /* So far so good. Since last_stmt was detected as a (summation) reduction,
1188 we know that oprnd1 is the reduction variable (defined by a loop-header
1189 phi), and oprnd0 is an ssa-name defined by a stmt in the loop body.
1190 Left to check that oprnd0 is defined by a cast from type 'type' to type
1191 'TYPE'. */
1193 if (!type_conversion_p (oprnd0, last_stmt, true, &half_type, &stmt,
1194 &promotion)
1195 || !promotion)
1196 return NULL;
1198 oprnd0 = gimple_assign_rhs1 (stmt);
1199 *type_in = half_type;
1200 *type_out = type;
1202 /* Pattern detected. Create a stmt to be used to replace the pattern: */
1203 var = vect_recog_temp_ssa_var (type, NULL);
1204 pattern_stmt = gimple_build_assign (var, WIDEN_SUM_EXPR, oprnd0, oprnd1);
1206 if (dump_enabled_p ())
1208 dump_printf_loc (MSG_NOTE, vect_location,
1209 "vect_recog_widen_sum_pattern: detected: ");
1210 dump_gimple_stmt (MSG_NOTE, TDF_SLIM, pattern_stmt, 0);
1213 return pattern_stmt;
1217 /* Return TRUE if the operation in STMT can be performed on a smaller type.
1219 Input:
1220 STMT - a statement to check.
1221 DEF - we support operations with two operands, one of which is constant.
1222 The other operand can be defined by a demotion operation, or by a
1223 previous statement in a sequence of over-promoted operations. In the
1224 later case DEF is used to replace that operand. (It is defined by a
1225 pattern statement we created for the previous statement in the
1226 sequence).
1228 Input/output:
1229 NEW_TYPE - Output: a smaller type that we are trying to use. Input: if not
1230 NULL, it's the type of DEF.
1231 STMTS - additional pattern statements. If a pattern statement (type
1232 conversion) is created in this function, its original statement is
1233 added to STMTS.
1235 Output:
1236 OP0, OP1 - if the operation fits a smaller type, OP0 and OP1 are the new
1237 operands to use in the new pattern statement for STMT (will be created
1238 in vect_recog_over_widening_pattern ()).
1239 NEW_DEF_STMT - in case DEF has to be promoted, we create two pattern
1240 statements for STMT: the first one is a type promotion and the second
1241 one is the operation itself. We return the type promotion statement
1242 in NEW_DEF_STMT and further store it in STMT_VINFO_PATTERN_DEF_SEQ of
1243 the second pattern statement. */
1245 static bool
1246 vect_operation_fits_smaller_type (gimple *stmt, tree def, tree *new_type,
1247 tree *op0, tree *op1, gimple **new_def_stmt,
1248 vec<gimple *> *stmts)
1250 enum tree_code code;
1251 tree const_oprnd, oprnd;
1252 tree interm_type = NULL_TREE, half_type, new_oprnd, type;
1253 gimple *def_stmt, *new_stmt;
1254 bool first = false;
1255 bool promotion;
1257 *op0 = NULL_TREE;
1258 *op1 = NULL_TREE;
1259 *new_def_stmt = NULL;
1261 if (!is_gimple_assign (stmt))
1262 return false;
1264 code = gimple_assign_rhs_code (stmt);
1265 if (code != LSHIFT_EXPR && code != RSHIFT_EXPR
1266 && code != BIT_IOR_EXPR && code != BIT_XOR_EXPR && code != BIT_AND_EXPR)
1267 return false;
1269 oprnd = gimple_assign_rhs1 (stmt);
1270 const_oprnd = gimple_assign_rhs2 (stmt);
1271 type = gimple_expr_type (stmt);
1273 if (TREE_CODE (oprnd) != SSA_NAME
1274 || TREE_CODE (const_oprnd) != INTEGER_CST)
1275 return false;
1277 /* If oprnd has other uses besides that in stmt we cannot mark it
1278 as being part of a pattern only. */
1279 if (!has_single_use (oprnd))
1280 return false;
1282 /* If we are in the middle of a sequence, we use DEF from a previous
1283 statement. Otherwise, OPRND has to be a result of type promotion. */
1284 if (*new_type)
1286 half_type = *new_type;
1287 oprnd = def;
1289 else
1291 first = true;
1292 if (!type_conversion_p (oprnd, stmt, false, &half_type, &def_stmt,
1293 &promotion)
1294 || !promotion
1295 || !vect_same_loop_or_bb_p (stmt, def_stmt))
1296 return false;
1299 /* Can we perform the operation on a smaller type? */
1300 switch (code)
1302 case BIT_IOR_EXPR:
1303 case BIT_XOR_EXPR:
1304 case BIT_AND_EXPR:
1305 if (!int_fits_type_p (const_oprnd, half_type))
1307 /* HALF_TYPE is not enough. Try a bigger type if possible. */
1308 if (TYPE_PRECISION (type) < (TYPE_PRECISION (half_type) * 4))
1309 return false;
1311 interm_type = build_nonstandard_integer_type (
1312 TYPE_PRECISION (half_type) * 2, TYPE_UNSIGNED (type));
1313 if (!int_fits_type_p (const_oprnd, interm_type))
1314 return false;
1317 break;
1319 case LSHIFT_EXPR:
1320 /* Try intermediate type - HALF_TYPE is not enough for sure. */
1321 if (TYPE_PRECISION (type) < (TYPE_PRECISION (half_type) * 4))
1322 return false;
1324 /* Check that HALF_TYPE size + shift amount <= INTERM_TYPE size.
1325 (e.g., if the original value was char, the shift amount is at most 8
1326 if we want to use short). */
1327 if (compare_tree_int (const_oprnd, TYPE_PRECISION (half_type)) == 1)
1328 return false;
1330 interm_type = build_nonstandard_integer_type (
1331 TYPE_PRECISION (half_type) * 2, TYPE_UNSIGNED (type));
1333 if (!vect_supportable_shift (code, interm_type))
1334 return false;
1336 break;
1338 case RSHIFT_EXPR:
1339 if (vect_supportable_shift (code, half_type))
1340 break;
1342 /* Try intermediate type - HALF_TYPE is not supported. */
1343 if (TYPE_PRECISION (type) < (TYPE_PRECISION (half_type) * 4))
1344 return false;
1346 interm_type = build_nonstandard_integer_type (
1347 TYPE_PRECISION (half_type) * 2, TYPE_UNSIGNED (type));
1349 if (!vect_supportable_shift (code, interm_type))
1350 return false;
1352 break;
1354 default:
1355 gcc_unreachable ();
1358 /* There are four possible cases:
1359 1. OPRND is defined by a type promotion (in that case FIRST is TRUE, it's
1360 the first statement in the sequence)
1361 a. The original, HALF_TYPE, is not enough - we replace the promotion
1362 from HALF_TYPE to TYPE with a promotion to INTERM_TYPE.
1363 b. HALF_TYPE is sufficient, OPRND is set as the RHS of the original
1364 promotion.
1365 2. OPRND is defined by a pattern statement we created.
1366 a. Its type is not sufficient for the operation, we create a new stmt:
1367 a type conversion for OPRND from HALF_TYPE to INTERM_TYPE. We store
1368 this statement in NEW_DEF_STMT, and it is later put in
1369 STMT_VINFO_PATTERN_DEF_SEQ of the pattern statement for STMT.
1370 b. OPRND is good to use in the new statement. */
1371 if (first)
1373 if (interm_type)
1375 /* Replace the original type conversion HALF_TYPE->TYPE with
1376 HALF_TYPE->INTERM_TYPE. */
1377 if (STMT_VINFO_RELATED_STMT (vinfo_for_stmt (def_stmt)))
1379 new_stmt = STMT_VINFO_RELATED_STMT (vinfo_for_stmt (def_stmt));
1380 /* Check if the already created pattern stmt is what we need. */
1381 if (!is_gimple_assign (new_stmt)
1382 || !CONVERT_EXPR_CODE_P (gimple_assign_rhs_code (new_stmt))
1383 || TREE_TYPE (gimple_assign_lhs (new_stmt)) != interm_type)
1384 return false;
1386 stmts->safe_push (def_stmt);
1387 oprnd = gimple_assign_lhs (new_stmt);
1389 else
1391 /* Create NEW_OPRND = (INTERM_TYPE) OPRND. */
1392 oprnd = gimple_assign_rhs1 (def_stmt);
1393 new_oprnd = make_ssa_name (interm_type);
1394 new_stmt = gimple_build_assign (new_oprnd, NOP_EXPR, oprnd);
1395 STMT_VINFO_RELATED_STMT (vinfo_for_stmt (def_stmt)) = new_stmt;
1396 stmts->safe_push (def_stmt);
1397 oprnd = new_oprnd;
1400 else
1402 /* Retrieve the operand before the type promotion. */
1403 oprnd = gimple_assign_rhs1 (def_stmt);
1406 else
1408 if (interm_type)
1410 /* Create a type conversion HALF_TYPE->INTERM_TYPE. */
1411 new_oprnd = make_ssa_name (interm_type);
1412 new_stmt = gimple_build_assign (new_oprnd, NOP_EXPR, oprnd);
1413 oprnd = new_oprnd;
1414 *new_def_stmt = new_stmt;
1417 /* Otherwise, OPRND is already set. */
1420 if (interm_type)
1421 *new_type = interm_type;
1422 else
1423 *new_type = half_type;
1425 *op0 = oprnd;
1426 *op1 = fold_convert (*new_type, const_oprnd);
1428 return true;
1432 /* Try to find a statement or a sequence of statements that can be performed
1433 on a smaller type:
1435 type x_t;
1436 TYPE x_T, res0_T, res1_T;
1437 loop:
1438 S1 x_t = *p;
1439 S2 x_T = (TYPE) x_t;
1440 S3 res0_T = op (x_T, C0);
1441 S4 res1_T = op (res0_T, C1);
1442 S5 ... = () res1_T; - type demotion
1444 where type 'TYPE' is at least double the size of type 'type', C0 and C1 are
1445 constants.
1446 Check if S3 and S4 can be done on a smaller type than 'TYPE', it can either
1447 be 'type' or some intermediate type. For now, we expect S5 to be a type
1448 demotion operation. We also check that S3 and S4 have only one use. */
1450 static gimple *
1451 vect_recog_over_widening_pattern (vec<gimple *> *stmts,
1452 tree *type_in, tree *type_out)
1454 gimple *stmt = stmts->pop ();
1455 gimple *pattern_stmt = NULL, *new_def_stmt, *prev_stmt = NULL,
1456 *use_stmt = NULL;
1457 tree op0, op1, vectype = NULL_TREE, use_lhs, use_type;
1458 tree var = NULL_TREE, new_type = NULL_TREE, new_oprnd;
1459 bool first;
1460 tree type = NULL;
1462 first = true;
1463 while (1)
1465 if (!vinfo_for_stmt (stmt)
1466 || STMT_VINFO_IN_PATTERN_P (vinfo_for_stmt (stmt)))
1467 return NULL;
1469 new_def_stmt = NULL;
1470 if (!vect_operation_fits_smaller_type (stmt, var, &new_type,
1471 &op0, &op1, &new_def_stmt,
1472 stmts))
1474 if (first)
1475 return NULL;
1476 else
1477 break;
1480 /* STMT can be performed on a smaller type. Check its uses. */
1481 use_stmt = vect_single_imm_use (stmt);
1482 if (!use_stmt || !is_gimple_assign (use_stmt))
1483 return NULL;
1485 /* Create pattern statement for STMT. */
1486 vectype = get_vectype_for_scalar_type (new_type);
1487 if (!vectype)
1488 return NULL;
1490 /* We want to collect all the statements for which we create pattern
1491 statetments, except for the case when the last statement in the
1492 sequence doesn't have a corresponding pattern statement. In such
1493 case we associate the last pattern statement with the last statement
1494 in the sequence. Therefore, we only add the original statement to
1495 the list if we know that it is not the last. */
1496 if (prev_stmt)
1497 stmts->safe_push (prev_stmt);
1499 var = vect_recog_temp_ssa_var (new_type, NULL);
1500 pattern_stmt
1501 = gimple_build_assign (var, gimple_assign_rhs_code (stmt), op0, op1);
1502 STMT_VINFO_RELATED_STMT (vinfo_for_stmt (stmt)) = pattern_stmt;
1503 new_pattern_def_seq (vinfo_for_stmt (stmt), new_def_stmt);
1505 if (dump_enabled_p ())
1507 dump_printf_loc (MSG_NOTE, vect_location,
1508 "created pattern stmt: ");
1509 dump_gimple_stmt (MSG_NOTE, TDF_SLIM, pattern_stmt, 0);
1512 type = gimple_expr_type (stmt);
1513 prev_stmt = stmt;
1514 stmt = use_stmt;
1516 first = false;
1519 /* We got a sequence. We expect it to end with a type demotion operation.
1520 Otherwise, we quit (for now). There are three possible cases: the
1521 conversion is to NEW_TYPE (we don't do anything), the conversion is to
1522 a type bigger than NEW_TYPE and/or the signedness of USE_TYPE and
1523 NEW_TYPE differs (we create a new conversion statement). */
1524 if (CONVERT_EXPR_CODE_P (gimple_assign_rhs_code (use_stmt)))
1526 use_lhs = gimple_assign_lhs (use_stmt);
1527 use_type = TREE_TYPE (use_lhs);
1528 /* Support only type demotion or signedess change. */
1529 if (!INTEGRAL_TYPE_P (use_type)
1530 || TYPE_PRECISION (type) <= TYPE_PRECISION (use_type))
1531 return NULL;
1533 /* Check that NEW_TYPE is not bigger than the conversion result. */
1534 if (TYPE_PRECISION (new_type) > TYPE_PRECISION (use_type))
1535 return NULL;
1537 if (TYPE_UNSIGNED (new_type) != TYPE_UNSIGNED (use_type)
1538 || TYPE_PRECISION (new_type) != TYPE_PRECISION (use_type))
1540 /* Create NEW_TYPE->USE_TYPE conversion. */
1541 new_oprnd = make_ssa_name (use_type);
1542 pattern_stmt = gimple_build_assign (new_oprnd, NOP_EXPR, var);
1543 STMT_VINFO_RELATED_STMT (vinfo_for_stmt (use_stmt)) = pattern_stmt;
1545 *type_in = get_vectype_for_scalar_type (new_type);
1546 *type_out = get_vectype_for_scalar_type (use_type);
1548 /* We created a pattern statement for the last statement in the
1549 sequence, so we don't need to associate it with the pattern
1550 statement created for PREV_STMT. Therefore, we add PREV_STMT
1551 to the list in order to mark it later in vect_pattern_recog_1. */
1552 if (prev_stmt)
1553 stmts->safe_push (prev_stmt);
1555 else
1557 if (prev_stmt)
1558 STMT_VINFO_PATTERN_DEF_SEQ (vinfo_for_stmt (use_stmt))
1559 = STMT_VINFO_PATTERN_DEF_SEQ (vinfo_for_stmt (prev_stmt));
1561 *type_in = vectype;
1562 *type_out = NULL_TREE;
1565 stmts->safe_push (use_stmt);
1567 else
1568 /* TODO: support general case, create a conversion to the correct type. */
1569 return NULL;
1571 /* Pattern detected. */
1572 if (dump_enabled_p ())
1574 dump_printf_loc (MSG_NOTE, vect_location,
1575 "vect_recog_over_widening_pattern: detected: ");
1576 dump_gimple_stmt (MSG_NOTE, TDF_SLIM, pattern_stmt, 0);
1579 return pattern_stmt;
1582 /* Detect widening shift pattern:
1584 type a_t;
1585 TYPE a_T, res_T;
1587 S1 a_t = ;
1588 S2 a_T = (TYPE) a_t;
1589 S3 res_T = a_T << CONST;
1591 where type 'TYPE' is at least double the size of type 'type'.
1593 Also detect cases where the shift result is immediately converted
1594 to another type 'result_type' that is no larger in size than 'TYPE'.
