]>
Commit | Line | Data |
---|---|---|
5f40b3cb | 1 | /* Loop autoparallelization. |
d1e082c2 | 2 | Copyright (C) 2006-2013 Free Software Foundation, Inc. |
70837b71 RL |
3 | Contributed by Sebastian Pop <pop@cri.ensmp.fr> |
4 | Zdenek Dvorak <dvorakz@suse.cz> and Razya Ladelsky <razya@il.ibm.com>. | |
5f40b3cb ZD |
5 | |
6 | This file is part of GCC. | |
7 | ||
8 | GCC is free software; you can redistribute it and/or modify it under | |
9 | the terms of the GNU General Public License as published by the Free | |
6da7fc87 | 10 | Software Foundation; either version 3, or (at your option) any later |
5f40b3cb ZD |
11 | version. |
12 | ||
13 | GCC is distributed in the hope that it will be useful, but WITHOUT ANY | |
14 | WARRANTY; without even the implied warranty of MERCHANTABILITY or | |
15 | FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License | |
16 | for more details. | |
17 | ||
18 | You should have received a copy of the GNU General Public License | |
6da7fc87 NC |
19 | along with GCC; see the file COPYING3. If not see |
20 | <http://www.gnu.org/licenses/>. */ | |
5f40b3cb ZD |
21 | |
22 | #include "config.h" | |
23 | #include "system.h" | |
24 | #include "coretypes.h" | |
4d648807 | 25 | #include "tree.h" |
2fb9a547 AM |
26 | #include "basic-block.h" |
27 | #include "tree-ssa-alias.h" | |
28 | #include "internal-fn.h" | |
29 | #include "gimple-expr.h" | |
30 | #include "is-a.h" | |
18f429e2 | 31 | #include "gimple.h" |
45b0be94 | 32 | #include "gimplify.h" |
5be5c238 | 33 | #include "gimple-iterator.h" |
18f429e2 | 34 | #include "gimplify-me.h" |
5be5c238 | 35 | #include "gimple-walk.h" |
d8a2d370 DN |
36 | #include "stor-layout.h" |
37 | #include "tree-nested.h" | |
442b4905 AM |
38 | #include "gimple-ssa.h" |
39 | #include "tree-cfg.h" | |
40 | #include "tree-phinodes.h" | |
41 | #include "ssa-iterators.h" | |
d8a2d370 | 42 | #include "stringpool.h" |
442b4905 | 43 | #include "tree-ssanames.h" |
e28030cf AM |
44 | #include "tree-ssa-loop-ivopts.h" |
45 | #include "tree-ssa-loop-manip.h" | |
46 | #include "tree-ssa-loop-niter.h" | |
442b4905 AM |
47 | #include "tree-ssa-loop.h" |
48 | #include "tree-into-ssa.h" | |
5f40b3cb | 49 | #include "cfgloop.h" |
5f40b3cb | 50 | #include "tree-data-ref.h" |
1bd6497c | 51 | #include "tree-scalar-evolution.h" |
cf835838 | 52 | #include "gimple-pretty-print.h" |
5f40b3cb | 53 | #include "tree-pass.h" |
5f40b3cb | 54 | #include "langhooks.h" |
a509ebb5 | 55 | #include "tree-vectorizer.h" |
4a8fb1a1 | 56 | #include "tree-hasher.h" |
c1bf2a39 | 57 | #include "tree-parloops.h" |
0645c1a2 | 58 | #include "omp-low.h" |
1fe37220 | 59 | #include "tree-nested.h" |
5f40b3cb ZD |
60 | |
61 | /* This pass tries to distribute iterations of loops into several threads. | |
62 | The implementation is straightforward -- for each loop we test whether its | |
63 | iterations are independent, and if it is the case (and some additional | |
64 | conditions regarding profitability and correctness are satisfied), we | |
726a989a RB |
65 | add GIMPLE_OMP_PARALLEL and GIMPLE_OMP_FOR codes and let omp expansion |
66 | machinery do its job. | |
b8698a0f | 67 | |
5f40b3cb ZD |
68 | The most of the complexity is in bringing the code into shape expected |
69 | by the omp expanders: | |
726a989a RB |
70 | -- for GIMPLE_OMP_FOR, ensuring that the loop has only one induction |
71 | variable and that the exit test is at the start of the loop body | |
72 | -- for GIMPLE_OMP_PARALLEL, replacing the references to local addressable | |
5f40b3cb ZD |
73 | variables by accesses through pointers, and breaking up ssa chains |
74 | by storing the values incoming to the parallelized loop to a structure | |
75 | passed to the new function as an argument (something similar is done | |
76 | in omp gimplification, unfortunately only a small part of the code | |
77 | can be shared). | |
78 | ||
79 | TODO: | |
80 | -- if there are several parallelizable loops in a function, it may be | |
81 | possible to generate the threads just once (using synchronization to | |
82 | ensure that cross-loop dependences are obeyed). | |
70837b71 RL |
83 | -- handling of common reduction patterns for outer loops. |
84 | ||
85 | More info can also be found at http://gcc.gnu.org/wiki/AutoParInGCC */ | |