1595 In those cases we perform a widen-shift that directly results in
1596 'result_type', to avoid a possible over-widening situation:
1598 type a_t;
1599 TYPE a_T, res_T;
1600 result_type res_result;
1602 S1 a_t = ;
1603 S2 a_T = (TYPE) a_t;
1604 S3 res_T = a_T << CONST;
1605 S4 res_result = (result_type) res_T;
1606 '--> res_result' = a_t w<< CONST;
1608 And a case when 'TYPE' is 4 times bigger than 'type'. In that case we
1609 create an additional pattern stmt for S2 to create a variable of an
1610 intermediate type, and perform widen-shift on the intermediate type:
1612 type a_t;
1613 interm_type a_it;
1614 TYPE a_T, res_T, res_T';
1616 S1 a_t = ;
1617 S2 a_T = (TYPE) a_t;
1618 '--> a_it = (interm_type) a_t;
1619 S3 res_T = a_T << CONST;
1620 '--> res_T' = a_it <<* CONST;
1622 Input/Output:
1624 * STMTS: Contains a stmt from which the pattern search begins.
1625 In case of unsigned widen-shift, the original stmt (S3) is replaced with S4
1626 in STMTS. When an intermediate type is used and a pattern statement is
1627 created for S2, we also put S2 here (before S3).
1629 Output:
1631 * TYPE_IN: The type of the input arguments to the pattern.
1633 * TYPE_OUT: The type of the output of this pattern.
1635 * Return value: A new stmt that will be used to replace the sequence of
1636 stmts that constitute the pattern. In this case it will be:
1637 WIDEN_LSHIFT_EXPR <a_t, CONST>. */
1639 static gimple *
1640 vect_recog_widen_shift_pattern (vec<gimple *> *stmts,
1641 tree *type_in, tree *type_out)
1643 gimple *last_stmt = stmts->pop ();
1644 gimple *def_stmt0;
1645 tree oprnd0, oprnd1;
1646 tree type, half_type0;
1647 gimple *pattern_stmt;
1648 tree vectype, vectype_out = NULL_TREE;
1649 tree var;
1650 enum tree_code dummy_code;
1651 int dummy_int;
1652 vec<tree> dummy_vec;
1653 gimple *use_stmt;
1654 bool promotion;
1656 if (!is_gimple_assign (last_stmt) || !vinfo_for_stmt (last_stmt))
1657 return NULL;
1659 if (STMT_VINFO_IN_PATTERN_P (vinfo_for_stmt (last_stmt)))
1660 return NULL;
1662 if (gimple_assign_rhs_code (last_stmt) != LSHIFT_EXPR)
1663 return NULL;
1665 oprnd0 = gimple_assign_rhs1 (last_stmt);
1666 oprnd1 = gimple_assign_rhs2 (last_stmt);
1667 if (TREE_CODE (oprnd0) != SSA_NAME || TREE_CODE (oprnd1) != INTEGER_CST)
1668 return NULL;
1670 /* Check operand 0: it has to be defined by a type promotion. */
1671 if (!type_conversion_p (oprnd0, last_stmt, false, &half_type0, &def_stmt0,
1672 &promotion)
1673 || !promotion)
1674 return NULL;
1676 /* Check operand 1: has to be positive. We check that it fits the type
1677 in vect_handle_widen_op_by_const (). */
1678 if (tree_int_cst_compare (oprnd1, size_zero_node) <= 0)
1679 return NULL;
1681 oprnd0 = gimple_assign_rhs1 (def_stmt0);
1682 type = gimple_expr_type (last_stmt);
1684 /* Check for subsequent conversion to another type. */
1685 use_stmt = vect_single_imm_use (last_stmt);
1686 if (use_stmt && is_gimple_assign (use_stmt)
1687 && CONVERT_EXPR_CODE_P (gimple_assign_rhs_code (use_stmt))
1688 && !STMT_VINFO_IN_PATTERN_P (vinfo_for_stmt (use_stmt)))
1690 tree use_lhs = gimple_assign_lhs (use_stmt);
1691 tree use_type = TREE_TYPE (use_lhs);
1693 if (INTEGRAL_TYPE_P (use_type)
1694 && TYPE_PRECISION (use_type) <= TYPE_PRECISION (type))
1696 last_stmt = use_stmt;
1697 type = use_type;
1701 /* Check if this a widening operation. */
1702 gimple *wstmt = NULL;
1703 if (!vect_handle_widen_op_by_const (last_stmt, LSHIFT_EXPR, oprnd1,
1704 &oprnd0, &wstmt,
1705 type, &half_type0, def_stmt0))
1706 return NULL;
1708 /* Pattern detected. */
1709 if (dump_enabled_p ())
1710 dump_printf_loc (MSG_NOTE, vect_location,
1711 "vect_recog_widen_shift_pattern: detected:\n");
1713 /* Check target support. */
1714 vectype = get_vectype_for_scalar_type (half_type0);
1715 vectype_out = get_vectype_for_scalar_type (type);
1717 if (!vectype
1718 || !vectype_out
1719 || !supportable_widening_operation (WIDEN_LSHIFT_EXPR, last_stmt,
1720 vectype_out, vectype,
1721 &dummy_code, &dummy_code,
1722 &dummy_int, &dummy_vec))
1723 return NULL;
1725 *type_in = vectype;
1726 *type_out = vectype_out;
1728 /* Pattern supported. Create a stmt to be used to replace the pattern. */
1729 var = vect_recog_temp_ssa_var (type, NULL);
1730 pattern_stmt =
1731 gimple_build_assign (var, WIDEN_LSHIFT_EXPR, oprnd0, oprnd1);
1732 if (wstmt)
1734 stmt_vec_info stmt_vinfo = vinfo_for_stmt (last_stmt);
1735 new_pattern_def_seq (stmt_vinfo, wstmt);
1736 stmt_vec_info new_stmt_info
1737 = new_stmt_vec_info (wstmt, stmt_vinfo->vinfo);
1738 set_vinfo_for_stmt (wstmt, new_stmt_info);
1739 STMT_VINFO_VECTYPE (new_stmt_info) = vectype;
1742 if (dump_enabled_p ())
1743 dump_gimple_stmt_loc (MSG_NOTE, vect_location, TDF_SLIM, pattern_stmt, 0);
1745 stmts->safe_push (last_stmt);
1746 return pattern_stmt;
1749 /* Detect a rotate pattern wouldn't be otherwise vectorized:
1751 type a_t, b_t, c_t;
1753 S0 a_t = b_t r<< c_t;
1755 Input/Output:
1757 * STMTS: Contains a stmt from which the pattern search begins,
1758 i.e. the shift/rotate stmt. The original stmt (S0) is replaced
1759 with a sequence:
1761 S1 d_t = -c_t;
1762 S2 e_t = d_t & (B - 1);
1763 S3 f_t = b_t << c_t;
1764 S4 g_t = b_t >> e_t;
1765 S0 a_t = f_t | g_t;
1767 where B is element bitsize of type.
1769 Output:
1771 * TYPE_IN: The type of the input arguments to the pattern.
1773 * TYPE_OUT: The type of the output of this pattern.
1775 * Return value: A new stmt that will be used to replace the rotate
1776 S0 stmt. */
1778 static gimple *
1779 vect_recog_rotate_pattern (vec<gimple *> *stmts, tree *type_in, tree *type_out)
1781 gimple *last_stmt = stmts->pop ();
1782 tree oprnd0, oprnd1, lhs, var, var1, var2, vectype, type, stype, def, def2;
1783 gimple *pattern_stmt, *def_stmt;
1784 enum tree_code rhs_code;
1785 stmt_vec_info stmt_vinfo = vinfo_for_stmt (last_stmt);
1786 vec_info *vinfo = stmt_vinfo->vinfo;
1787 enum vect_def_type dt;
1788 optab optab1, optab2;
1789 edge ext_def = NULL;
1791 if (!is_gimple_assign (last_stmt))
1792 return NULL;
1794 rhs_code = gimple_assign_rhs_code (last_stmt);
1795 switch (rhs_code)
1797 case LROTATE_EXPR:
1798 case RROTATE_EXPR:
1799 break;
1800 default:
1801 return NULL;
1804 if (STMT_VINFO_IN_PATTERN_P (stmt_vinfo))
1805 return NULL;
1807 lhs = gimple_assign_lhs (last_stmt);
1808 oprnd0 = gimple_assign_rhs1 (last_stmt);
1809 type = TREE_TYPE (oprnd0);
1810 oprnd1 = gimple_assign_rhs2 (last_stmt);
1811 if (TREE_CODE (oprnd0) != SSA_NAME
1812 || TYPE_PRECISION (TREE_TYPE (lhs)) != TYPE_PRECISION (type)
1813 || !INTEGRAL_TYPE_P (type)
1814 || !TYPE_UNSIGNED (type))
1815 return NULL;
1817 if (!vect_is_simple_use (oprnd1, vinfo, &def_stmt, &dt))
1818 return NULL;
1820 if (dt != vect_internal_def
1821 && dt != vect_constant_def
1822 && dt != vect_external_def)
1823 return NULL;
1825 vectype = get_vectype_for_scalar_type (type);
1826 if (vectype == NULL_TREE)
1827 return NULL;
1829 /* If vector/vector or vector/scalar rotate is supported by the target,
1830 don't do anything here. */
1831 optab1 = optab_for_tree_code (rhs_code, vectype, optab_vector);
1832 if (optab1
1833 && optab_handler (optab1, TYPE_MODE (vectype)) != CODE_FOR_nothing)
1834 return NULL;
1836 if (is_a <bb_vec_info> (vinfo) || dt != vect_internal_def)
1838 optab2 = optab_for_tree_code (rhs_code, vectype, optab_scalar);
1839 if (optab2
1840 && optab_handler (optab2, TYPE_MODE (vectype)) != CODE_FOR_nothing)
1841 return NULL;
1844 /* If vector/vector or vector/scalar shifts aren't supported by the target,
1845 don't do anything here either. */
1846 optab1 = optab_for_tree_code (LSHIFT_EXPR, vectype, optab_vector);
1847 optab2 = optab_for_tree_code (RSHIFT_EXPR, vectype, optab_vector);
1848 if (!optab1
1849 || optab_handler (optab1, TYPE_MODE (vectype)) == CODE_FOR_nothing
1850 || !optab2
1851 || optab_handler (optab2, TYPE_MODE (vectype)) == CODE_FOR_nothing)
1853 if (! is_a <bb_vec_info> (vinfo) && dt == vect_internal_def)
1854 return NULL;
1855 optab1 = optab_for_tree_code (LSHIFT_EXPR, vectype, optab_scalar);
1856 optab2 = optab_for_tree_code (RSHIFT_EXPR, vectype, optab_scalar);
1857 if (!optab1
1858 || optab_handler (optab1, TYPE_MODE (vectype)) == CODE_FOR_nothing
1859 || !optab2
1860 || optab_handler (optab2, TYPE_MODE (vectype)) == CODE_FOR_nothing)
1861 return NULL;
1864 *type_in = vectype;
1865 *type_out = vectype;
1866 if (*type_in == NULL_TREE)
1867 return NULL;
1869 if (dt == vect_external_def
1870 && TREE_CODE (oprnd1) == SSA_NAME
1871 && is_a <loop_vec_info> (vinfo))
1873 struct loop *loop = as_a <loop_vec_info> (vinfo)->loop;
1874 ext_def = loop_preheader_edge (loop);
1875 if (!SSA_NAME_IS_DEFAULT_DEF (oprnd1))
1877 basic_block bb = gimple_bb (SSA_NAME_DEF_STMT (oprnd1));
1878 if (bb == NULL
1879 || !dominated_by_p (CDI_DOMINATORS, ext_def->dest, bb))
1880 ext_def = NULL;
1884 def = NULL_TREE;
1885 if (TREE_CODE (oprnd1) == INTEGER_CST
1886 || TYPE_MODE (TREE_TYPE (oprnd1)) == TYPE_MODE (type))
1887 def = oprnd1;
1888 else if (def_stmt && gimple_assign_cast_p (def_stmt))
1890 tree rhs1 = gimple_assign_rhs1 (def_stmt);
1891 if (TYPE_MODE (TREE_TYPE (rhs1)) == TYPE_MODE (type)
1892 && TYPE_PRECISION (TREE_TYPE (rhs1))
1893 == TYPE_PRECISION (type))
1894 def = rhs1;
1897 STMT_VINFO_PATTERN_DEF_SEQ (stmt_vinfo) = NULL;
1898 if (def == NULL_TREE)
1900 def = vect_recog_temp_ssa_var (type, NULL);
1901 def_stmt = gimple_build_assign (def, NOP_EXPR, oprnd1);
1902 if (ext_def)
1904 basic_block new_bb
1905 = gsi_insert_on_edge_immediate (ext_def, def_stmt);
1906 gcc_assert (!new_bb);
1908 else
1909 append_pattern_def_seq (stmt_vinfo, def_stmt);
1911 stype = TREE_TYPE (def);
1913 if (TREE_CODE (def) == INTEGER_CST)
1915 if (!tree_fits_uhwi_p (def)
1916 || tree_to_uhwi (def) >= GET_MODE_PRECISION (TYPE_MODE (type))
1917 || integer_zerop (def))
1918 return NULL;
1919 def2 = build_int_cst (stype,
1920 GET_MODE_PRECISION (TYPE_MODE (type))
1921 - tree_to_uhwi (def));
1923 else
1925 tree vecstype = get_vectype_for_scalar_type (stype);
1926 stmt_vec_info def_stmt_vinfo;
1928 if (vecstype == NULL_TREE)
1929 return NULL;
1930 def2 = vect_recog_temp_ssa_var (stype, NULL);
1931 def_stmt = gimple_build_assign (def2, NEGATE_EXPR, def);
1932 if (ext_def)
1934 basic_block new_bb
1935 = gsi_insert_on_edge_immediate (ext_def, def_stmt);
1936 gcc_assert (!new_bb);
1938 else
1940 def_stmt_vinfo = new_stmt_vec_info (def_stmt, vinfo);
1941 set_vinfo_for_stmt (def_stmt, def_stmt_vinfo);
1942 STMT_VINFO_VECTYPE (def_stmt_vinfo) = vecstype;
1943 append_pattern_def_seq (stmt_vinfo, def_stmt);
1946 def2 = vect_recog_temp_ssa_var (stype, NULL);
1947 tree mask
1948 = build_int_cst (stype, GET_MODE_PRECISION (TYPE_MODE (stype)) - 1);
1949 def_stmt = gimple_build_assign (def2, BIT_AND_EXPR,
1950 gimple_assign_lhs (def_stmt), mask);
1951 if (ext_def)
1953 basic_block new_bb
1954 = gsi_insert_on_edge_immediate (ext_def, def_stmt);
1955 gcc_assert (!new_bb);
1957 else
1959 def_stmt_vinfo = new_stmt_vec_info (def_stmt, vinfo);
1960 set_vinfo_for_stmt (def_stmt, def_stmt_vinfo);
1961 STMT_VINFO_VECTYPE (def_stmt_vinfo) = vecstype;
1962 append_pattern_def_seq (stmt_vinfo, def_stmt);
1966 var1 = vect_recog_temp_ssa_var (type, NULL);
1967 def_stmt = gimple_build_assign (var1, rhs_code == LROTATE_EXPR
1968 ? LSHIFT_EXPR : RSHIFT_EXPR,
1969 oprnd0, def);
1970 append_pattern_def_seq (stmt_vinfo, def_stmt);
1972 var2 = vect_recog_temp_ssa_var (type, NULL);
1973 def_stmt = gimple_build_assign (var2, rhs_code == LROTATE_EXPR
1974 ? RSHIFT_EXPR : LSHIFT_EXPR,
1975 oprnd0, def2);
1976 append_pattern_def_seq (stmt_vinfo, def_stmt);
1978 /* Pattern detected. */
1979 if (dump_enabled_p ())
1980 dump_printf_loc (MSG_NOTE, vect_location,
1981 "vect_recog_rotate_pattern: detected:\n");
1983 /* Pattern supported. Create a stmt to be used to replace the pattern. */
1984 var = vect_recog_temp_ssa_var (type, NULL);
1985 pattern_stmt = gimple_build_assign (var, BIT_IOR_EXPR, var1, var2);
1987 if (dump_enabled_p ())
1988 dump_gimple_stmt_loc (MSG_NOTE, vect_location, TDF_SLIM, pattern_stmt, 0);
1990 stmts->safe_push (last_stmt);
1991 return pattern_stmt;
1994 /* Detect a vector by vector shift pattern that wouldn't be otherwise
1995 vectorized:
1997 type a_t;
1998 TYPE b_T, res_T;
2000 S1 a_t = ;
2001 S2 b_T = ;
2002 S3 res_T = b_T op a_t;
2004 where type 'TYPE' is a type with different size than 'type',
2005 and op is <<, >> or rotate.
2007 Also detect cases:
2009 type a_t;
2010 TYPE b_T, c_T, res_T;
2012 S0 c_T = ;
2013 S1 a_t = (type) c_T;
2014 S2 b_T = ;
2015 S3 res_T = b_T op a_t;
2017 Input/Output:
2019 * STMTS: Contains a stmt from which the pattern search begins,
2020 i.e. the shift/rotate stmt. The original stmt (S3) is replaced
2021 with a shift/rotate which has same type on both operands, in the
2022 second case just b_T op c_T, in the first case with added cast
2023 from a_t to c_T in STMT_VINFO_PATTERN_DEF_SEQ.
2025 Output:
2027 * TYPE_IN: The type of the input arguments to the pattern.
2029 * TYPE_OUT: The type of the output of this pattern.
2031 * Return value: A new stmt that will be used to replace the shift/rotate
2032 S3 stmt. */
2034 static gimple *
2035 vect_recog_vector_vector_shift_pattern (vec<gimple *> *stmts,
2036 tree *type_in, tree *type_out)
2038 gimple *last_stmt = stmts->pop ();
2039 tree oprnd0, oprnd1, lhs, var;
2040 gimple *pattern_stmt, *def_stmt;
2041 enum tree_code rhs_code;
2042 stmt_vec_info stmt_vinfo = vinfo_for_stmt (last_stmt);
2043 vec_info *vinfo = stmt_vinfo->vinfo;
2044 enum vect_def_type dt;
2046 if (!is_gimple_assign (last_stmt))
2047 return NULL;
2049 rhs_code = gimple_assign_rhs_code (last_stmt);
2050 switch (rhs_code)
2052 case LSHIFT_EXPR:
2053 case RSHIFT_EXPR:
2054 case LROTATE_EXPR:
2055 case RROTATE_EXPR:
2056 break;
2057 default:
2058 return NULL;
2061 if (STMT_VINFO_IN_PATTERN_P (stmt_vinfo))
2062 return NULL;
2064 lhs = gimple_assign_lhs (last_stmt);
2065 oprnd0 = gimple_assign_rhs1 (last_stmt);
2066 oprnd1 = gimple_assign_rhs2 (last_stmt);
2067 if (TREE_CODE (oprnd0) != SSA_NAME
2068 || TREE_CODE (oprnd1) != SSA_NAME
2069 || TYPE_MODE (TREE_TYPE (oprnd0)) == TYPE_MODE (TREE_TYPE (oprnd1))
2070 || TYPE_PRECISION (TREE_TYPE (oprnd1))
2071 != GET_MODE_PRECISION (TYPE_MODE (TREE_TYPE (oprnd1)))
2072 || TYPE_PRECISION (TREE_TYPE (lhs))
2073 != TYPE_PRECISION (TREE_TYPE (oprnd0)))
2074 return NULL;
2076 if (!vect_is_simple_use (oprnd1, vinfo, &def_stmt, &dt))
2077 return NULL;
2079 if (dt != vect_internal_def)
2080 return NULL;
2082 *type_in = get_vectype_for_scalar_type (TREE_TYPE (oprnd0));
2083 *type_out = *type_in;
2084 if (*type_in == NULL_TREE)
2085 return NULL;
2087 tree def = NULL_TREE;
2088 stmt_vec_info def_vinfo = vinfo_for_stmt (def_stmt);
2089 if (!STMT_VINFO_IN_PATTERN_P (def_vinfo) && gimple_assign_cast_p (def_stmt))
2091 tree rhs1 = gimple_assign_rhs1 (def_stmt);
2092 if (TYPE_MODE (TREE_TYPE (rhs1)) == TYPE_MODE (TREE_TYPE (oprnd0))
2093 && TYPE_PRECISION (TREE_TYPE (rhs1))
2094 == TYPE_PRECISION (TREE_TYPE (oprnd0)))
2096 if (TYPE_PRECISION (TREE_TYPE (oprnd1))
2097 >= TYPE_PRECISION (TREE_TYPE (rhs1)))
2098 def = rhs1;
2099 else
2101 tree mask
2102 = build_low_bits_mask (TREE_TYPE (rhs1),
2103 TYPE_PRECISION (TREE_TYPE (oprnd1)));
2104 def = vect_recog_temp_ssa_var (TREE_TYPE (rhs1), NULL);
2105 def_stmt = gimple_build_assign (def, BIT_AND_EXPR, rhs1, mask);
2106 new_pattern_def_seq (stmt_vinfo, def_stmt);
2111 if (def == NULL_TREE)
2113 def = vect_recog_temp_ssa_var (TREE_TYPE (oprnd0), NULL);
2114 def_stmt = gimple_build_assign (def, NOP_EXPR, oprnd1);
2115 new_pattern_def_seq (stmt_vinfo, def_stmt);
2118 /* Pattern detected. */
2119 if (dump_enabled_p ())
2120 dump_printf_loc (MSG_NOTE, vect_location,
2121 "vect_recog_vector_vector_shift_pattern: detected:\n");
2123 /* Pattern supported. Create a stmt to be used to replace the pattern. */
2124 var = vect_recog_temp_ssa_var (TREE_TYPE (oprnd0), NULL);
2125 pattern_stmt = gimple_build_assign (var, rhs_code, oprnd0, def);
2127 if (dump_enabled_p ())
2128 dump_gimple_stmt_loc (MSG_NOTE, vect_location, TDF_SLIM, pattern_stmt, 0);
2130 stmts->safe_push (last_stmt);
2131 return pattern_stmt;
2134 /* Return true iff the target has a vector optab implementing the operation
2135 CODE on type VECTYPE. */
2137 static bool
2138 target_has_vecop_for_code (tree_code code, tree vectype)
2140 optab voptab = optab_for_tree_code (code, vectype, optab_vector);
2141 return voptab
2142 && optab_handler (voptab, TYPE_MODE (vectype)) != CODE_FOR_nothing;
2145 /* Verify that the target has optabs of VECTYPE to perform all the steps
2146 needed by the multiplication-by-immediate synthesis algorithm described by
2147 ALG and VAR. If SYNTH_SHIFT_P is true ensure that vector addition is
2148 present. Return true iff the target supports all the steps. */
2150 static bool
2151 target_supports_mult_synth_alg (struct algorithm *alg, mult_variant var,
2152 tree vectype, bool synth_shift_p)
2154 if (alg->op[0] != alg_zero && alg->op[0] != alg_m)
2155 return false;
2157 bool supports_vminus = target_has_vecop_for_code (MINUS_EXPR, vectype);
2158 bool supports_vplus = target_has_vecop_for_code (PLUS_EXPR, vectype);
2160 if (var == negate_variant
2161 && !target_has_vecop_for_code (NEGATE_EXPR, vectype))
2162 return false;
2164 /* If we must synthesize shifts with additions make sure that vector
2165 addition is available. */
2166 if ((var == add_variant || synth_shift_p) && !supports_vplus)
2167 return false;
2169 for (int i = 1; i < alg->ops; i++)
2171 switch (alg->op[i])
2173 case alg_shift:
2174 break;
2175 case alg_add_t_m2:
2176 case alg_add_t2_m:
2177 case alg_add_factor:
2178 if (!supports_vplus)
2179 return false;
2180 break;
2181 case alg_sub_t_m2:
2182 case alg_sub_t2_m:
2183 case alg_sub_factor:
2184 if (!supports_vminus)
2185 return false;
2186 break;
2187 case alg_unknown:
2188 case alg_m:
2189 case alg_zero:
2190 case alg_impossible:
2191 return false;
2192 default:
2193 gcc_unreachable ();
2197 return true;
2200 /* Synthesize a left shift of OP by AMNT bits using a series of additions and
2201 putting the final result in DEST. Append all statements but the last into
2202 VINFO. Return the last statement. */
2204 static gimple *
2205 synth_lshift_by_additions (tree dest, tree op, HOST_WIDE_INT amnt,
2206 stmt_vec_info vinfo)
2208 HOST_WIDE_INT i;
2209 tree itype = TREE_TYPE (op);
2210 tree prev_res = op;
2211 gcc_assert (amnt >= 0);
2212 for (i = 0; i < amnt; i++)
2214 tree tmp_var = (i < amnt - 1) ? vect_recog_temp_ssa_var (itype, NULL)
2215 : dest;
2216 gimple *stmt
2217 = gimple_build_assign (tmp_var, PLUS_EXPR, prev_res, prev_res);
2218 prev_res = tmp_var;
2219 if (i < amnt - 1)
2220 append_pattern_def_seq (vinfo, stmt);
2221 else
2222 return stmt;
2224 gcc_unreachable ();
2225 return NULL;
2228 /* Helper for vect_synth_mult_by_constant. Apply a binary operation
2229 CODE to operands OP1 and OP2, creating a new temporary SSA var in
2230 the process if necessary. Append the resulting assignment statements
2231 to the sequence in STMT_VINFO. Return the SSA variable that holds the
2232 result of the binary operation. If SYNTH_SHIFT_P is true synthesize
2233 left shifts using additions. */
2235 static tree
2236 apply_binop_and_append_stmt (tree_code code, tree op1, tree op2,
2237 stmt_vec_info stmt_vinfo, bool synth_shift_p)
2239 if (integer_zerop (op2)
2240 && (code == LSHIFT_EXPR
2241 || code == PLUS_EXPR))
2243 gcc_assert (TREE_CODE (op1) == SSA_NAME);
2244 return op1;
2247 gimple *stmt;
2248 tree itype = TREE_TYPE (op1);
2249 tree tmp_var = vect_recog_temp_ssa_var (itype, NULL);
2251 if (code == LSHIFT_EXPR
2252 && synth_shift_p)
2254 stmt = synth_lshift_by_additions (tmp_var, op1, TREE_INT_CST_LOW (op2),
2255 stmt_vinfo);
2256 append_pattern_def_seq (stmt_vinfo, stmt);
2257 return tmp_var;
2260 stmt = gimple_build_assign (tmp_var, code, op1, op2);
2261 append_pattern_def_seq (stmt_vinfo, stmt);
2262 return tmp_var;
2265 /* Synthesize a multiplication of OP by an INTEGER_CST VAL using shifts
2266 and simple arithmetic operations to be vectorized. Record the statements
2267 produced in STMT_VINFO and return the last statement in the sequence or
2268 NULL if it's not possible to synthesize such a multiplication.
2269 This function mirrors the behavior of expand_mult_const in expmed.c but
2270 works on tree-ssa form. */
2272 static gimple *
2273 vect_synth_mult_by_constant (tree op, tree val,
2274 stmt_vec_info stmt_vinfo)
2276 tree itype = TREE_TYPE (op);
2277 machine_mode mode = TYPE_MODE (itype);
2278 struct algorithm alg;
2279 mult_variant variant;
2280 if (!tree_fits_shwi_p (val))
2281 return NULL;
2283 /* Multiplication synthesis by shifts, adds and subs can introduce
2284 signed overflow where the original operation didn't. Perform the
2285 operations on an unsigned type and cast back to avoid this.
2286 In the future we may want to relax this for synthesis algorithms
2287 that we can prove do not cause unexpected overflow. */
2288 bool cast_to_unsigned_p = !TYPE_OVERFLOW_WRAPS (itype);
2290 tree multtype = cast_to_unsigned_p ? unsigned_type_for (itype) : itype;
2292 /* Targets that don't support vector shifts but support vector additions
2293 can synthesize shifts that way. */
2294 bool synth_shift_p = !vect_supportable_shift (LSHIFT_EXPR, multtype);
2296 HOST_WIDE_INT hwval = tree_to_shwi (val);
2297 /* Use MAX_COST here as we don't want to limit the sequence on rtx costs.
2298 The vectorizer's benefit analysis will decide whether it's beneficial
2299 to do this. */
2300 bool possible = choose_mult_variant (mode, hwval, &alg,
2301 &variant, MAX_COST);
2302 if (!possible)
2303 return NULL;
2305 tree vectype = get_vectype_for_scalar_type (multtype);
2307 if (!vectype
2308 || !target_supports_mult_synth_alg (&alg, variant,
2309 vectype, synth_shift_p))
2310 return NULL;
2312 tree accumulator;
2314 /* Clear out the sequence of statements so we can populate it below. */
2315 STMT_VINFO_PATTERN_DEF_SEQ (stmt_vinfo) = NULL;
2316 gimple *stmt = NULL;
2318 if (cast_to_unsigned_p)
2320 tree tmp_op = vect_recog_temp_ssa_var (multtype, NULL);
2321 stmt = gimple_build_assign (tmp_op, CONVERT_EXPR, op);
2322 append_pattern_def_seq (stmt_vinfo, stmt);
2323 op = tmp_op;
2326 if (alg.op[0] == alg_zero)
2327 accumulator = build_int_cst (multtype, 0);
2328 else
2329 accumulator = op;
2331 bool needs_fixup = (variant == negate_variant)
2332 || (variant == add_variant);
2334 for (int i = 1; i < alg.ops; i++)
2336 tree shft_log = build_int_cst (multtype, alg.log[i]);
2337 tree accum_tmp = vect_recog_temp_ssa_var (multtype, NULL);
2338 tree tmp_var = NULL_TREE;
2340 switch (alg.op[i])
2342 case alg_shift:
2343 if (synth_shift_p)
2344 stmt
2345 = synth_lshift_by_additions (accum_tmp, accumulator, alg.log[i],
2346 stmt_vinfo);
2347 else
2348 stmt = gimple_build_assign (accum_tmp, LSHIFT_EXPR, accumulator,
2349 shft_log);
2350 break;
2351 case alg_add_t_m2:
2352 tmp_var
2353 = apply_binop_and_append_stmt (LSHIFT_EXPR, op, shft_log,
2354 stmt_vinfo, synth_shift_p);
2355 stmt = gimple_build_assign (accum_tmp, PLUS_EXPR, accumulator,
2356 tmp_var);
2357 break;
2358 case alg_sub_t_m2:
2359 tmp_var = apply_binop_and_append_stmt (LSHIFT_EXPR, op,
2360 shft_log, stmt_vinfo,
2361 synth_shift_p);
2362 /* In some algorithms the first step involves zeroing the
2363 accumulator. If subtracting from such an accumulator
2364 just emit the negation directly. */
2365 if (integer_zerop (accumulator))
2366 stmt = gimple_build_assign (accum_tmp, NEGATE_EXPR, tmp_var);
2367 else
2368 stmt = gimple_build_assign (accum_tmp, MINUS_EXPR, accumulator,
2369 tmp_var);
2370 break;
2371 case alg_add_t2_m:
2372 tmp_var
2373 = apply_binop_and_append_stmt (LSHIFT_EXPR, accumulator, shft_log,
2374 stmt_vinfo, synth_shift_p);
2375 stmt = gimple_build_assign (accum_tmp, PLUS_EXPR, tmp_var, op);
2376 break;
2377 case alg_sub_t2_m:
2378 tmp_var
2379 = apply_binop_and_append_stmt (LSHIFT_EXPR, accumulator, shft_log,
2380 stmt_vinfo, synth_shift_p);
2381 stmt = gimple_build_assign (accum_tmp, MINUS_EXPR, tmp_var, op);
2382 break;
2383 case alg_add_factor:
2384 tmp_var
2385 = apply_binop_and_append_stmt (LSHIFT_EXPR, accumulator, shft_log,
2386 stmt_vinfo, synth_shift_p);
2387 stmt = gimple_build_assign (accum_tmp, PLUS_EXPR, accumulator,
2388 tmp_var);
2389 break;
2390 case alg_sub_factor:
2391 tmp_var
2392 = apply_binop_and_append_stmt (LSHIFT_EXPR, accumulator, shft_log,
2393 stmt_vinfo, synth_shift_p);
2394 stmt = gimple_build_assign (accum_tmp, MINUS_EXPR, tmp_var,
2395 accumulator);
2396 break;
2397 default:
2398 gcc_unreachable ();
2400 /* We don't want to append the last stmt in the sequence to stmt_vinfo
2401 but rather return it directly. */
2403 if ((i < alg.ops - 1) || needs_fixup || cast_to_unsigned_p)
2404 append_pattern_def_seq (stmt_vinfo, stmt);
2405 accumulator = accum_tmp;
2407 if (variant == negate_variant)
2409 tree accum_tmp = vect_recog_temp_ssa_var (multtype, NULL);
2410 stmt = gimple_build_assign (accum_tmp, NEGATE_EXPR, accumulator);
2411 accumulator = accum_tmp;
2412 if (cast_to_unsigned_p)
2413 append_pattern_def_seq (stmt_vinfo, stmt);
2415 else if (variant == add_variant)
2417 tree accum_tmp = vect_recog_temp_ssa_var (multtype, NULL);
2418 stmt = gimple_build_assign (accum_tmp, PLUS_EXPR, accumulator, op);
2419 accumulator = accum_tmp;
2420 if (cast_to_unsigned_p)
2421 append_pattern_def_seq (stmt_vinfo, stmt);
2423 /* Move back to a signed if needed. */
2424 if (cast_to_unsigned_p)
2426 tree accum_tmp = vect_recog_temp_ssa_var (itype, NULL);
2427 stmt = gimple_build_assign (accum_tmp, CONVERT_EXPR, accumulator);
2430 return stmt;
2433 /* Detect multiplication by constant and convert it into a sequence of
2434 shifts and additions, subtractions, negations. We reuse the
2435 choose_mult_variant algorithms from expmed.c
2437 Input/Output:
2439 STMTS: Contains a stmt from which the pattern search begins,
2440 i.e. the mult stmt.