b8698a0f | 86 | /* |
a509ebb5 | 87 | Reduction handling: |
8a9ecffd | 88 | currently we use vect_force_simple_reduction() to detect reduction patterns. |
a509ebb5 | 89 | The code transformation will be introduced by an example. |
b8698a0f L |
90 | |
91 | ||
a509ebb5 RL |
92 | parloop |
93 | { | |
94 | int sum=1; | |
95 | ||
0eb7e7aa | 96 | for (i = 0; i < N; i++) |
a509ebb5 RL |
97 | { |
98 | x[i] = i + 3; | |
99 | sum+=x[i]; | |
100 | } | |
101 | } | |
102 | ||
0eb7e7aa | 103 | gimple-like code: |
a509ebb5 RL |
104 | header_bb: |
105 | ||
0eb7e7aa RL |
106 | # sum_29 = PHI <sum_11(5), 1(3)> |
107 | # i_28 = PHI <i_12(5), 0(3)> | |
108 | D.1795_8 = i_28 + 3; | |
109 | x[i_28] = D.1795_8; | |
110 | sum_11 = D.1795_8 + sum_29; | |
111 | i_12 = i_28 + 1; | |
112 | if (N_6(D) > i_12) | |
113 | goto header_bb; | |
114 | ||
a509ebb5 RL |
115 | |
116 | exit_bb: | |
117 | ||
0eb7e7aa RL |
118 | # sum_21 = PHI <sum_11(4)> |
119 | printf (&"%d"[0], sum_21); | |
a509ebb5 RL |
120 | |
121 | ||
122 | after reduction transformation (only relevant parts): | |
123 | ||
124 | parloop | |
125 | { | |
126 | ||
127 | .... | |
128 | ||
0eb7e7aa | 129 | |
fa10beec | 130 | # Storing the initial value given by the user. # |
0eb7e7aa | 131 | |
ae0bce62 | 132 | .paral_data_store.32.sum.27 = 1; |
b8698a0f L |
133 | |
134 | #pragma omp parallel num_threads(4) | |
a509ebb5 | 135 | |
0eb7e7aa | 136 | #pragma omp for schedule(static) |
ae0bce62 RL |
137 | |
138 | # The neutral element corresponding to the particular | |
139 | reduction's operation, e.g. 0 for PLUS_EXPR, | |
140 | 1 for MULT_EXPR, etc. replaces the user's initial value. # | |
141 | ||
142 | # sum.27_29 = PHI <sum.27_11, 0> | |
143 | ||
0eb7e7aa | 144 | sum.27_11 = D.1827_8 + sum.27_29; |
ae0bce62 | 145 | |
726a989a | 146 | GIMPLE_OMP_CONTINUE |
a509ebb5 | 147 | |
0eb7e7aa RL |
148 | # Adding this reduction phi is done at create_phi_for_local_result() # |
149 | # sum.27_56 = PHI <sum.27_11, 0> | |
726a989a | 150 | GIMPLE_OMP_RETURN |
b8698a0f L |
151 | |
152 | # Creating the atomic operation is done at | |
0eb7e7aa | 153 | create_call_for_reduction_1() # |
a509ebb5 | 154 | |
0eb7e7aa RL |
155 | #pragma omp atomic_load |
156 | D.1839_59 = *&.paral_data_load.33_51->reduction.23; | |
157 | D.1840_60 = sum.27_56 + D.1839_59; | |
158 | #pragma omp atomic_store (D.1840_60); | |
b8698a0f | 159 | |
726a989a | 160 | GIMPLE_OMP_RETURN |
b8698a0f | 161 | |
0eb7e7aa RL |
162 | # collecting the result after the join of the threads is done at |
163 | create_loads_for_reductions(). | |
ae0bce62 RL |
164 | The value computed by the threads is loaded from the |
165 | shared struct. # | |
166 | ||
b8698a0f | 167 | |
0eb7e7aa | 168 | .paral_data_load.33_52 = &.paral_data_store.32; |
ae0bce62 | 169 | sum_37 = .paral_data_load.33_52->sum.27; |
0eb7e7aa RL |
170 | sum_43 = D.1795_41 + sum_37; |
171 | ||
172 | exit bb: | |
173 | # sum_21 = PHI <sum_43, sum_26> | |
174 | printf (&"%d"[0], sum_21); | |
175 | ||
176 | ... | |
177 | ||
a509ebb5 RL |
178 | } |
179 | ||
180 | */ | |
181 | ||
5f40b3cb ZD |
182 | /* Minimal number of iterations of a loop that should be executed in each |
183 | thread. */ | |
184 | #define MIN_PER_THREAD 100 | |
185 | ||
b8698a0f | 186 | /* Element of the hashtable, representing a |
a509ebb5 RL |
187 | reduction in the current loop. */ |
188 | struct reduction_info | |
189 | { | |
726a989a RB |
190 | gimple reduc_stmt; /* reduction statement. */ |
191 | gimple reduc_phi; /* The phi node defining the reduction. */ | |
192 | enum tree_code reduction_code;/* code for the reduction operation. */ | |
5d1fd1de JJ |
193 | unsigned reduc_version; /* SSA_NAME_VERSION of original reduc_phi |
194 | result. */ | |
b8698a0f | 195 | gimple keep_res; /* The PHI_RESULT of this phi is the resulting value |
a509ebb5 | 196 | of the reduction variable when existing the loop. */ |
ae0bce62 | 197 | tree initial_value; /* The initial value of the reduction var before entering the loop. */ |
a509ebb5 | 198 | tree field; /* the name of the field in the parloop data structure intended for reduction. */ |
a509ebb5 | 199 | tree init; /* reduction initialization value. */ |
b8698a0f | 200 | gimple new_phi; /* (helper field) Newly created phi node whose result |