2442 Output:
2444 * TYPE_IN: The type of the input arguments to the pattern.
2446 * TYPE_OUT: The type of the output of this pattern.
2448 * Return value: A new stmt that will be used to replace
2449 the multiplication. */
2451 static gimple *
2452 vect_recog_mult_pattern (vec<gimple *> *stmts,
2453 tree *type_in, tree *type_out)
2455 gimple *last_stmt = stmts->pop ();
2456 tree oprnd0, oprnd1, vectype, itype;
2457 gimple *pattern_stmt;
2458 stmt_vec_info stmt_vinfo = vinfo_for_stmt (last_stmt);
2460 if (!is_gimple_assign (last_stmt))
2461 return NULL;
2463 if (gimple_assign_rhs_code (last_stmt) != MULT_EXPR)
2464 return NULL;
2466 oprnd0 = gimple_assign_rhs1 (last_stmt);
2467 oprnd1 = gimple_assign_rhs2 (last_stmt);
2468 itype = TREE_TYPE (oprnd0);
2470 if (TREE_CODE (oprnd0) != SSA_NAME
2471 || TREE_CODE (oprnd1) != INTEGER_CST
2472 || !INTEGRAL_TYPE_P (itype)
2473 || TYPE_PRECISION (itype) != GET_MODE_PRECISION (TYPE_MODE (itype)))
2474 return NULL;
2476 vectype = get_vectype_for_scalar_type (itype);
2477 if (vectype == NULL_TREE)
2478 return NULL;
2480 /* If the target can handle vectorized multiplication natively,
2481 don't attempt to optimize this. */
2482 optab mul_optab = optab_for_tree_code (MULT_EXPR, vectype, optab_default);
2483 if (mul_optab != unknown_optab)
2485 machine_mode vec_mode = TYPE_MODE (vectype);
2486 int icode = (int) optab_handler (mul_optab, vec_mode);
2487 if (icode != CODE_FOR_nothing)
2488 return NULL;
2491 pattern_stmt = vect_synth_mult_by_constant (oprnd0, oprnd1, stmt_vinfo);
2492 if (!pattern_stmt)
2493 return NULL;
2495 /* Pattern detected. */
2496 if (dump_enabled_p ())
2497 dump_printf_loc (MSG_NOTE, vect_location,
2498 "vect_recog_mult_pattern: detected:\n");
2500 if (dump_enabled_p ())
2501 dump_gimple_stmt_loc (MSG_NOTE, vect_location, TDF_SLIM,
2502 pattern_stmt,0);
2504 stmts->safe_push (last_stmt);
2505 *type_in = vectype;
2506 *type_out = vectype;
2508 return pattern_stmt;
2511 /* Detect a signed division by a constant that wouldn't be
2512 otherwise vectorized:
2514 type a_t, b_t;
2516 S1 a_t = b_t / N;
2518 where type 'type' is an integral type and N is a constant.
2520 Similarly handle modulo by a constant:
2522 S4 a_t = b_t % N;
2524 Input/Output:
2526 * STMTS: Contains a stmt from which the pattern search begins,
2527 i.e. the division stmt. S1 is replaced by if N is a power
2528 of two constant and type is signed:
2529 S3 y_t = b_t < 0 ? N - 1 : 0;
2530 S2 x_t = b_t + y_t;
2531 S1' a_t = x_t >> log2 (N);
2533 S4 is replaced if N is a power of two constant and
2534 type is signed by (where *_T temporaries have unsigned type):
2535 S9 y_T = b_t < 0 ? -1U : 0U;
2536 S8 z_T = y_T >> (sizeof (type_t) * CHAR_BIT - log2 (N));
2537 S7 z_t = (type) z_T;
2538 S6 w_t = b_t + z_t;
2539 S5 x_t = w_t & (N - 1);
2540 S4' a_t = x_t - z_t;
2542 Output:
2544 * TYPE_IN: The type of the input arguments to the pattern.
2546 * TYPE_OUT: The type of the output of this pattern.
2548 * Return value: A new stmt that will be used to replace the division
2549 S1 or modulo S4 stmt. */
2551 static gimple *
2552 vect_recog_divmod_pattern (vec<gimple *> *stmts,
2553 tree *type_in, tree *type_out)
2555 gimple *last_stmt = stmts->pop ();
2556 tree oprnd0, oprnd1, vectype, itype, cond;
2557 gimple *pattern_stmt, *def_stmt;
2558 enum tree_code rhs_code;
2559 stmt_vec_info stmt_vinfo = vinfo_for_stmt (last_stmt);
2560 vec_info *vinfo = stmt_vinfo->vinfo;
2561 optab optab;
2562 tree q;
2563 int dummy_int, prec;
2564 stmt_vec_info def_stmt_vinfo;
2566 if (!is_gimple_assign (last_stmt))
2567 return NULL;
2569 rhs_code = gimple_assign_rhs_code (last_stmt);
2570 switch (rhs_code)
2572 case TRUNC_DIV_EXPR:
2573 case TRUNC_MOD_EXPR:
2574 break;
2575 default:
2576 return NULL;
2579 if (STMT_VINFO_IN_PATTERN_P (stmt_vinfo))
2580 return NULL;
2582 oprnd0 = gimple_assign_rhs1 (last_stmt);
2583 oprnd1 = gimple_assign_rhs2 (last_stmt);
2584 itype = TREE_TYPE (oprnd0);
2585 if (TREE_CODE (oprnd0) != SSA_NAME
2586 || TREE_CODE (oprnd1) != INTEGER_CST
2587 || TREE_CODE (itype) != INTEGER_TYPE
2588 || TYPE_PRECISION (itype) != GET_MODE_PRECISION (TYPE_MODE (itype)))
2589 return NULL;
2591 vectype = get_vectype_for_scalar_type (itype);
2592 if (vectype == NULL_TREE)
2593 return NULL;
2595 /* If the target can handle vectorized division or modulo natively,
2596 don't attempt to optimize this. */
2597 optab = optab_for_tree_code (rhs_code, vectype, optab_default);
2598 if (optab != unknown_optab)
2600 machine_mode vec_mode = TYPE_MODE (vectype);
2601 int icode = (int) optab_handler (optab, vec_mode);
2602 if (icode != CODE_FOR_nothing)
2603 return NULL;
2606 prec = TYPE_PRECISION (itype);
2607 if (integer_pow2p (oprnd1))
2609 if (TYPE_UNSIGNED (itype) || tree_int_cst_sgn (oprnd1) != 1)
2610 return NULL;
2612 /* Pattern detected. */
2613 if (dump_enabled_p ())
2614 dump_printf_loc (MSG_NOTE, vect_location,
2615 "vect_recog_divmod_pattern: detected:\n");
2617 cond = build2 (LT_EXPR, boolean_type_node, oprnd0,
2618 build_int_cst (itype, 0));
2619 if (rhs_code == TRUNC_DIV_EXPR)
2621 tree var = vect_recog_temp_ssa_var (itype, NULL);
2622 tree shift;
2623 def_stmt
2624 = gimple_build_assign (var, COND_EXPR, cond,
2625 fold_build2 (MINUS_EXPR, itype, oprnd1,
2626 build_int_cst (itype, 1)),
2627 build_int_cst (itype, 0));
2628 new_pattern_def_seq (stmt_vinfo, def_stmt);
2629 var = vect_recog_temp_ssa_var (itype, NULL);
2630 def_stmt
2631 = gimple_build_assign (var, PLUS_EXPR, oprnd0,
2632 gimple_assign_lhs (def_stmt));
2633 append_pattern_def_seq (stmt_vinfo, def_stmt);
2635 shift = build_int_cst (itype, tree_log2 (oprnd1));
2636 pattern_stmt
2637 = gimple_build_assign (vect_recog_temp_ssa_var (itype, NULL),
2638 RSHIFT_EXPR, var, shift);
2640 else
2642 tree signmask;
2643 STMT_VINFO_PATTERN_DEF_SEQ (stmt_vinfo) = NULL;
2644 if (compare_tree_int (oprnd1, 2) == 0)
2646 signmask = vect_recog_temp_ssa_var (itype, NULL);
2647 def_stmt = gimple_build_assign (signmask, COND_EXPR, cond,
2648 build_int_cst (itype, 1),
2649 build_int_cst (itype, 0));
2650 append_pattern_def_seq (stmt_vinfo, def_stmt);
2652 else
2654 tree utype
2655 = build_nonstandard_integer_type (prec, 1);
2656 tree vecutype = get_vectype_for_scalar_type (utype);
2657 tree shift
2658 = build_int_cst (utype, GET_MODE_BITSIZE (TYPE_MODE (itype))
2659 - tree_log2 (oprnd1));
2660 tree var = vect_recog_temp_ssa_var (utype, NULL);
2662 def_stmt = gimple_build_assign (var, COND_EXPR, cond,
2663 build_int_cst (utype, -1),
2664 build_int_cst (utype, 0));
2665 def_stmt_vinfo = new_stmt_vec_info (def_stmt, vinfo);
2666 set_vinfo_for_stmt (def_stmt, def_stmt_vinfo);
2667 STMT_VINFO_VECTYPE (def_stmt_vinfo) = vecutype;
2668 append_pattern_def_seq (stmt_vinfo, def_stmt);
2669 var = vect_recog_temp_ssa_var (utype, NULL);
2670 def_stmt = gimple_build_assign (var, RSHIFT_EXPR,
2671 gimple_assign_lhs (def_stmt),
2672 shift);
2673 def_stmt_vinfo = new_stmt_vec_info (def_stmt, vinfo);
2674 set_vinfo_for_stmt (def_stmt, def_stmt_vinfo);
2675 STMT_VINFO_VECTYPE (def_stmt_vinfo) = vecutype;
2676 append_pattern_def_seq (stmt_vinfo, def_stmt);
2677 signmask = vect_recog_temp_ssa_var (itype, NULL);
2678 def_stmt
2679 = gimple_build_assign (signmask, NOP_EXPR, var);
2680 append_pattern_def_seq (stmt_vinfo, def_stmt);
2682 def_stmt
2683 = gimple_build_assign (vect_recog_temp_ssa_var (itype, NULL),
2684 PLUS_EXPR, oprnd0, signmask);
2685 append_pattern_def_seq (stmt_vinfo, def_stmt);
2686 def_stmt
2687 = gimple_build_assign (vect_recog_temp_ssa_var (itype, NULL),
2688 BIT_AND_EXPR, gimple_assign_lhs (def_stmt),
2689 fold_build2 (MINUS_EXPR, itype, oprnd1,
2690 build_int_cst (itype, 1)));
2691 append_pattern_def_seq (stmt_vinfo, def_stmt);
2693 pattern_stmt
2694 = gimple_build_assign (vect_recog_temp_ssa_var (itype, NULL),
2695 MINUS_EXPR, gimple_assign_lhs (def_stmt),
2696 signmask);
2699 if (dump_enabled_p ())
2700 dump_gimple_stmt_loc (MSG_NOTE, vect_location, TDF_SLIM, pattern_stmt,
2703 stmts->safe_push (last_stmt);
2705 *type_in = vectype;
2706 *type_out = vectype;
2707 return pattern_stmt;
2710 if (prec > HOST_BITS_PER_WIDE_INT
2711 || integer_zerop (oprnd1))
2712 return NULL;
2714 if (!can_mult_highpart_p (TYPE_MODE (vectype), TYPE_UNSIGNED (itype)))
2715 return NULL;
2717 STMT_VINFO_PATTERN_DEF_SEQ (stmt_vinfo) = NULL;
2719 if (TYPE_UNSIGNED (itype))
2721 unsigned HOST_WIDE_INT mh, ml;
2722 int pre_shift, post_shift;
2723 unsigned HOST_WIDE_INT d = (TREE_INT_CST_LOW (oprnd1)
2724 & GET_MODE_MASK (TYPE_MODE (itype)));
2725 tree t1, t2, t3, t4;
2727 if (d >= (HOST_WIDE_INT_1U << (prec - 1)))
2728 /* FIXME: Can transform this into oprnd0 >= oprnd1 ? 1 : 0. */
2729 return NULL;
2731 /* Find a suitable multiplier and right shift count
2732 instead of multiplying with D. */
2733 mh = choose_multiplier (d, prec, prec, &ml, &post_shift, &dummy_int);
2735 /* If the suggested multiplier is more than SIZE bits, we can do better
2736 for even divisors, using an initial right shift. */
2737 if (mh != 0 && (d & 1) == 0)
2739 pre_shift = ctz_or_zero (d);
2740 mh = choose_multiplier (d >> pre_shift, prec, prec - pre_shift,
2741 &ml, &post_shift, &dummy_int);
2742 gcc_assert (!mh);
2744 else
2745 pre_shift = 0;
2747 if (mh != 0)
2749 if (post_shift - 1 >= prec)
2750 return NULL;
2752 /* t1 = oprnd0 h* ml;
2753 t2 = oprnd0 - t1;
2754 t3 = t2 >> 1;
2755 t4 = t1 + t3;
2756 q = t4 >> (post_shift - 1); */
2757 t1 = vect_recog_temp_ssa_var (itype, NULL);
2758 def_stmt = gimple_build_assign (t1, MULT_HIGHPART_EXPR, oprnd0,
2759 build_int_cst (itype, ml));
2760 append_pattern_def_seq (stmt_vinfo, def_stmt);
2762 t2 = vect_recog_temp_ssa_var (itype, NULL);
2763 def_stmt
2764 = gimple_build_assign (t2, MINUS_EXPR, oprnd0, t1);
2765 append_pattern_def_seq (stmt_vinfo, def_stmt);
2767 t3 = vect_recog_temp_ssa_var (itype, NULL);
2768 def_stmt
2769 = gimple_build_assign (t3, RSHIFT_EXPR, t2, integer_one_node);
2770 append_pattern_def_seq (stmt_vinfo, def_stmt);
2772 t4 = vect_recog_temp_ssa_var (itype, NULL);
2773 def_stmt
2774 = gimple_build_assign (t4, PLUS_EXPR, t1, t3);
2776 if (post_shift != 1)
2778 append_pattern_def_seq (stmt_vinfo, def_stmt);
2780 q = vect_recog_temp_ssa_var (itype, NULL);
2781 pattern_stmt
2782 = gimple_build_assign (q, RSHIFT_EXPR, t4,
2783 build_int_cst (itype, post_shift - 1));
2785 else
2787 q = t4;
2788 pattern_stmt = def_stmt;
2791 else
2793 if (pre_shift >= prec || post_shift >= prec)
2794 return NULL;
2796 /* t1 = oprnd0 >> pre_shift;
2797 t2 = t1 h* ml;
2798 q = t2 >> post_shift; */
2799 if (pre_shift)
2801 t1 = vect_recog_temp_ssa_var (itype, NULL);
2802 def_stmt
2803 = gimple_build_assign (t1, RSHIFT_EXPR, oprnd0,
2804 build_int_cst (NULL, pre_shift));
2805 append_pattern_def_seq (stmt_vinfo, def_stmt);
2807 else
2808 t1 = oprnd0;
2810 t2 = vect_recog_temp_ssa_var (itype, NULL);
2811 def_stmt = gimple_build_assign (t2, MULT_HIGHPART_EXPR, t1,
2812 build_int_cst (itype, ml));
2814 if (post_shift)
2816 append_pattern_def_seq (stmt_vinfo, def_stmt);
2818 q = vect_recog_temp_ssa_var (itype, NULL);
2819 def_stmt
2820 = gimple_build_assign (q, RSHIFT_EXPR, t2,
2821 build_int_cst (itype, post_shift));
2823 else
2824 q = t2;
2826 pattern_stmt = def_stmt;
2829 else
2831 unsigned HOST_WIDE_INT ml;
2832 int post_shift;
2833 HOST_WIDE_INT d = TREE_INT_CST_LOW (oprnd1);
2834 unsigned HOST_WIDE_INT abs_d;
2835 bool add = false;
2836 tree t1, t2, t3, t4;
2838 /* Give up for -1. */
2839 if (d == -1)
2840 return NULL;
2842 /* Since d might be INT_MIN, we have to cast to
2843 unsigned HOST_WIDE_INT before negating to avoid
2844 undefined signed overflow. */