a509ebb5 RL |
201 | will be passed to the atomic operation. Represents |
202 | the local result each thread computed for the reduction | |
203 | operation. */ | |
204 | }; | |
205 | ||
4a8fb1a1 | 206 | /* Reduction info hashtable helpers. */ |
a509ebb5 | 207 | |
4a8fb1a1 | 208 | struct reduction_hasher : typed_free_remove <reduction_info> |
a509ebb5 | 209 | { |
4a8fb1a1 LC |
210 | typedef reduction_info value_type; |
211 | typedef reduction_info compare_type; | |
212 | static inline hashval_t hash (const value_type *); | |
213 | static inline bool equal (const value_type *, const compare_type *); | |
214 | }; | |
215 | ||
216 | /* Equality and hash functions for hashtab code. */ | |
a509ebb5 | 217 | |
4a8fb1a1 LC |
218 | inline bool |
219 | reduction_hasher::equal (const value_type *a, const compare_type *b) | |
220 | { | |
a509ebb5 RL |
221 | return (a->reduc_phi == b->reduc_phi); |
222 | } | |
223 | ||
4a8fb1a1 LC |
224 | inline hashval_t |
225 | reduction_hasher::hash (const value_type *a) | |
a509ebb5 | 226 | { |
5d1fd1de | 227 | return a->reduc_version; |
a509ebb5 RL |
228 | } |
229 | ||
4a8fb1a1 LC |
230 | typedef hash_table <reduction_hasher> reduction_info_table_type; |
231 | ||
232 | ||
a509ebb5 | 233 | static struct reduction_info * |
4a8fb1a1 | 234 | reduction_phi (reduction_info_table_type reduction_list, gimple phi) |
a509ebb5 RL |
235 | { |
236 | struct reduction_info tmpred, *red; | |
237 | ||
4a8fb1a1 | 238 | if (reduction_list.elements () == 0 || phi == NULL) |
a509ebb5 RL |
239 | return NULL; |
240 | ||
241 | tmpred.reduc_phi = phi; | |
5d1fd1de | 242 | tmpred.reduc_version = gimple_uid (phi); |
4a8fb1a1 | 243 | red = reduction_list.find (&tmpred); |
a509ebb5 RL |
244 | |
245 | return red; | |
246 | } | |
247 | ||
5f40b3cb ZD |
248 | /* Element of hashtable of names to copy. */ |
249 | ||
250 | struct name_to_copy_elt | |
251 | { | |
252 | unsigned version; /* The version of the name to copy. */ | |
253 | tree new_name; /* The new name used in the copy. */ | |
254 | tree field; /* The field of the structure used to pass the | |
255 | value. */ | |
256 | }; | |
257 | ||
4a8fb1a1 | 258 | /* Name copies hashtable helpers. */ |
5f40b3cb | 259 | |
4a8fb1a1 | 260 | struct name_to_copy_hasher : typed_free_remove <name_to_copy_elt> |
5f40b3cb | 261 | { |
4a8fb1a1 LC |
262 | typedef name_to_copy_elt value_type; |
263 | typedef name_to_copy_elt compare_type; | |
264 | static inline hashval_t hash (const value_type *); | |
265 | static inline bool equal (const value_type *, const compare_type *); | |
266 | }; | |
267 | ||
268 | /* Equality and hash functions for hashtab code. */ | |
5f40b3cb | 269 | |
4a8fb1a1 LC |
270 | inline bool |
271 | name_to_copy_hasher::equal (const value_type *a, const compare_type *b) | |
272 | { | |
5f40b3cb ZD |
273 | return a->version == b->version; |
274 | } | |
275 | ||
4a8fb1a1 LC |
276 | inline hashval_t |
277 | name_to_copy_hasher::hash (const value_type *a) | |
5f40b3cb | 278 | { |
5f40b3cb ZD |
279 | return (hashval_t) a->version; |
280 | } | |
281 | ||
4a8fb1a1 LC |
282 | typedef hash_table <name_to_copy_hasher> name_to_copy_table_type; |
283 | ||
b305e3da SP |
284 | /* A transformation matrix, which is a self-contained ROWSIZE x COLSIZE |
285 | matrix. Rather than use floats, we simply keep a single DENOMINATOR that | |
286 | represents the denominator for every element in the matrix. */ | |
287 | typedef struct lambda_trans_matrix_s | |
288 | { | |
289 | lambda_matrix matrix; | |
290 | int rowsize; | |
291 | int colsize; | |
292 | int denominator; | |
293 | } *lambda_trans_matrix; | |
294 | #define LTM_MATRIX(T) ((T)->matrix) | |
295 | #define LTM_ROWSIZE(T) ((T)->rowsize) | |
296 | #define LTM_COLSIZE(T) ((T)->colsize) | |
297 | #define LTM_DENOMINATOR(T) ((T)->denominator) | |
298 | ||
299 | /* Allocate a new transformation matrix. */ | |
300 | ||
301 | static lambda_trans_matrix | |
302 | lambda_trans_matrix_new (int colsize, int rowsize, | |
303 | struct obstack * lambda_obstack) | |
304 | { | |
305 | lambda_trans_matrix ret; | |
306 | ||
307 | ret = (lambda_trans_matrix) | |
308 | obstack_alloc (lambda_obstack, sizeof (struct lambda_trans_matrix_s)); | |
309 | LTM_MATRIX (ret) = lambda_matrix_new (rowsize, colsize, lambda_obstack); | |
310 | LTM_ROWSIZE (ret) = rowsize; | |
311 | LTM_COLSIZE (ret) = colsize; | |
312 | LTM_DENOMINATOR (ret) = 1; | |
313 | return ret; | |
314 | } | |
315 | ||
316 | /* Multiply a vector VEC by a matrix MAT. | |
317 | MAT is an M*N matrix, and VEC is a vector with length N. The result | |
318 | is stored in DEST which must be a vector of length M. */ | |
319 | ||
320 | static void | |
321 | lambda_matrix_vector_mult (lambda_matrix matrix, int m, int n, | |
322 | lambda_vector vec, lambda_vector dest) | |
323 | { | |
324 | int i, j; | |
325 | ||
326 | lambda_vector_clear (dest, m); | |
327 | for (i = 0; i < m; i++) | |
328 | for (j = 0; j < n; j++) | |
329 | dest[i] += matrix[i][j] * vec[j]; | |
330 | } | |
331 | ||
332 | /* Return true if TRANS is a legal transformation matrix that respects | |
333 | the dependence vectors in DISTS and DIRS. The conservative answer | |
334 | is false. | |
335 | ||
336 | "Wolfe proves that a unimodular transformation represented by the | |
337 | matrix T is legal when applied to a loop nest with a set of | |
338 | lexicographically non-negative distance vectors RDG if and only if | |
339 | for each vector d in RDG, (T.d >= 0) is lexicographically positive. | |
340 | i.e.: if and only if it transforms the lexicographically positive | |
341 | distance vectors to lexicographically positive vectors. Note that | |
342 | a unimodular matrix must transform the zero vector (and only it) to | |
343 | the zero vector." S.Muchnick. */ | |
344 | ||
345 | static bool | |
346 | lambda_transform_legal_p (lambda_trans_matrix trans, | |
347 | int nb_loops, | |
9771b263 | 348 | vec<ddr_p> dependence_relations) |
b305e3da SP |
349 | { |
350 | unsigned int i, j; | |
351 | lambda_vector distres; | |
352 | struct data_dependence_relation *ddr; | |
353 | ||
354 | gcc_assert (LTM_COLSIZE (trans) == nb_loops | |
355 | && LTM_ROWSIZE (trans) == nb_loops); | |
356 | ||
357 | /* When there are no dependences, the transformation is correct. */ | |
9771b263 | 358 | if (dependence_relations.length () == 0) |
b305e3da SP |
359 | return true; |
360 | ||
9771b263 | 361 | ddr = dependence_relations[0]; |
b305e3da SP |
362 | if (ddr == NULL) |
363 | return true; | |
364 | ||
365 | /* When there is an unknown relation in the dependence_relations, we | |
366 | know that it is no worth looking at this loop nest: give up. */ | |
367 | if (DDR_ARE_DEPENDENT (ddr) == chrec_dont_know) | |
368 | return false; | |
369 | ||
370 | distres = lambda_vector_new (nb_loops); | |
371 | ||
372 | /* For each distance vector in the dependence graph. */ | |
9771b263 | 373 | FOR_EACH_VEC_ELT (dependence_relations, i, ddr) |
b305e3da SP |
374 | { |
375 | /* Don't care about relations for which we know that there is no | |
376 | dependence, nor about read-read (aka. output-dependences): | |
377 | these data accesses can happen in any order. */ | |
378 | if (DDR_ARE_DEPENDENT (ddr) == chrec_known | |
379 | || (DR_IS_READ (DDR_A (ddr)) && DR_IS_READ (DDR_B (ddr)))) | |
380 | continue; | |
381 | ||
382 | /* Conservatively answer: "this transformation is not valid". */ | |
383 | if (DDR_ARE_DEPENDENT (ddr) == chrec_dont_know) | |
384 | return false; | |
385 | ||
386 | /* If the dependence could not be captured by a distance vector, | |
387 | conservatively answer that the transform is not valid. */ | |
388 | if (DDR_NUM_DIST_VECTS (ddr) == 0) | |
389 | return false; | |
390 | ||
391 | /* Compute trans.dist_vect */ | |
392 | for (j = 0; j < DDR_NUM_DIST_VECTS (ddr); j++) | |
393 | { | |
394 | lambda_matrix_vector_mult (LTM_MATRIX (trans), nb_loops, nb_loops, | |
395 | DDR_DIST_VECT (ddr, j), distres); | |
396 | ||
397 | if (!lambda_vector_lexico_pos (distres, nb_loops)) | |
398 | return false; | |
399 | } | |
400 | } | |
401 | return true; | |
402 | } | |
08dab97a RL |
403 | |
404 | /* Data dependency analysis. Returns true if the iterations of LOOP | |
405 | are independent on each other (that is, if we can execute them | |
406 | in parallel). */ | |
5f40b3cb ZD |
407 | |
408 | static bool | |
f873b205 | 409 | loop_parallel_p (struct loop *loop, struct obstack * parloop_obstack) |
5f40b3cb | 410 | { |
9771b263 DN |
411 | vec<ddr_p> dependence_relations; |