2845 abs_d = (d >= 0
2846 ? (unsigned HOST_WIDE_INT) d
2847 : - (unsigned HOST_WIDE_INT) d);
2849 /* n rem d = n rem -d */
2850 if (rhs_code == TRUNC_MOD_EXPR && d < 0)
2852 d = abs_d;
2853 oprnd1 = build_int_cst (itype, abs_d);
2855 else if (HOST_BITS_PER_WIDE_INT >= prec
2856 && abs_d == HOST_WIDE_INT_1U << (prec - 1))
2857 /* This case is not handled correctly below. */
2858 return NULL;
2860 choose_multiplier (abs_d, prec, prec - 1, &ml, &post_shift, &dummy_int);
2861 if (ml >= HOST_WIDE_INT_1U << (prec - 1))
2863 add = true;
2864 ml |= HOST_WIDE_INT_M1U << (prec - 1);
2866 if (post_shift >= prec)
2867 return NULL;
2869 /* t1 = oprnd0 h* ml; */
2870 t1 = vect_recog_temp_ssa_var (itype, NULL);
2871 def_stmt = gimple_build_assign (t1, MULT_HIGHPART_EXPR, oprnd0,
2872 build_int_cst (itype, ml));
2874 if (add)
2876 /* t2 = t1 + oprnd0; */
2877 append_pattern_def_seq (stmt_vinfo, def_stmt);
2878 t2 = vect_recog_temp_ssa_var (itype, NULL);
2879 def_stmt = gimple_build_assign (t2, PLUS_EXPR, t1, oprnd0);
2881 else
2882 t2 = t1;
2884 if (post_shift)
2886 /* t3 = t2 >> post_shift; */
2887 append_pattern_def_seq (stmt_vinfo, def_stmt);
2888 t3 = vect_recog_temp_ssa_var (itype, NULL);
2889 def_stmt = gimple_build_assign (t3, RSHIFT_EXPR, t2,
2890 build_int_cst (itype, post_shift));
2892 else
2893 t3 = t2;
2895 wide_int oprnd0_min, oprnd0_max;
2896 int msb = 1;
2897 if (get_range_info (oprnd0, &oprnd0_min, &oprnd0_max) == VR_RANGE)
2899 if (!wi::neg_p (oprnd0_min, TYPE_SIGN (itype)))
2900 msb = 0;
2901 else if (wi::neg_p (oprnd0_max, TYPE_SIGN (itype)))
2902 msb = -1;
2905 if (msb == 0 && d >= 0)
2907 /* q = t3; */
2908 q = t3;
2909 pattern_stmt = def_stmt;
2911 else
2913 /* t4 = oprnd0 >> (prec - 1);
2914 or if we know from VRP that oprnd0 >= 0
2915 t4 = 0;
2916 or if we know from VRP that oprnd0 < 0
2917 t4 = -1; */
2918 append_pattern_def_seq (stmt_vinfo, def_stmt);
2919 t4 = vect_recog_temp_ssa_var (itype, NULL);
2920 if (msb != 1)
2921 def_stmt = gimple_build_assign (t4, INTEGER_CST,
2922 build_int_cst (itype, msb));
2923 else
2924 def_stmt = gimple_build_assign (t4, RSHIFT_EXPR, oprnd0,
2925 build_int_cst (itype, prec - 1));
2926 append_pattern_def_seq (stmt_vinfo, def_stmt);
2928 /* q = t3 - t4; or q = t4 - t3; */
2929 q = vect_recog_temp_ssa_var (itype, NULL);
2930 pattern_stmt = gimple_build_assign (q, MINUS_EXPR, d < 0 ? t4 : t3,
2931 d < 0 ? t3 : t4);
2935 if (rhs_code == TRUNC_MOD_EXPR)
2937 tree r, t1;
2939 /* We divided. Now finish by:
2940 t1 = q * oprnd1;
2941 r = oprnd0 - t1; */
2942 append_pattern_def_seq (stmt_vinfo, pattern_stmt);
2944 t1 = vect_recog_temp_ssa_var (itype, NULL);
2945 def_stmt = gimple_build_assign (t1, MULT_EXPR, q, oprnd1);
2946 append_pattern_def_seq (stmt_vinfo, def_stmt);
2948 r = vect_recog_temp_ssa_var (itype, NULL);
2949 pattern_stmt = gimple_build_assign (r, MINUS_EXPR, oprnd0, t1);
2952 /* Pattern detected. */
2953 if (dump_enabled_p ())
2955 dump_printf_loc (MSG_NOTE, vect_location,
2956 "vect_recog_divmod_pattern: detected: ");
2957 dump_gimple_stmt (MSG_NOTE, TDF_SLIM, pattern_stmt, 0);
2960 stmts->safe_push (last_stmt);
2962 *type_in = vectype;
2963 *type_out = vectype;
2964 return pattern_stmt;
2967 /* Function vect_recog_mixed_size_cond_pattern
2969 Try to find the following pattern:
2971 type x_t, y_t;
2972 TYPE a_T, b_T, c_T;
2973 loop:
2974 S1 a_T = x_t CMP y_t ? b_T : c_T;
2976 where type 'TYPE' is an integral type which has different size
2977 from 'type'. b_T and c_T are either constants (and if 'TYPE' is wider
2978 than 'type', the constants need to fit into an integer type
2979 with the same width as 'type') or results of conversion from 'type'.
2981 Input:
2983 * LAST_STMT: A stmt from which the pattern search begins.
2985 Output:
2987 * TYPE_IN: The type of the input arguments to the pattern.
2989 * TYPE_OUT: The type of the output of this pattern.
2991 * Return value: A new stmt that will be used to replace the pattern.
2992 Additionally a def_stmt is added.
2994 a_it = x_t CMP y_t ? b_it : c_it;
2995 a_T = (TYPE) a_it; */
2997 static gimple *
2998 vect_recog_mixed_size_cond_pattern (vec<gimple *> *stmts, tree *type_in,
2999 tree *type_out)
3001 gimple *last_stmt = (*stmts)[0];
3002 tree cond_expr, then_clause, else_clause;
3003 stmt_vec_info stmt_vinfo = vinfo_for_stmt (last_stmt), def_stmt_info;
3004 tree type, vectype, comp_vectype, itype = NULL_TREE, vecitype;
3005 gimple *pattern_stmt, *def_stmt;
3006 vec_info *vinfo = stmt_vinfo->vinfo;
3007 tree orig_type0 = NULL_TREE, orig_type1 = NULL_TREE;
3008 gimple *def_stmt0 = NULL, *def_stmt1 = NULL;
3009 bool promotion;
3010 tree comp_scalar_type;
3012 if (!is_gimple_assign (last_stmt)
3013 || gimple_assign_rhs_code (last_stmt) != COND_EXPR
3014 || STMT_VINFO_DEF_TYPE (stmt_vinfo) != vect_internal_def)
3015 return NULL;
3017 cond_expr = gimple_assign_rhs1 (last_stmt);
3018 then_clause = gimple_assign_rhs2 (last_stmt);
3019 else_clause = gimple_assign_rhs3 (last_stmt);
3021 if (!COMPARISON_CLASS_P (cond_expr))
3022 return NULL;
3024 comp_scalar_type = TREE_TYPE (TREE_OPERAND (cond_expr, 0));
3025 comp_vectype = get_vectype_for_scalar_type (comp_scalar_type);
3026 if (comp_vectype == NULL_TREE)
3027 return NULL;
3029 type = gimple_expr_type (last_stmt);
3030 if (types_compatible_p (type, comp_scalar_type)
3031 || ((TREE_CODE (then_clause) != INTEGER_CST
3032 || TREE_CODE (else_clause) != INTEGER_CST)
3033 && !INTEGRAL_TYPE_P (comp_scalar_type))
3034 || !INTEGRAL_TYPE_P (type))
3035 return NULL;
3037 if ((TREE_CODE (then_clause) != INTEGER_CST
3038 && !type_conversion_p (then_clause, last_stmt, false, &orig_type0,
3039 &def_stmt0, &promotion))
3040 || (TREE_CODE (else_clause) != INTEGER_CST
3041 && !type_conversion_p (else_clause, last_stmt, false, &orig_type1,
3042 &def_stmt1, &promotion)))
3043 return NULL;
3045 if (orig_type0 && orig_type1
3046 && !types_compatible_p (orig_type0, orig_type1))
3047 return NULL;
3049 if (orig_type0)
3051 if (!types_compatible_p (orig_type0, comp_scalar_type))
3052 return NULL;
3053 then_clause = gimple_assign_rhs1 (def_stmt0);
3054 itype = orig_type0;
3057 if (orig_type1)
3059 if (!types_compatible_p (orig_type1, comp_scalar_type))
3060 return NULL;
3061 else_clause = gimple_assign_rhs1 (def_stmt1);
3062 itype = orig_type1;
3066 HOST_WIDE_INT cmp_mode_size
3067 = GET_MODE_UNIT_BITSIZE (TYPE_MODE (comp_vectype));
3069 if (GET_MODE_BITSIZE (TYPE_MODE (type)) == cmp_mode_size)
3070 return NULL;
3072 vectype = get_vectype_for_scalar_type (type);
3073 if (vectype == NULL_TREE)
3074 return NULL;
3076 if (expand_vec_cond_expr_p (vectype, comp_vectype))
3077 return NULL;
3079 if (itype == NULL_TREE)
3080 itype = build_nonstandard_integer_type (cmp_mode_size,
3081 TYPE_UNSIGNED (type));
3083 if (itype == NULL_TREE
3084 || GET_MODE_BITSIZE (TYPE_MODE (itype)) != cmp_mode_size)
3085 return NULL;
3087 vecitype = get_vectype_for_scalar_type (itype);
3088 if (vecitype == NULL_TREE)
3089 return NULL;
3091 if (!expand_vec_cond_expr_p (vecitype, comp_vectype))
3092 return NULL;
3094 if (GET_MODE_BITSIZE (TYPE_MODE (type)) > cmp_mode_size)
3096 if ((TREE_CODE (then_clause) == INTEGER_CST
3097 && !int_fits_type_p (then_clause, itype))
3098 || (TREE_CODE (else_clause) == INTEGER_CST
3099 && !int_fits_type_p (else_clause, itype)))
3100 return NULL;
3103 def_stmt = gimple_build_assign (vect_recog_temp_ssa_var (itype, NULL),
3104 COND_EXPR, unshare_expr (cond_expr),
3105 fold_convert (itype, then_clause),
3106 fold_convert (itype, else_clause));
3107 pattern_stmt = gimple_build_assign (vect_recog_temp_ssa_var (type, NULL),
3108 NOP_EXPR, gimple_assign_lhs (def_stmt));
3110 new_pattern_def_seq (stmt_vinfo, def_stmt);
3111 def_stmt_info = new_stmt_vec_info (def_stmt, vinfo);
3112 set_vinfo_for_stmt (def_stmt, def_stmt_info);
3113 STMT_VINFO_VECTYPE (def_stmt_info) = vecitype;
3114 *type_in = vecitype;
3115 *type_out = vectype;
3117 if (dump_enabled_p ())
3118 dump_printf_loc (MSG_NOTE, vect_location,
3119 "vect_recog_mixed_size_cond_pattern: detected:\n");
3121 return pattern_stmt;
3125 /* Helper function of vect_recog_bool_pattern. Called recursively, return
3126 true if bool VAR can and should be optimized that way. Assume it shouldn't
3127 in case it's a result of a comparison which can be directly vectorized into
3128 a vector comparison. Fills in STMTS with all stmts visited during the
3129 walk. */
3131 static bool
3132 check_bool_pattern (tree var, vec_info *vinfo, hash_set<gimple *> &stmts)
3134 gimple *def_stmt;
3135 enum vect_def_type dt;
3136 tree rhs1;
3137 enum tree_code rhs_code;
3139 if (!vect_is_simple_use (var, vinfo, &def_stmt, &dt))
3140 return false;
3142 if (dt != vect_internal_def)
3143 return false;
3145 if (!is_gimple_assign (def_stmt))
3146 return false;
3148 if (stmts.contains (def_stmt))
3149 return true;
3151 rhs1 = gimple_assign_rhs1 (def_stmt);
3152 rhs_code = gimple_assign_rhs_code (def_stmt);
3153 switch (rhs_code)
3155 case SSA_NAME:
3156 if (! check_bool_pattern (rhs1, vinfo, stmts))
3157 return false;
3158 break;
3160 CASE_CONVERT:
3161 if ((TYPE_PRECISION (TREE_TYPE (rhs1)) != 1
3162 || !TYPE_UNSIGNED (TREE_TYPE (rhs1)))
3163 && TREE_CODE (TREE_TYPE (rhs1)) != BOOLEAN_TYPE)
3164 return false;
3165 if (! check_bool_pattern (rhs1, vinfo, stmts))
3166 return false;
3167 break;
3169 case BIT_NOT_EXPR:
3170 if (! check_bool_pattern (rhs1, vinfo, stmts))
3171 return false;
3172 break;
3174 case BIT_AND_EXPR:
3175 case BIT_IOR_EXPR:
3176 case BIT_XOR_EXPR:
3177 if (! check_bool_pattern (rhs1, vinfo, stmts)
3178 || ! check_bool_pattern (gimple_assign_rhs2 (def_stmt), vinfo, stmts))
3179 return false;
3180 break;
3182 default:
3183 if (TREE_CODE_CLASS (rhs_code) == tcc_comparison)
3185 tree vecitype, comp_vectype;
3187 /* If the comparison can throw, then is_gimple_condexpr will be
3188 false and we can't make a COND_EXPR/VEC_COND_EXPR out of it. */
3189 if (stmt_could_throw_p (def_stmt))
3190 return false;
3192 comp_vectype = get_vectype_for_scalar_type (TREE_TYPE (rhs1));
3193 if (comp_vectype == NULL_TREE)
3194 return false;
3196 tree mask_type = get_mask_type_for_scalar_type (TREE_TYPE (rhs1));
3197 if (mask_type
3198 && expand_vec_cmp_expr_p (comp_vectype, mask_type))
3199 return false;
3201 if (TREE_CODE (TREE_TYPE (rhs1)) != INTEGER_TYPE)
3203 machine_mode mode = TYPE_MODE (TREE_TYPE (rhs1));
3204 tree itype
3205 = build_nonstandard_integer_type (GET_MODE_BITSIZE (mode), 1);
3206 vecitype = get_vectype_for_scalar_type (itype);
3207 if (vecitype == NULL_TREE)
3208 return false;
3210 else
3211 vecitype = comp_vectype;
3212 if (! expand_vec_cond_expr_p (vecitype, comp_vectype))
3213 return false;
3215 else
3216 return false;
3217 break;
3220 bool res = stmts.add (def_stmt);
3221 /* We can't end up recursing when just visiting SSA defs but not PHIs. */
3222 gcc_assert (!res);
3224 return true;
3228 /* Helper function of adjust_bool_pattern. Add a cast to TYPE to a previous
3229 stmt (SSA_NAME_DEF_STMT of VAR) adding a cast to STMT_INFOs
3230 pattern sequence. */
3232 static tree
3233 adjust_bool_pattern_cast (tree type, tree var, stmt_vec_info stmt_info)
3235 gimple *cast_stmt = gimple_build_assign (vect_recog_temp_ssa_var (type, NULL),
3236 NOP_EXPR, var);
3237 stmt_vec_info patt_vinfo = new_stmt_vec_info (cast_stmt, stmt_info->vinfo);
3238 set_vinfo_for_stmt (cast_stmt, patt_vinfo);
3239 STMT_VINFO_VECTYPE (patt_vinfo) = get_vectype_for_scalar_type (type);
3240 append_pattern_def_seq (stmt_info, cast_stmt);
3241 return gimple_assign_lhs (cast_stmt);
3244 /* Helper function of vect_recog_bool_pattern. Do the actual transformations.
3245 VAR is an SSA_NAME that should be transformed from bool to a wider integer
3246 type, OUT_TYPE is the desired final integer type of the whole pattern.
3247 STMT_INFO is the info of the pattern root and is where pattern stmts should
3248 be associated with. DEFS is a map of pattern defs. */
3250 static void
3251 adjust_bool_pattern (tree var, tree out_type,
3252 stmt_vec_info stmt_info, hash_map <tree, tree> &defs)
3254 gimple *stmt = SSA_NAME_DEF_STMT (var);
3255 enum tree_code rhs_code, def_rhs_code;
3256 tree itype, cond_expr, rhs1, rhs2, irhs1, irhs2;
3257 location_t loc;
3258 gimple *pattern_stmt, *def_stmt;
3259 tree trueval = NULL_TREE;
3261 rhs1 = gimple_assign_rhs1 (stmt);
3262 rhs2 = gimple_assign_rhs2 (stmt);
3263 rhs_code = gimple_assign_rhs_code (stmt);
3264 loc = gimple_location (stmt);
3265 switch (rhs_code)
3267 case SSA_NAME:
3268 CASE_CONVERT:
3269 irhs1 = *defs.get (rhs1);
3270 itype = TREE_TYPE (irhs1);
3271 pattern_stmt
3272 = gimple_build_assign (vect_recog_temp_ssa_var (itype, NULL),
3273 SSA_NAME, irhs1);
3274 break;
3276 case BIT_NOT_EXPR:
3277 irhs1 = *defs.get (rhs1);
3278 itype = TREE_TYPE (irhs1);
3279 pattern_stmt
3280 = gimple_build_assign (vect_recog_temp_ssa_var (itype, NULL),
3281 BIT_XOR_EXPR, irhs1, build_int_cst (itype, 1));
3282 break;
3284 case BIT_AND_EXPR:
3285 /* Try to optimize x = y & (a < b ? 1 : 0); into
3286 x = (a < b ? y : 0);
3288 E.g. for:
3289 bool a_b, b_b, c_b;
3290 TYPE d_T;
3292 S1 a_b = x1 CMP1 y1;
3293 S2 b_b = x2 CMP2 y2;
3294 S3 c_b = a_b & b_b;
3295 S4 d_T = (TYPE) c_b;
3297 we would normally emit:
3299 S1' a_T = x1 CMP1 y1 ? 1 : 0;
3300 S2' b_T = x2 CMP2 y2 ? 1 : 0;
3301 S3' c_T = a_T & b_T;
3302 S4' d_T = c_T;
3304 but we can save one stmt by using the
3305 result of one of the COND_EXPRs in the other COND_EXPR and leave
3306 BIT_AND_EXPR stmt out:
3308 S1' a_T = x1 CMP1 y1 ? 1 : 0;
3309 S3' c_T = x2 CMP2 y2 ? a_T : 0;
3310 S4' f_T = c_T;
3312 At least when VEC_COND_EXPR is implemented using masks
3313 cond ? 1 : 0 is as expensive as cond ? var : 0, in both cases it
3314 computes the comparison masks and ands it, in one case with
3315 all ones vector, in the other case with a vector register.
3316 Don't do this for BIT_IOR_EXPR, because cond ? 1 : var; is
3317 often more expensive. */
3318 def_stmt = SSA_NAME_DEF_STMT (rhs2);
3319 def_rhs_code = gimple_assign_rhs_code (def_stmt);
3320 if (TREE_CODE_CLASS (def_rhs_code) == tcc_comparison)
3322 irhs1 = *defs.get (rhs1);
3323 tree def_rhs1 = gimple_assign_rhs1 (def_stmt);
3324 if (TYPE_PRECISION (TREE_TYPE (irhs1))
3325 == GET_MODE_BITSIZE (TYPE_MODE (TREE_TYPE (def_rhs1))))
3327 rhs_code = def_rhs_code;
3328 rhs1 = def_rhs1;
3329 rhs2 = gimple_assign_rhs2 (def_stmt);
3330 trueval = irhs1;
3331 goto do_compare;
3333 else
3334 irhs2 = *defs.get (rhs2);
3335 goto and_ior_xor;
3337 def_stmt = SSA_NAME_DEF_STMT (rhs1);
3338 def_rhs_code = gimple_assign_rhs_code (def_stmt);
3339 if (TREE_CODE_CLASS (def_rhs_code) == tcc_comparison)
3341 irhs2 = *defs.get (rhs2);
3342 tree def_rhs1 = gimple_assign_rhs1 (def_stmt);
3343 if (TYPE_PRECISION (TREE_TYPE (irhs2))
3344 == GET_MODE_BITSIZE (TYPE_MODE (TREE_TYPE (def_rhs1))))
3346 rhs_code = def_rhs_code;
3347 rhs1 = def_rhs1;
3348 rhs2 = gimple_assign_rhs2 (def_stmt);
3349 trueval = irhs2;
3350 goto do_compare;
3352 else
3353 irhs1 = *defs.get (rhs1);
3354 goto and_ior_xor;
3356 /* FALLTHRU */
3357 case BIT_IOR_EXPR:
3358 case BIT_XOR_EXPR:
3359 irhs1 = *defs.get (rhs1);
3360 irhs2 = *defs.get (rhs2);
3361 and_ior_xor:
3362 if (TYPE_PRECISION (TREE_TYPE (irhs1))
3363 != TYPE_PRECISION (TREE_TYPE (irhs2)))
3365 int prec1 = TYPE_PRECISION (TREE_TYPE (irhs1));
3366 int prec2 = TYPE_PRECISION (TREE_TYPE (irhs2));
3367 int out_prec = TYPE_PRECISION (out_type);
3368 if (absu_hwi (out_prec - prec1) < absu_hwi (out_prec - prec2))
3369 irhs2 = adjust_bool_pattern_cast (TREE_TYPE (irhs1), irhs2,
3370 stmt_info);
3371 else if (absu_hwi (out_prec - prec1) > absu_hwi (out_prec - prec2))
3372 irhs1 = adjust_bool_pattern_cast (TREE_TYPE (irhs2), irhs1,
3373 stmt_info);
3374 else
3376 irhs1 = adjust_bool_pattern_cast (out_type, irhs1, stmt_info);
3377 irhs2 = adjust_bool_pattern_cast (out_type, irhs2, stmt_info);
3380 itype = TREE_TYPE (irhs1);
3381 pattern_stmt
3382 = gimple_build_assign (vect_recog_temp_ssa_var (itype, NULL),
3383 rhs_code, irhs1, irhs2);
3384 break;
3386 default:
3387 do_compare:
3388 gcc_assert (TREE_CODE_CLASS (rhs_code) == tcc_comparison);
3389 if (TREE_CODE (TREE_TYPE (rhs1)) != INTEGER_TYPE
3390 || !TYPE_UNSIGNED (TREE_TYPE (rhs1))
3391 || (TYPE_PRECISION (TREE_TYPE (rhs1))
3392 != GET_MODE_BITSIZE (TYPE_MODE (TREE_TYPE (rhs1)))))
3394 machine_mode mode = TYPE_MODE (TREE_TYPE (rhs1));
3395 itype
3396 = build_nonstandard_integer_type (GET_MODE_BITSIZE (mode), 1);
3398 else
3399 itype = TREE_TYPE (rhs1);
3400 cond_expr = build2_loc (loc, rhs_code, itype, rhs1, rhs2);
3401 if (trueval == NULL_TREE)
3402 trueval = build_int_cst (itype, 1);
3403 else
3404 gcc_checking_assert (useless_type_conversion_p (itype,
3405 TREE_TYPE (trueval)));
3406 pattern_stmt
3407 = gimple_build_assign (vect_recog_temp_ssa_var (itype, NULL),
3408 COND_EXPR, cond_expr, trueval,
3409 build_int_cst (itype, 0));
3410 break;
3413 gimple_set_location (pattern_stmt, loc);
3414 /* ??? Why does vect_mark_pattern_stmts set the vector type on all
3415 pattern def seq stmts instead of just letting auto-detection do
3416 its work? */
3417 stmt_vec_info patt_vinfo = new_stmt_vec_info (pattern_stmt, stmt_info->vinfo);
3418 set_vinfo_for_stmt (pattern_stmt, patt_vinfo);
3419 STMT_VINFO_VECTYPE (patt_vinfo) = get_vectype_for_scalar_type (itype);
3420 append_pattern_def_seq (stmt_info, pattern_stmt);
3421 defs.put (var, gimple_assign_lhs (pattern_stmt));
3424 /* Comparison function to qsort a vector of gimple stmts after UID. */
3426 static int
3427 sort_after_uid (const void *p1, const void *p2)
3429 const gimple *stmt1 = *(const gimple * const *)p1;
3430 const gimple *stmt2 = *(const gimple * const *)p2;
3431 return gimple_uid (stmt1) - gimple_uid (stmt2);
3434 /* Create pattern stmts for all stmts participating in the bool pattern
3435 specified by BOOL_STMT_SET and its root STMT with the desired type
3436 OUT_TYPE. Return the def of the pattern root. */
3438 static tree
3439 adjust_bool_stmts (hash_set <gimple *> &bool_stmt_set,
3440 tree out_type, gimple *stmt)
3442 /* Gather original stmts in the bool pattern in their order of appearance
3443 in the IL. */
3444 auto_vec<gimple *> bool_stmts (bool_stmt_set.elements ());
3445 for (hash_set <gimple *>::iterator i = bool_stmt_set.begin ();
3446 i != bool_stmt_set.end (); ++i)
3447 bool_stmts.quick_push (*i);
3448 bool_stmts.qsort (sort_after_uid);
3450 /* Now process them in that order, producing pattern stmts. */
3451 hash_map <tree, tree> defs;
3452 for (unsigned i = 0; i < bool_stmts.length (); ++i)
3453 adjust_bool_pattern (gimple_assign_lhs (bool_stmts[i]),
3454 out_type, vinfo_for_stmt (stmt), defs);
3456 /* Pop the last pattern seq stmt and install it as pattern root for STMT. */
3457 gimple *pattern_stmt
3458 = gimple_seq_last_stmt (STMT_VINFO_PATTERN_DEF_SEQ (vinfo_for_stmt (stmt)));
3459 return gimple_assign_lhs (pattern_stmt);
3462 /* Helper for search_type_for_mask. */
3464 static tree
3465 search_type_for_mask_1 (tree var, vec_info *vinfo,
3466 hash_map<gimple *, tree> &cache)
3468 gimple *def_stmt;
3469 enum vect_def_type dt;
3470 tree rhs1;
3471 enum tree_code rhs_code;
3472 tree res = NULL_TREE, res2;
3474 if (TREE_CODE (var) != SSA_NAME)
3475 return NULL_TREE;
3477 if ((TYPE_PRECISION (TREE_TYPE (var)) != 1
3478 || !TYPE_UNSIGNED (TREE_TYPE (var)))
3479 && TREE_CODE (TREE_TYPE (var)) != BOOLEAN_TYPE)
3480 return NULL_TREE;
3482 if (!vect_is_simple_use (var, vinfo, &def_stmt, &dt))
3483 return NULL_TREE;
3485 if (dt != vect_internal_def)
3486 return NULL_TREE;
3488 if (!is_gimple_assign (def_stmt))
3489 return NULL_TREE;
3491 tree *c = cache.get (def_stmt);
3492 if (c)
3493 return *c;
3495 rhs_code = gimple_assign_rhs_code (def_stmt);
3496 rhs1 = gimple_assign_rhs1 (def_stmt);
3498 switch (rhs_code)
3500 case SSA_NAME:
3501 case BIT_NOT_EXPR:
3502 CASE_CONVERT:
3503 res = search_type_for_mask_1 (rhs1, vinfo, cache);
3504 break;
3506 case BIT_AND_EXPR:
3507 case BIT_IOR_EXPR:
3508 case BIT_XOR_EXPR:
3509 res = search_type_for_mask_1 (rhs1, vinfo, cache);
3510 res2 = search_type_for_mask_1 (gimple_assign_rhs2 (def_stmt), vinfo,
3511 cache);
3512 if (!res || (res2 && TYPE_PRECISION (res) > TYPE_PRECISION (res2)))
3513 res = res2;
3514 break;
3516 default:
3517 if (TREE_CODE_CLASS (rhs_code) == tcc_comparison)
3519 tree comp_vectype, mask_type;
3521 if (TREE_CODE (TREE_TYPE (rhs1)) == BOOLEAN_TYPE)
3523 res = search_type_for_mask_1 (rhs1, vinfo, cache);
3524 res2 = search_type_for_mask_1 (gimple_assign_rhs2 (def_stmt),
3525 vinfo, cache);
3526 if (!res || (res2 && TYPE_PRECISION (res) > TYPE_PRECISION (res2)))
3527 res = res2;
3528 break;
3531 comp_vectype = get_vectype_for_scalar_type (TREE_TYPE (rhs1));
3532 if (comp_vectype == NULL_TREE)
3534 res = NULL_TREE;
3535 break;
3538 mask_type = get_mask_type_for_scalar_type (TREE_TYPE (rhs1));
3539 if (!mask_type
3540 || !expand_vec_cmp_expr_p (comp_vectype, mask_type))
3542 res = NULL_TREE;
3543 break;
3546 if (TREE_CODE (TREE_TYPE (rhs1)) != INTEGER_TYPE
3547 || !TYPE_UNSIGNED (TREE_TYPE (rhs1)))
3549 machine_mode mode = TYPE_MODE (TREE_TYPE (rhs1));
3550 res = build_nonstandard_integer_type (GET_MODE_BITSIZE (mode), 1);
3552 else
3553 res = TREE_TYPE (rhs1);
3557 cache.put (def_stmt, res);
3558 return res;
3561 /* Return the proper type for converting bool VAR into
3562 an integer value or NULL_TREE if no such type exists.
3563 The type is chosen so that converted value has the
3564 same number of elements as VAR's vector type. */
3566 static tree
3567 search_type_for_mask (tree var, vec_info *vinfo)
3569 hash_map<gimple *, tree> cache;
3570 return search_type_for_mask_1 (var, vinfo, cache);
3573 /* Function vect_recog_bool_pattern
3575 Try to find pattern like following:
3577 bool a_b, b_b, c_b, d_b, e_b;
3578 TYPE f_T;
3579 loop:
3580 S1 a_b = x1 CMP1 y1;
3581 S2 b_b = x2 CMP2 y2;
3582 S3 c_b = a_b & b_b;
3583 S4 d_b = x3 CMP3 y3;
3584 S5 e_b = c_b | d_b;
3585 S6 f_T = (TYPE) e_b;
3587 where type 'TYPE' is an integral type. Or a similar pattern
3588 ending in
3590 S6 f_Y = e_b ? r_Y : s_Y;
3592 as results from if-conversion of a complex condition.
3594 Input:
3596 * LAST_STMT: A stmt at the end from which the pattern
3597 search begins, i.e. cast of a bool to
3598 an integer type.
3600 Output:
3602 * TYPE_IN: The type of the input arguments to the pattern.
3604 * TYPE_OUT: The type of the output of this pattern.
3606 * Return value: A new stmt that will be used to replace the pattern.