412 | vec<data_reference_p> datarefs; | |
5f40b3cb ZD |
413 | lambda_trans_matrix trans; |
414 | bool ret = false; | |
5f40b3cb ZD |
415 | |
416 | if (dump_file && (dump_flags & TDF_DETAILS)) | |
48710229 RL |
417 | { |
418 | fprintf (dump_file, "Considering loop %d\n", loop->num); | |
419 | if (!loop->inner) | |
420 | fprintf (dump_file, "loop is innermost\n"); | |
b8698a0f | 421 | else |
48710229 RL |
422 | fprintf (dump_file, "loop NOT innermost\n"); |
423 | } | |
5f40b3cb | 424 | |
5f40b3cb ZD |
425 | /* Check for problems with dependences. If the loop can be reversed, |
426 | the iterations are independent. */ | |
00f96dc9 | 427 | auto_vec<loop_p, 3> loop_nest; |
9771b263 | 428 | datarefs.create (10); |
07687835 | 429 | dependence_relations.create (100); |
9ca3d00e AB |
430 | if (! compute_data_dependences_for_loop (loop, true, &loop_nest, &datarefs, |
431 | &dependence_relations)) | |
432 | { | |
433 | if (dump_file && (dump_flags & TDF_DETAILS)) | |
434 | fprintf (dump_file, " FAILED: cannot analyze data dependencies\n"); | |
435 | ret = false; | |
436 | goto end; | |
437 | } | |
5f40b3cb ZD |
438 | if (dump_file && (dump_flags & TDF_DETAILS)) |
439 | dump_data_dependence_relations (dump_file, dependence_relations); | |
440 | ||
f873b205 | 441 | trans = lambda_trans_matrix_new (1, 1, parloop_obstack); |
5f40b3cb ZD |
442 | LTM_MATRIX (trans)[0][0] = -1; |
443 | ||
444 | if (lambda_transform_legal_p (trans, 1, dependence_relations)) | |
445 | { | |
446 | ret = true; | |
447 | if (dump_file && (dump_flags & TDF_DETAILS)) | |
448 | fprintf (dump_file, " SUCCESS: may be parallelized\n"); | |
449 | } | |
450 | else if (dump_file && (dump_flags & TDF_DETAILS)) | |
a509ebb5 RL |
451 | fprintf (dump_file, |
452 | " FAILED: data dependencies exist across iterations\n"); | |
5f40b3cb | 453 | |
9ca3d00e | 454 | end: |
5f40b3cb ZD |
455 | free_dependence_relations (dependence_relations); |
456 | free_data_refs (datarefs); | |
457 | ||
458 | return ret; | |
459 | } | |
460 | ||
1d4af1e8 SP |
461 | /* Return true when LOOP contains basic blocks marked with the |
462 | BB_IRREDUCIBLE_LOOP flag. */ | |
463 | ||
464 | static inline bool | |
465 | loop_has_blocks_with_irreducible_flag (struct loop *loop) | |
466 | { | |
467 | unsigned i; | |
468 | basic_block *bbs = get_loop_body_in_dom_order (loop); | |
469 | bool res = true; | |
470 | ||
471 | for (i = 0; i < loop->num_nodes; i++) | |
472 | if (bbs[i]->flags & BB_IRREDUCIBLE_LOOP) | |
473 | goto end; | |
474 | ||
475 | res = false; | |
476 | end: | |
477 | free (bbs); | |
478 | return res; | |
479 | } | |
480 | ||
8a171a59 | 481 | /* Assigns the address of OBJ in TYPE to an ssa name, and returns this name. |
9f9f72aa | 482 | The assignment statement is placed on edge ENTRY. DECL_ADDRESS maps decls |
8a171a59 | 483 | to their addresses that can be reused. The address of OBJ is known to |
cba1eb61 JJ |
484 | be invariant in the whole function. Other needed statements are placed |
485 | right before GSI. */ | |
5f40b3cb ZD |
486 | |
487 | static tree | |
4a8fb1a1 LC |
488 | take_address_of (tree obj, tree type, edge entry, |
489 | int_tree_htab_type decl_address, gimple_stmt_iterator *gsi) | |
5f40b3cb | 490 | { |
8a171a59 | 491 | int uid; |
4a8fb1a1 | 492 | int_tree_map **dslot; |
5f40b3cb | 493 | struct int_tree_map ielt, *nielt; |
83d5977e | 494 | tree *var_p, name, addr; |
726a989a RB |
495 | gimple stmt; |
496 | gimple_seq stmts; | |
5f40b3cb | 497 | |
8a171a59 ZD |
498 | /* Since the address of OBJ is invariant, the trees may be shared. |
499 | Avoid rewriting unrelated parts of the code. */ | |
500 | obj = unshare_expr (obj); | |
501 | for (var_p = &obj; | |
502 | handled_component_p (*var_p); | |
503 | var_p = &TREE_OPERAND (*var_p, 0)) | |
504 | continue; | |
8a171a59 | 505 | |
c9a410f0 RG |
506 | /* Canonicalize the access to base on a MEM_REF. */ |
507 | if (DECL_P (*var_p)) | |
508 | *var_p = build_simple_mem_ref (build_fold_addr_expr (*var_p)); | |
509 | ||
510 | /* Assign a canonical SSA name to the address of the base decl used | |
511 | in the address and share it for all accesses and addresses based | |
512 | on it. */ | |
513 | uid = DECL_UID (TREE_OPERAND (TREE_OPERAND (*var_p, 0), 0)); | |
5f40b3cb | 514 | ielt.uid = uid; |