3608 Assuming size of TYPE is the same as size of all comparisons
3609 (otherwise some casts would be added where needed), the above
3610 sequence we create related pattern stmts:
3611 S1' a_T = x1 CMP1 y1 ? 1 : 0;
3612 S3' c_T = x2 CMP2 y2 ? a_T : 0;
3613 S4' d_T = x3 CMP3 y3 ? 1 : 0;
3614 S5' e_T = c_T | d_T;
3615 S6' f_T = e_T;
3617 Instead of the above S3' we could emit:
3618 S2' b_T = x2 CMP2 y2 ? 1 : 0;
3619 S3' c_T = a_T | b_T;
3620 but the above is more efficient. */
3622 static gimple *
3623 vect_recog_bool_pattern (vec<gimple *> *stmts, tree *type_in,
3624 tree *type_out)
3626 gimple *last_stmt = stmts->pop ();
3627 enum tree_code rhs_code;
3628 tree var, lhs, rhs, vectype;
3629 stmt_vec_info stmt_vinfo = vinfo_for_stmt (last_stmt);
3630 stmt_vec_info new_stmt_info;
3631 vec_info *vinfo = stmt_vinfo->vinfo;
3632 gimple *pattern_stmt;
3634 if (!is_gimple_assign (last_stmt))
3635 return NULL;
3637 var = gimple_assign_rhs1 (last_stmt);
3638 lhs = gimple_assign_lhs (last_stmt);
3640 if ((TYPE_PRECISION (TREE_TYPE (var)) != 1
3641 || !TYPE_UNSIGNED (TREE_TYPE (var)))
3642 && TREE_CODE (TREE_TYPE (var)) != BOOLEAN_TYPE)
3643 return NULL;
3645 hash_set<gimple *> bool_stmts;
3647 rhs_code = gimple_assign_rhs_code (last_stmt);
3648 if (CONVERT_EXPR_CODE_P (rhs_code))
3650 if (TREE_CODE (TREE_TYPE (lhs)) != INTEGER_TYPE
3651 || TYPE_PRECISION (TREE_TYPE (lhs)) == 1)
3652 return NULL;
3653 vectype = get_vectype_for_scalar_type (TREE_TYPE (lhs));
3654 if (vectype == NULL_TREE)
3655 return NULL;
3657 if (check_bool_pattern (var, vinfo, bool_stmts))
3659 rhs = adjust_bool_stmts (bool_stmts, TREE_TYPE (lhs), last_stmt);
3660 lhs = vect_recog_temp_ssa_var (TREE_TYPE (lhs), NULL);
3661 if (useless_type_conversion_p (TREE_TYPE (lhs), TREE_TYPE (rhs)))
3662 pattern_stmt = gimple_build_assign (lhs, SSA_NAME, rhs);
3663 else
3664 pattern_stmt
3665 = gimple_build_assign (lhs, NOP_EXPR, rhs);
3667 else
3669 tree type = search_type_for_mask (var, vinfo);
3670 tree cst0, cst1, tmp;
3672 if (!type)
3673 return NULL;
3675 /* We may directly use cond with narrowed type to avoid
3676 multiple cond exprs with following result packing and
3677 perform single cond with packed mask instead. In case
3678 of widening we better make cond first and then extract
3679 results. */
3680 if (TYPE_MODE (type) == TYPE_MODE (TREE_TYPE (lhs)))
3681 type = TREE_TYPE (lhs);
3683 cst0 = build_int_cst (type, 0);
3684 cst1 = build_int_cst (type, 1);
3685 tmp = vect_recog_temp_ssa_var (type, NULL);
3686 pattern_stmt = gimple_build_assign (tmp, COND_EXPR, var, cst1, cst0);
3688 if (!useless_type_conversion_p (type, TREE_TYPE (lhs)))
3690 tree new_vectype = get_vectype_for_scalar_type (type);
3691 new_stmt_info = new_stmt_vec_info (pattern_stmt, vinfo);
3692 set_vinfo_for_stmt (pattern_stmt, new_stmt_info);
3693 STMT_VINFO_VECTYPE (new_stmt_info) = new_vectype;
3694 new_pattern_def_seq (stmt_vinfo, pattern_stmt);
3696 lhs = vect_recog_temp_ssa_var (TREE_TYPE (lhs), NULL);
3697 pattern_stmt = gimple_build_assign (lhs, CONVERT_EXPR, tmp);
3701 *type_out = vectype;
3702 *type_in = vectype;
3703 stmts->safe_push (last_stmt);
3704 if (dump_enabled_p ())
3705 dump_printf_loc (MSG_NOTE, vect_location,
3706 "vect_recog_bool_pattern: detected:\n");
3708 return pattern_stmt;
3710 else if (rhs_code == COND_EXPR
3711 && TREE_CODE (var) == SSA_NAME)
3713 vectype = get_vectype_for_scalar_type (TREE_TYPE (lhs));
3714 if (vectype == NULL_TREE)
3715 return NULL;
3717 /* Build a scalar type for the boolean result that when
3718 vectorized matches the vector type of the result in
3719 size and number of elements. */
3720 unsigned prec
3721 = wi::udiv_trunc (TYPE_SIZE (vectype),
3722 TYPE_VECTOR_SUBPARTS (vectype)).to_uhwi ();
3723 tree type
3724 = build_nonstandard_integer_type (prec,
3725 TYPE_UNSIGNED (TREE_TYPE (var)));
3726 if (get_vectype_for_scalar_type (type) == NULL_TREE)
3727 return NULL;
3729 if (!check_bool_pattern (var, vinfo, bool_stmts))
3730 return NULL;
3732 rhs = adjust_bool_stmts (bool_stmts, type, last_stmt);
3734 lhs = vect_recog_temp_ssa_var (TREE_TYPE (lhs), NULL);
3735 pattern_stmt
3736 = gimple_build_assign (lhs, COND_EXPR,
3737 build2 (NE_EXPR, boolean_type_node,
3738 rhs, build_int_cst (type, 0)),
3739 gimple_assign_rhs2 (last_stmt),
3740 gimple_assign_rhs3 (last_stmt));
3741 *type_out = vectype;
3742 *type_in = vectype;
3743 stmts->safe_push (last_stmt);
3744 if (dump_enabled_p ())
3745 dump_printf_loc (MSG_NOTE, vect_location,
3746 "vect_recog_bool_pattern: detected:\n");
3748 return pattern_stmt;
3750 else if (rhs_code == SSA_NAME
3751 && STMT_VINFO_DATA_REF (stmt_vinfo))
3753 stmt_vec_info pattern_stmt_info;
3754 vectype = STMT_VINFO_VECTYPE (stmt_vinfo);
3755 gcc_assert (vectype != NULL_TREE);
3756 if (!VECTOR_MODE_P (TYPE_MODE (vectype)))
3757 return NULL;
3759 if (check_bool_pattern (var, vinfo, bool_stmts))
3760 rhs = adjust_bool_stmts (bool_stmts, TREE_TYPE (vectype), last_stmt);
3761 else
3763 tree type = search_type_for_mask (var, vinfo);
3764 tree cst0, cst1, new_vectype;
3766 if (!type)
3767 return NULL;
3769 if (TYPE_MODE (type) == TYPE_MODE (TREE_TYPE (vectype)))
3770 type = TREE_TYPE (vectype);
3772 cst0 = build_int_cst (type, 0);
3773 cst1 = build_int_cst (type, 1);
3774 new_vectype = get_vectype_for_scalar_type (type);
3776 rhs = vect_recog_temp_ssa_var (type, NULL);
3777 pattern_stmt = gimple_build_assign (rhs, COND_EXPR, var, cst1, cst0);
3779 pattern_stmt_info = new_stmt_vec_info (pattern_stmt, vinfo);
3780 set_vinfo_for_stmt (pattern_stmt, pattern_stmt_info);
3781 STMT_VINFO_VECTYPE (pattern_stmt_info) = new_vectype;
3782 append_pattern_def_seq (stmt_vinfo, pattern_stmt);
3785 lhs = build1 (VIEW_CONVERT_EXPR, TREE_TYPE (vectype), lhs);
3786 if (!useless_type_conversion_p (TREE_TYPE (lhs), TREE_TYPE (rhs)))
3788 tree rhs2 = vect_recog_temp_ssa_var (TREE_TYPE (lhs), NULL);
3789 gimple *cast_stmt = gimple_build_assign (rhs2, NOP_EXPR, rhs);
3790 append_pattern_def_seq (stmt_vinfo, cast_stmt);
3791 rhs = rhs2;
3793 pattern_stmt = gimple_build_assign (lhs, SSA_NAME, rhs);
3794 pattern_stmt_info = new_stmt_vec_info (pattern_stmt, vinfo);
3795 set_vinfo_for_stmt (pattern_stmt, pattern_stmt_info);
3796 STMT_VINFO_DATA_REF (pattern_stmt_info)
3797 = STMT_VINFO_DATA_REF (stmt_vinfo);
3798 STMT_VINFO_DR_BASE_ADDRESS (pattern_stmt_info)
3799 = STMT_VINFO_DR_BASE_ADDRESS (stmt_vinfo);
3800 STMT_VINFO_DR_INIT (pattern_stmt_info) = STMT_VINFO_DR_INIT (stmt_vinfo);
3801 STMT_VINFO_DR_OFFSET (pattern_stmt_info)
3802 = STMT_VINFO_DR_OFFSET (stmt_vinfo);
3803 STMT_VINFO_DR_STEP (pattern_stmt_info) = STMT_VINFO_DR_STEP (stmt_vinfo);
3804 STMT_VINFO_DR_ALIGNED_TO (pattern_stmt_info)
3805 = STMT_VINFO_DR_ALIGNED_TO (stmt_vinfo);
3806 DR_STMT (STMT_VINFO_DATA_REF (stmt_vinfo)) = pattern_stmt;
3807 *type_out = vectype;
3808 *type_in = vectype;
3809 stmts->safe_push (last_stmt);
3810 if (dump_enabled_p ())
3811 dump_printf_loc (MSG_NOTE, vect_location,
3812 "vect_recog_bool_pattern: detected:\n");
3813 return pattern_stmt;
3815 else
3816 return NULL;
3820 /* A helper for vect_recog_mask_conversion_pattern. Build
3821 conversion of MASK to a type suitable for masking VECTYPE.
3822 Built statement gets required vectype and is appended to
3823 a pattern sequence of STMT_VINFO.
3825 Return converted mask. */
3827 static tree
3828 build_mask_conversion (tree mask, tree vectype, stmt_vec_info stmt_vinfo,
3829 vec_info *vinfo)
3831 gimple *stmt;
3832 tree masktype, tmp;
3833 stmt_vec_info new_stmt_info;
3835 masktype = build_same_sized_truth_vector_type (vectype);
3836 tmp = vect_recog_temp_ssa_var (TREE_TYPE (masktype), NULL);
3837 stmt = gimple_build_assign (tmp, CONVERT_EXPR, mask);
3838 new_stmt_info = new_stmt_vec_info (stmt, vinfo);
3839 set_vinfo_for_stmt (stmt, new_stmt_info);
3840 STMT_VINFO_VECTYPE (new_stmt_info) = masktype;
3841 append_pattern_def_seq (stmt_vinfo, stmt);
3843 return tmp;
3847 /* Function vect_recog_mask_conversion_pattern
3849 Try to find statements which require boolean type
3850 converison. Additional conversion statements are
3851 added to handle such cases. For example:
3853 bool m_1, m_2, m_3;
3854 int i_4, i_5;
3855 double d_6, d_7;
3856 char c_1, c_2, c_3;
3858 S1 m_1 = i_4 > i_5;
3859 S2 m_2 = d_6 < d_7;
3860 S3 m_3 = m_1 & m_2;
3861 S4 c_1 = m_3 ? c_2 : c_3;
3863 Will be transformed into:
3865 S1 m_1 = i_4 > i_5;
3866 S2 m_2 = d_6 < d_7;
3867 S3'' m_2' = (_Bool[bitsize=32])m_2
3868 S3' m_3' = m_1 & m_2';
3869 S4'' m_3'' = (_Bool[bitsize=8])m_3'
3870 S4' c_1' = m_3'' ? c_2 : c_3; */
3872 static gimple *
3873 vect_recog_mask_conversion_pattern (vec<gimple *> *stmts, tree *type_in,
3874 tree *type_out)
3876 gimple *last_stmt = stmts->pop ();
3877 enum tree_code rhs_code;
3878 tree lhs = NULL_TREE, rhs1, rhs2, tmp, rhs1_type, rhs2_type;
3879 tree vectype1, vectype2;
3880 stmt_vec_info stmt_vinfo = vinfo_for_stmt (last_stmt);
3881 stmt_vec_info pattern_stmt_info;
3882 vec_info *vinfo = stmt_vinfo->vinfo;
3883 gimple *pattern_stmt;
3885 /* Check for MASK_LOAD ans MASK_STORE calls requiring mask conversion. */
3886 if (is_gimple_call (last_stmt)
3887 && gimple_call_internal_p (last_stmt)
3888 && (gimple_call_internal_fn (last_stmt) == IFN_MASK_STORE
3889 || gimple_call_internal_fn (last_stmt) == IFN_MASK_LOAD))
3891 bool load = (gimple_call_internal_fn (last_stmt) == IFN_MASK_LOAD);
3893 if (load)
3895 lhs = gimple_call_lhs (last_stmt);
3896 vectype1 = get_vectype_for_scalar_type (TREE_TYPE (lhs));
3898 else
3900 rhs2 = gimple_call_arg (last_stmt, 3);
3901 vectype1 = get_vectype_for_scalar_type (TREE_TYPE (rhs2));
3904 rhs1 = gimple_call_arg (last_stmt, 2);
3905 rhs1_type = search_type_for_mask (rhs1, vinfo);
3906 if (!rhs1_type)
3907 return NULL;
3908 vectype2 = get_mask_type_for_scalar_type (rhs1_type);
3910 if (!vectype1 || !vectype2
3911 || TYPE_VECTOR_SUBPARTS (vectype1) == TYPE_VECTOR_SUBPARTS (vectype2))
3912 return NULL;
3914 tmp = build_mask_conversion (rhs1, vectype1, stmt_vinfo, vinfo);
3916 if (load)
3918 lhs = vect_recog_temp_ssa_var (TREE_TYPE (lhs), NULL);
3919 pattern_stmt
3920 = gimple_build_call_internal (IFN_MASK_LOAD, 3,
3921 gimple_call_arg (last_stmt, 0),
3922 gimple_call_arg (last_stmt, 1),
3923 tmp);
3924 gimple_call_set_lhs (pattern_stmt, lhs);
3926 else
3927 pattern_stmt
3928 = gimple_build_call_internal (IFN_MASK_STORE, 4,
3929 gimple_call_arg (last_stmt, 0),
3930 gimple_call_arg (last_stmt, 1),
3931 tmp,
3932 gimple_call_arg (last_stmt, 3));
3935 pattern_stmt_info = new_stmt_vec_info (pattern_stmt, vinfo);
3936 set_vinfo_for_stmt (pattern_stmt, pattern_stmt_info);
3937 STMT_VINFO_DATA_REF (pattern_stmt_info)
3938 = STMT_VINFO_DATA_REF (stmt_vinfo);
3939 STMT_VINFO_DR_BASE_ADDRESS (pattern_stmt_info)
3940 = STMT_VINFO_DR_BASE_ADDRESS (stmt_vinfo);
3941 STMT_VINFO_DR_INIT (pattern_stmt_info) = STMT_VINFO_DR_INIT (stmt_vinfo);
3942 STMT_VINFO_DR_OFFSET (pattern_stmt_info)
3943 = STMT_VINFO_DR_OFFSET (stmt_vinfo);
3944 STMT_VINFO_DR_STEP (pattern_stmt_info) = STMT_VINFO_DR_STEP (stmt_vinfo);
3945 STMT_VINFO_DR_ALIGNED_TO (pattern_stmt_info)
3946 = STMT_VINFO_DR_ALIGNED_TO (stmt_vinfo);
3947 DR_STMT (STMT_VINFO_DATA_REF (stmt_vinfo)) = pattern_stmt;
3949 *type_out = vectype1;
3950 *type_in = vectype1;
3951 stmts->safe_push (last_stmt);
3952 if (dump_enabled_p ())
3953 dump_printf_loc (MSG_NOTE, vect_location,
3954 "vect_recog_mask_conversion_pattern: detected:\n");
3956 return pattern_stmt;
3959 if (!is_gimple_assign (last_stmt))
3960 return NULL;
3962 lhs = gimple_assign_lhs (last_stmt);
3963 rhs1 = gimple_assign_rhs1 (last_stmt);
3964 rhs_code = gimple_assign_rhs_code (last_stmt);
3966 /* Check for cond expression requiring mask conversion. */
3967 if (rhs_code == COND_EXPR)
3969 /* vect_recog_mixed_size_cond_pattern could apply.