4a8fb1a1 | 515 | dslot = decl_address.find_slot_with_hash (&ielt, uid, INSERT); |
5f40b3cb ZD |
516 | if (!*dslot) |
517 | { | |
cba1eb61 JJ |
518 | if (gsi == NULL) |
519 | return NULL; | |
c9a410f0 | 520 | addr = TREE_OPERAND (*var_p, 0); |
29b89442 JJ |
521 | const char *obj_name |
522 | = get_name (TREE_OPERAND (TREE_OPERAND (*var_p, 0), 0)); | |
523 | if (obj_name) | |
524 | name = make_temp_ssa_name (TREE_TYPE (addr), NULL, obj_name); | |
525 | else | |
526 | name = make_ssa_name (TREE_TYPE (addr), NULL); | |
83d5977e | 527 | stmt = gimple_build_assign (name, addr); |
726a989a | 528 | gsi_insert_on_edge_immediate (entry, stmt); |
5f40b3cb ZD |
529 | |
530 | nielt = XNEW (struct int_tree_map); | |
531 | nielt->uid = uid; | |
532 | nielt->to = name; | |
533 | *dslot = nielt; | |
5f40b3cb | 534 | } |
8a171a59 | 535 | else |
4a8fb1a1 | 536 | name = (*dslot)->to; |
5f40b3cb | 537 | |
c9a410f0 RG |
538 | /* Express the address in terms of the canonical SSA name. */ |
539 | TREE_OPERAND (*var_p, 0) = name; | |
cba1eb61 JJ |
540 | if (gsi == NULL) |
541 | return build_fold_addr_expr_with_type (obj, type); | |
542 | ||
c9a410f0 RG |
543 | name = force_gimple_operand (build_addr (obj, current_function_decl), |
544 | &stmts, true, NULL_TREE); | |
545 | if (!gimple_seq_empty_p (stmts)) | |
cba1eb61 | 546 | gsi_insert_seq_before (gsi, stmts, GSI_SAME_STMT); |
5f40b3cb | 547 | |
c9a410f0 | 548 | if (!useless_type_conversion_p (type, TREE_TYPE (name))) |
8a171a59 | 549 | { |
726a989a | 550 | name = force_gimple_operand (fold_convert (type, name), &stmts, true, |
8a171a59 | 551 | NULL_TREE); |
726a989a | 552 | if (!gimple_seq_empty_p (stmts)) |
cba1eb61 | 553 | gsi_insert_seq_before (gsi, stmts, GSI_SAME_STMT); |
8a171a59 | 554 | } |
5f40b3cb ZD |
555 | |
556 | return name; | |
557 | } | |
558 | ||
a509ebb5 | 559 | /* Callback for htab_traverse. Create the initialization statement |
b8698a0f | 560 | for reduction described in SLOT, and place it at the preheader of |
a509ebb5 RL |
561 | the loop described in DATA. */ |
562 | ||
4a8fb1a1 LC |
563 | int |
564 | initialize_reductions (reduction_info **slot, struct loop *loop) | |
a509ebb5 | 565 | { |
a509ebb5 | 566 | tree init, c; |
a509ebb5 RL |
567 | tree bvar, type, arg; |
568 | edge e; | |
569 | ||
4a8fb1a1 | 570 | struct reduction_info *const reduc = *slot; |
a509ebb5 | 571 | |
b8698a0f | 572 | /* Create initialization in preheader: |
a509ebb5 RL |
573 | reduction_variable = initialization value of reduction. */ |
574 | ||
b8698a0f | 575 | /* In the phi node at the header, replace the argument coming |
a509ebb5 RL |
576 | from the preheader with the reduction initialization value. */ |
577 | ||
578 | /* Create a new variable to initialize the reduction. */ | |
579 | type = TREE_TYPE (PHI_RESULT (reduc->reduc_phi)); | |
580 | bvar = create_tmp_var (type, "reduction"); | |
a509ebb5 | 581 | |
c2255bc4 AH |
582 | c = build_omp_clause (gimple_location (reduc->reduc_stmt), |
583 | OMP_CLAUSE_REDUCTION); | |
a509ebb5 | 584 | OMP_CLAUSE_REDUCTION_CODE (c) = reduc->reduction_code; |
726a989a | 585 | OMP_CLAUSE_DECL (c) = SSA_NAME_VAR (gimple_assign_lhs (reduc->reduc_stmt)); |
a509ebb5 RL |
586 | |
587 | init = omp_reduction_init (c, TREE_TYPE (bvar)); | |
588 | reduc->init = init; | |
589 | ||
b8698a0f L |
590 | /* Replace the argument representing the initialization value |
591 | with the initialization value for the reduction (neutral | |
592 | element for the particular operation, e.g. 0 for PLUS_EXPR, | |
593 | 1 for MULT_EXPR, etc). | |
594 | Keep the old value in a new variable "reduction_initial", | |
595 | that will be taken in consideration after the parallel | |
0eb7e7aa | 596 | computing is done. */ |
a509ebb5 RL |
597 | |
598 | e = loop_preheader_edge (loop); | |
599 | arg = PHI_ARG_DEF_FROM_EDGE (reduc->reduc_phi, e); | |
600 | /* Create new variable to hold the initial value. */ | |
a509ebb5 | 601 | |
a509ebb5 | 602 | SET_USE (PHI_ARG_DEF_PTR_FROM_EDGE |
0eb7e7aa | 603 | (reduc->reduc_phi, loop_preheader_edge (loop)), init); |
ae0bce62 | 604 | reduc->initial_value = arg; |
a509ebb5 RL |
605 | return 1; |
606 | } | |
5f40b3cb ZD |
607 | |
608 | struct elv_data | |
609 | { | |
726a989a | 610 | struct walk_stmt_info info; |
9f9f72aa | 611 | edge entry; |
4a8fb1a1 | 612 | int_tree_htab_type decl_address; |
cba1eb61 | 613 | gimple_stmt_iterator *gsi; |
5f40b3cb | 614 | bool changed; |
cba1eb61 | 615 | bool reset; |
5f40b3cb ZD |
616 | }; |
617 | ||
9f9f72aa AP |
618 | /* Eliminates references to local variables in *TP out of the single |
619 | entry single exit region starting at DTA->ENTRY. | |
620 | DECL_ADDRESS contains addresses of the references that had their | |
621 | address taken already. If the expression is changed, CHANGED is | |
622 | set to true. Callback for walk_tree. */ | |
a509ebb5 | 623 | |
5f40b3cb | 624 | static tree |
8a171a59 | 625 | eliminate_local_variables_1 (tree *tp, int *walk_subtrees, void *data) |
5f40b3cb | 626 | { |
3d9a9f94 | 627 | struct elv_data *const dta = (struct elv_data *) data; |
8a171a59 | 628 | tree t = *tp, var, addr, addr_type, type, obj; |
5f40b3cb ZD |
629 | |
630 | if (DECL_P (t)) | |
631 | { | |
632 | *walk_subtrees = 0; | |
633 | ||
634 | if (!SSA_VAR_P (t) || DECL_EXTERNAL (t)) | |
635 | return NULL_TREE; | |
636 | ||
637 | type = TREE_TYPE (t); | |
638 | addr_type = build_pointer_type (type); | |
cba1eb61 JJ |
639 | addr = take_address_of (t, addr_type, dta->entry, dta->decl_address, |
640 | dta->gsi); | |
641 | if (dta->gsi == NULL && addr == NULL_TREE) | |
642 | { | |
643 | dta->reset = true; | |
644 | return NULL_TREE; | |
645 | } | |
646 | ||
70f34814 | 647 | *tp = build_simple_mem_ref (addr); |
5f40b3cb ZD |
648 | |
649 | dta->changed = true; | |
650 | return NULL_TREE; | |
651 | } | |
652 | ||
653 | if (TREE_CODE (t) == ADDR_EXPR) | |
654 | { | |
8a171a59 ZD |
655 | /* ADDR_EXPR may appear in two contexts: |
656 | -- as a gimple operand, when the address taken is a function invariant | |
657 | -- as gimple rhs, when the resulting address in not a function | |
658 | invariant | |
659 | We do not need to do anything special in the latter case (the base of | |
660 | the memory reference whose address is taken may be replaced in the | |
661 | DECL_P case). The former case is more complicated, as we need to | |
662 | ensure that the new address is still a gimple operand. Thus, it | |
663 | is not sufficient to replace just the base of the memory reference -- | |
664 | we need to move the whole computation of the address out of the | |
665 | loop. */ | |
666 | if (!is_gimple_val (t)) | |
5f40b3cb ZD |
667 | return NULL_TREE; |
668 | ||
669 | *walk_subtrees = 0; | |
8a171a59 ZD |
670 | obj = TREE_OPERAND (t, 0); |
671 | var = get_base_address (obj); | |
672 | if (!var || !SSA_VAR_P (var) || DECL_EXTERNAL (var)) | |
5f40b3cb ZD |
673 | return NULL_TREE; |
674 | ||
675 | addr_type = TREE_TYPE (t); | |
cba1eb61 JJ |
676 | addr = take_address_of (obj, addr_type, dta->entry, dta->decl_address, |
677 | dta->gsi); | |
678 | if (dta->gsi == NULL && addr == NULL_TREE) | |
679 | { | |
680 | dta->reset = true; | |
681 | return NULL_TREE; | |
682 | } | |
5f40b3cb ZD |
683 | *tp = addr; |
684 | ||
685 | dta->changed = true; | |
686 | return NULL_TREE; | |
687 | } | |
688 | ||
726a989a | 689 | if (!EXPR_P (t)) |
5f40b3cb ZD |
690 | *walk_subtrees = 0; |
691 | ||
692 | return NULL_TREE; | |
693 | } | |
694 | ||
cba1eb61 | 695 | /* Moves the references to local variables in STMT at *GSI out of the single |
9f9f72aa AP |
696 | entry single exit region starting at ENTRY. DECL_ADDRESS contains |
697 | addresses of the references that had their address taken | |
698 | already. */ | |
5f40b3cb ZD |
699 | |
700 | static void | |
cba1eb61 | 701 | eliminate_local_variables_stmt (edge entry, gimple_stmt_iterator *gsi, |
4a8fb1a1 | 702 | int_tree_htab_type decl_address) |
5f40b3cb ZD |
703 | { |
704 | struct elv_data dta; | |
cba1eb61 | 705 | gimple stmt = gsi_stmt (*gsi); |
5f40b3cb | 706 | |
726a989a | 707 | memset (&dta.info, '\0', sizeof (dta.info)); |
9f9f72aa | 708 | dta.entry = entry; |
5f40b3cb ZD |
709 | dta.decl_address = decl_address; |
710 | dta.changed = false; | |
cba1eb61 | 711 | dta.reset = false; |
5f40b3cb | 712 | |
b5b8b0ac | 713 | if (gimple_debug_bind_p (stmt)) |
cba1eb61 JJ |
714 | { |
715 | dta.gsi = NULL; | |
716 | walk_tree (gimple_debug_bind_get_value_ptr (stmt), | |
717 | eliminate_local_variables_1, &dta.info, NULL); | |
718 | if (dta.reset) | |