3970 Do nothing then. */
3971 if (STMT_VINFO_IN_PATTERN_P (stmt_vinfo))
3972 return NULL;
3974 vectype1 = get_vectype_for_scalar_type (TREE_TYPE (lhs));
3976 if (TREE_CODE (rhs1) == SSA_NAME)
3978 rhs1_type = search_type_for_mask (rhs1, vinfo);
3979 if (!rhs1_type)
3980 return NULL;
3982 else if (COMPARISON_CLASS_P (rhs1))
3983 rhs1_type = TREE_TYPE (TREE_OPERAND (rhs1, 0));
3984 else
3985 return NULL;
3987 vectype2 = get_mask_type_for_scalar_type (rhs1_type);
3989 if (!vectype1 || !vectype2
3990 || TYPE_VECTOR_SUBPARTS (vectype1) == TYPE_VECTOR_SUBPARTS (vectype2))
3991 return NULL;
3993 /* If rhs1 is a comparison we need to move it into a
3994 separate statement. */
3995 if (TREE_CODE (rhs1) != SSA_NAME)
3997 tmp = vect_recog_temp_ssa_var (TREE_TYPE (rhs1), NULL);
3998 pattern_stmt = gimple_build_assign (tmp, rhs1);
3999 rhs1 = tmp;
4001 pattern_stmt_info = new_stmt_vec_info (pattern_stmt, vinfo);
4002 set_vinfo_for_stmt (pattern_stmt, pattern_stmt_info);
4003 STMT_VINFO_VECTYPE (pattern_stmt_info) = vectype2;
4004 append_pattern_def_seq (stmt_vinfo, pattern_stmt);
4007 tmp = build_mask_conversion (rhs1, vectype1, stmt_vinfo, vinfo);
4009 lhs = vect_recog_temp_ssa_var (TREE_TYPE (lhs), NULL);
4010 pattern_stmt = gimple_build_assign (lhs, COND_EXPR, tmp,
4011 gimple_assign_rhs2 (last_stmt),
4012 gimple_assign_rhs3 (last_stmt));
4014 *type_out = vectype1;
4015 *type_in = vectype1;
4016 stmts->safe_push (last_stmt);
4017 if (dump_enabled_p ())
4018 dump_printf_loc (MSG_NOTE, vect_location,
4019 "vect_recog_mask_conversion_pattern: detected:\n");
4021 return pattern_stmt;
4024 /* Now check for binary boolean operations requiring conversion for
4025 one of operands. */
4026 if (TREE_CODE (TREE_TYPE (lhs)) != BOOLEAN_TYPE)
4027 return NULL;
4029 if (rhs_code != BIT_IOR_EXPR
4030 && rhs_code != BIT_XOR_EXPR
4031 && rhs_code != BIT_AND_EXPR
4032 && TREE_CODE_CLASS (rhs_code) != tcc_comparison)
4033 return NULL;
4035 rhs2 = gimple_assign_rhs2 (last_stmt);
4037 rhs1_type = search_type_for_mask (rhs1, vinfo);
4038 rhs2_type = search_type_for_mask (rhs2, vinfo);
4040 if (!rhs1_type || !rhs2_type
4041 || TYPE_PRECISION (rhs1_type) == TYPE_PRECISION (rhs2_type))
4042 return NULL;
4044 if (TYPE_PRECISION (rhs1_type) < TYPE_PRECISION (rhs2_type))
4046 vectype1 = get_mask_type_for_scalar_type (rhs1_type);
4047 if (!vectype1)
4048 return NULL;
4049 rhs2 = build_mask_conversion (rhs2, vectype1, stmt_vinfo, vinfo);
4051 else
4053 vectype1 = get_mask_type_for_scalar_type (rhs2_type);
4054 if (!vectype1)
4055 return NULL;
4056 rhs1 = build_mask_conversion (rhs1, vectype1, stmt_vinfo, vinfo);
4059 lhs = vect_recog_temp_ssa_var (TREE_TYPE (lhs), NULL);
4060 pattern_stmt = gimple_build_assign (lhs, rhs_code, rhs1, rhs2);
4062 *type_out = vectype1;
4063 *type_in = vectype1;
4064 stmts->safe_push (last_stmt);
4065 if (dump_enabled_p ())
4066 dump_printf_loc (MSG_NOTE, vect_location,
4067 "vect_recog_mask_conversion_pattern: detected:\n");
4069 return pattern_stmt;
4073 /* Mark statements that are involved in a pattern. */
4075 static inline void
4076 vect_mark_pattern_stmts (gimple *orig_stmt, gimple *pattern_stmt,
4077 tree pattern_vectype)
4079 stmt_vec_info pattern_stmt_info, def_stmt_info;
4080 stmt_vec_info orig_stmt_info = vinfo_for_stmt (orig_stmt);
4081 vec_info *vinfo = orig_stmt_info->vinfo;
4082 gimple *def_stmt;
4084 pattern_stmt_info = vinfo_for_stmt (pattern_stmt);
4085 if (pattern_stmt_info == NULL)
4087 pattern_stmt_info = new_stmt_vec_info (pattern_stmt, vinfo);
4088 set_vinfo_for_stmt (pattern_stmt, pattern_stmt_info);
4090 gimple_set_bb (pattern_stmt, gimple_bb (orig_stmt));
4092 STMT_VINFO_RELATED_STMT (pattern_stmt_info) = orig_stmt;
4093 STMT_VINFO_DEF_TYPE (pattern_stmt_info)
4094 = STMT_VINFO_DEF_TYPE (orig_stmt_info);
4095 STMT_VINFO_VECTYPE (pattern_stmt_info) = pattern_vectype;
4096 STMT_VINFO_IN_PATTERN_P (orig_stmt_info) = true;
4097 STMT_VINFO_RELATED_STMT (orig_stmt_info) = pattern_stmt;
4098 STMT_VINFO_PATTERN_DEF_SEQ (pattern_stmt_info)
4099 = STMT_VINFO_PATTERN_DEF_SEQ (orig_stmt_info);
4100 if (STMT_VINFO_PATTERN_DEF_SEQ (pattern_stmt_info))
4102 gimple_stmt_iterator si;
4103 for (si = gsi_start (STMT_VINFO_PATTERN_DEF_SEQ (pattern_stmt_info));
4104 !gsi_end_p (si); gsi_next (&si))
4106 def_stmt = gsi_stmt (si);
4107 def_stmt_info = vinfo_for_stmt (def_stmt);
4108 if (def_stmt_info == NULL)
4110 def_stmt_info = new_stmt_vec_info (def_stmt, vinfo);
4111 set_vinfo_for_stmt (def_stmt, def_stmt_info);
4113 gimple_set_bb (def_stmt, gimple_bb (orig_stmt));
4114 STMT_VINFO_RELATED_STMT (def_stmt_info) = orig_stmt;
4115 STMT_VINFO_DEF_TYPE (def_stmt_info) = vect_internal_def;
4116 if (STMT_VINFO_VECTYPE (def_stmt_info) == NULL_TREE)
4117 STMT_VINFO_VECTYPE (def_stmt_info) = pattern_vectype;
4122 /* Function vect_pattern_recog_1
4124 Input:
4125 PATTERN_RECOG_FUNC: A pointer to a function that detects a certain
4126 computation pattern.
4127 STMT: A stmt from which the pattern search should start.
4129 If PATTERN_RECOG_FUNC successfully detected the pattern, it creates an
4130 expression that computes the same functionality and can be used to
4131 replace the sequence of stmts that are involved in the pattern.
4133 Output:
4134 This function checks if the expression returned by PATTERN_RECOG_FUNC is
4135 supported in vector form by the target. We use 'TYPE_IN' to obtain the
4136 relevant vector type. If 'TYPE_IN' is already a vector type, then this
4137 indicates that target support had already been checked by PATTERN_RECOG_FUNC.
4138 If 'TYPE_OUT' is also returned by PATTERN_RECOG_FUNC, we check that it fits
4139 to the available target pattern.
4141 This function also does some bookkeeping, as explained in the documentation
4142 for vect_recog_pattern. */
4144 static bool
4145 vect_pattern_recog_1 (vect_recog_func *recog_func,
4146 gimple_stmt_iterator si,
4147 vec<gimple *> *stmts_to_replace)
4149 gimple *stmt = gsi_stmt (si), *pattern_stmt;
4150 stmt_vec_info stmt_info;
4151 loop_vec_info loop_vinfo;
4152 tree pattern_vectype;
4153 tree type_in, type_out;
4154 enum tree_code code;
4155 int i;
4156 gimple *next;
4158 stmts_to_replace->truncate (0);
4159 stmts_to_replace->quick_push (stmt);
4160 pattern_stmt = recog_func->fn (stmts_to_replace, &type_in, &type_out);
4161 if (!pattern_stmt)
4162 return false;
4164 stmt = stmts_to_replace->last ();
4165 stmt_info = vinfo_for_stmt (stmt);
4166 loop_vinfo = STMT_VINFO_LOOP_VINFO (stmt_info);
4168 if (VECTOR_BOOLEAN_TYPE_P (type_in)
4169 || VECTOR_MODE_P (TYPE_MODE (type_in)))
4171 /* No need to check target support (already checked by the pattern
4172 recognition function). */
4173 pattern_vectype = type_out ? type_out : type_in;
4175 else
4177 machine_mode vec_mode;
4178 enum insn_code icode;
4179 optab optab;
4181 /* Check target support */
4182 type_in = get_vectype_for_scalar_type (type_in);
4183 if (!type_in)
4184 return false;
4185 if (type_out)
4186 type_out = get_vectype_for_scalar_type (type_out);
4187 else
4188 type_out = type_in;
4189 if (!type_out)
4190 return false;
4191 pattern_vectype = type_out;
4193 if (is_gimple_assign (pattern_stmt))
4194 code = gimple_assign_rhs_code (pattern_stmt);
4195 else
4197 gcc_assert (is_gimple_call (pattern_stmt));
4198 code = CALL_EXPR;
4201 optab = optab_for_tree_code (code, type_in, optab_default);
4202 vec_mode = TYPE_MODE (type_in);
4203 if (!optab
4204 || (icode = optab_handler (optab, vec_mode)) == CODE_FOR_nothing
4205 || (insn_data[icode].operand[0].mode != TYPE_MODE (type_out)))
4206 return false;
4209 /* Found a vectorizable pattern. */
4210 if (dump_enabled_p ())
4212 dump_printf_loc (MSG_NOTE, vect_location,
4213 "%s pattern recognized: ", recog_func->name);
4214 dump_gimple_stmt (MSG_NOTE, TDF_SLIM, pattern_stmt, 0);
4217 /* Mark the stmts that are involved in the pattern. */
4218 vect_mark_pattern_stmts (stmt, pattern_stmt, pattern_vectype);
4220 /* Patterns cannot be vectorized using SLP, because they change the order of
4221 computation. */
4222 if (loop_vinfo)
4223 FOR_EACH_VEC_ELT (LOOP_VINFO_REDUCTIONS (loop_vinfo), i, next)
4224 if (next == stmt)
4225 LOOP_VINFO_REDUCTIONS (loop_vinfo).ordered_remove (i);
4227 /* It is possible that additional pattern stmts are created and inserted in
4228 STMTS_TO_REPLACE. We create a stmt_info for each of them, and mark the
4229 relevant statements. */
4230 for (i = 0; stmts_to_replace->iterate (i, &stmt)
4231 && (unsigned) i < (stmts_to_replace->length () - 1);
4232 i++)
4234 stmt_info = vinfo_for_stmt (stmt);
4235 pattern_stmt = STMT_VINFO_RELATED_STMT (stmt_info);
4236 if (dump_enabled_p ())
4238 dump_printf_loc (MSG_NOTE, vect_location,
4239 "additional pattern stmt: ");
4240 dump_gimple_stmt (MSG_NOTE, TDF_SLIM, pattern_stmt, 0);
4243 vect_mark_pattern_stmts (stmt, pattern_stmt, NULL_TREE);
4246 return true;
4250 /* Function vect_pattern_recog
4252 Input:
4253 LOOP_VINFO - a struct_loop_info of a loop in which we want to look for
4254 computation idioms.
4256 Output - for each computation idiom that is detected we create a new stmt
4257 that provides the same functionality and that can be vectorized. We
4258 also record some information in the struct_stmt_info of the relevant
4259 stmts, as explained below:
4261 At the entry to this function we have the following stmts, with the
4262 following initial value in the STMT_VINFO fields:
4264 stmt in_pattern_p related_stmt vec_stmt
4265 S1: a_i = .... - - -
4266 S2: a_2 = ..use(a_i).. - - -
4267 S3: a_1 = ..use(a_2).. - - -
4268 S4: a_0 = ..use(a_1).. - - -
4269 S5: ... = ..use(a_0).. - - -
4271 Say the sequence {S1,S2,S3,S4} was detected as a pattern that can be
4272 represented by a single stmt. We then:
4273 - create a new stmt S6 equivalent to the pattern (the stmt is not
4274 inserted into the code)
4275 - fill in the STMT_VINFO fields as follows:
4277 in_pattern_p related_stmt vec_stmt
4278 S1: a_i = .... - - -
4279 S2: a_2 = ..use(a_i).. - - -
4280 S3: a_1 = ..use(a_2).. - - -
4281 S4: a_0 = ..use(a_1).. true S6 -
4282 '---> S6: a_new = .... - S4 -
4283 S5: ... = ..use(a_0).. - - -
4285 (the last stmt in the pattern (S4) and the new pattern stmt (S6) point
4286 to each other through the RELATED_STMT field).
4288 S6 will be marked as relevant in vect_mark_stmts_to_be_vectorized instead
4289 of S4 because it will replace all its uses. Stmts {S1,S2,S3} will
4290 remain irrelevant unless used by stmts other than S4.
4292 If vectorization succeeds, vect_transform_stmt will skip over {S1,S2,S3}
4293 (because they are marked as irrelevant). It will vectorize S6, and record
4294 a pointer to the new vector stmt VS6 from S6 (as usual).
4295 S4 will be skipped, and S5 will be vectorized as usual:
4297 in_pattern_p related_stmt vec_stmt
4298 S1: a_i = .... - - -
4299 S2: a_2 = ..use(a_i).. - - -
4300 S3: a_1 = ..use(a_2).. - - -
4301 > VS6: va_new = .... - - -
4302 S4: a_0 = ..use(a_1).. true S6 VS6
4303 '---> S6: a_new = .... - S4 VS6
4304 > VS5: ... = ..vuse(va_new).. - - -
4305 S5: ... = ..use(a_0).. - - -
4307 DCE could then get rid of {S1,S2,S3,S4,S5} (if their defs are not used
4308 elsewhere), and we'll end up with:
4310 VS6: va_new = ....
4311 VS5: ... = ..vuse(va_new)..
4313 In case of more than one pattern statements, e.g., widen-mult with
4314 intermediate type:
4316 S1 a_t = ;
4317 S2 a_T = (TYPE) a_t;
4318 '--> S3: a_it = (interm_type) a_t;
4319 S4 prod_T = a_T * CONST;
4320 '--> S5: prod_T' = a_it w* CONST;
4322 there may be other users of a_T outside the pattern. In that case S2 will
4323 be marked as relevant (as well as S3), and both S2 and S3 will be analyzed
4324 and vectorized. The vector stmt VS2 will be recorded in S2, and VS3 will
4325 be recorded in S3. */
4327 void
4328 vect_pattern_recog (vec_info *vinfo)
4330 struct loop *loop;
4331 basic_block *bbs;
4332 unsigned int nbbs;
4333 gimple_stmt_iterator si;
4334 unsigned int i, j;
4335 auto_vec<gimple *, 1> stmts_to_replace;
4336 gimple *stmt;
4338 if (dump_enabled_p ())
4339 dump_printf_loc (MSG_NOTE, vect_location,
4340 "=== vect_pattern_recog ===\n");
4342 if (loop_vec_info loop_vinfo = dyn_cast <loop_vec_info> (vinfo))
4344 loop = LOOP_VINFO_LOOP (loop_vinfo);
4345 bbs = LOOP_VINFO_BBS (loop_vinfo);
4346 nbbs = loop->num_nodes;
4348 /* Scan through the loop stmts, applying the pattern recognition
4349 functions starting at each stmt visited: */
4350 for (i = 0; i < nbbs; i++)
4352 basic_block bb = bbs[i];
4353 for (si = gsi_start_bb (bb); !gsi_end_p (si); gsi_next (&si))
4355 /* Scan over all generic vect_recog_xxx_pattern functions. */
4356 for (j = 0; j < NUM_PATTERNS; j++)
4357 if (vect_pattern_recog_1 (&vect_vect_recog_func_ptrs[j], si,
4358 &stmts_to_replace))
4359 break;
4363 else
4365 bb_vec_info bb_vinfo = as_a <bb_vec_info> (vinfo);
4366 for (si = bb_vinfo->region_begin;
4367 gsi_stmt (si) != gsi_stmt (bb_vinfo->region_end); gsi_next (&si))
4369 if ((stmt = gsi_stmt (si))
4370 && vinfo_for_stmt (stmt)
4371 && !STMT_VINFO_VECTORIZABLE (vinfo_for_stmt (stmt)))
4372 continue;
4374 /* Scan over all generic vect_recog_xxx_pattern functions. */
4375 for (j = 0; j < NUM_PATTERNS; j++)
4376 if (vect_pattern_recog_1 (&vect_vect_recog_func_ptrs[j], si,
4377 &stmts_to_replace))
4378 break;