719 | { | |
720 | gimple_debug_bind_reset_value (stmt); | |
721 | dta.changed = true; | |
722 | } | |
723 | } | |
29b89442 JJ |
724 | else if (gimple_clobber_p (stmt)) |
725 | { | |
726 | stmt = gimple_build_nop (); | |
727 | gsi_replace (gsi, stmt, false); | |
728 | dta.changed = true; | |
729 | } | |
b5b8b0ac | 730 | else |
cba1eb61 JJ |
731 | { |
732 | dta.gsi = gsi; | |
733 | walk_gimple_op (stmt, eliminate_local_variables_1, &dta.info); | |
734 | } | |
5f40b3cb ZD |
735 | |
736 | if (dta.changed) | |
737 | update_stmt (stmt); | |
738 | } | |
739 | ||
9f9f72aa AP |
740 | /* Eliminates the references to local variables from the single entry |
741 | single exit region between the ENTRY and EXIT edges. | |
b8698a0f | 742 | |
a509ebb5 | 743 | This includes: |
b8698a0f L |
744 | 1) Taking address of a local variable -- these are moved out of the |
745 | region (and temporary variable is created to hold the address if | |
a509ebb5 | 746 | necessary). |
9f9f72aa | 747 | |
5f40b3cb | 748 | 2) Dereferencing a local variable -- these are replaced with indirect |
a509ebb5 | 749 | references. */ |
5f40b3cb ZD |
750 | |
751 | static void | |
9f9f72aa | 752 | eliminate_local_variables (edge entry, edge exit) |
5f40b3cb | 753 | { |
9f9f72aa | 754 | basic_block bb; |
00f96dc9 | 755 | auto_vec<basic_block, 3> body; |
5f40b3cb | 756 | unsigned i; |
726a989a | 757 | gimple_stmt_iterator gsi; |
cba1eb61 | 758 | bool has_debug_stmt = false; |
4a8fb1a1 LC |
759 | int_tree_htab_type decl_address; |
760 | decl_address.create (10); | |
9f9f72aa AP |
761 | basic_block entry_bb = entry->src; |
762 | basic_block exit_bb = exit->dest; | |
5f40b3cb | 763 | |
9f9f72aa | 764 | gather_blocks_in_sese_region (entry_bb, exit_bb, &body); |
5f40b3cb | 765 | |
9771b263 | 766 | FOR_EACH_VEC_ELT (body, i, bb) |
9f9f72aa | 767 | if (bb != entry_bb && bb != exit_bb) |
726a989a | 768 | for (gsi = gsi_start_bb (bb); !gsi_end_p (gsi); gsi_next (&gsi)) |
ddb555ed JJ |
769 | if (is_gimple_debug (gsi_stmt (gsi))) |
770 | { | |
771 | if (gimple_debug_bind_p (gsi_stmt (gsi))) | |
772 | has_debug_stmt = true; | |
773 | } | |
cba1eb61 JJ |
774 | else |
775 | eliminate_local_variables_stmt (entry, &gsi, decl_address); | |
776 | ||
777 | if (has_debug_stmt) | |
9771b263 | 778 | FOR_EACH_VEC_ELT (body, i, bb) |
cba1eb61 JJ |
779 | if (bb != entry_bb && bb != exit_bb) |
780 | for (gsi = gsi_start_bb (bb); !gsi_end_p (gsi); gsi_next (&gsi)) | |
781 | if (gimple_debug_bind_p (gsi_stmt (gsi))) | |
782 | eliminate_local_variables_stmt (entry, &gsi, decl_address); | |
5f40b3cb | 783 | |
4a8fb1a1 | 784 | decl_address.dispose (); |
9f9f72aa AP |
785 | } |
786 | ||
787 | /* Returns true if expression EXPR is not defined between ENTRY and | |
788 | EXIT, i.e. if all its operands are defined outside of the region. */ | |
789 | ||
790 | static bool | |
791 | expr_invariant_in_region_p (edge entry, edge exit, tree expr) | |
792 | { | |
793 | basic_block entry_bb = entry->src; | |
794 | basic_block exit_bb = exit->dest; | |
795 | basic_block def_bb; | |
9f9f72aa AP |
796 | |
797 | if (is_gimple_min_invariant (expr)) | |
798 | return true; | |
799 | ||
800 | if (TREE_CODE (expr) == SSA_NAME) | |
801 | { | |
726a989a | 802 | def_bb = gimple_bb (SSA_NAME_DEF_STMT (expr)); |
9f9f72aa AP |
803 | if (def_bb |
804 | && dominated_by_p (CDI_DOMINATORS, def_bb, entry_bb) | |
805 | && !dominated_by_p (CDI_DOMINATORS, def_bb, exit_bb)) | |
806 | return false; | |
807 | ||
808 | return true; | |
809 | } | |
810 | ||
726a989a | 811 | return false; |
5f40b3cb ZD |
812 | } |
813 | ||
814 | /* If COPY_NAME_P is true, creates and returns a duplicate of NAME. | |
815 | The copies are stored to NAME_COPIES, if NAME was already duplicated, | |
816 | its duplicate stored in NAME_COPIES is returned. | |
b8698a0f | 817 | |
5f40b3cb ZD |
818 | Regardless of COPY_NAME_P, the decl used as a base of the ssa name is also |
819 | duplicated, storing the copies in DECL_COPIES. */ | |
820 | ||
821 | static tree | |
4a8fb1a1 LC |
822 | separate_decls_in_region_name (tree name, name_to_copy_table_type name_copies, |
823 | int_tree_htab_type decl_copies, bool copy_name_p) | |
5f40b3cb ZD |
824 | { |
825 | tree copy, var, var_copy; | |
826 | unsigned idx, uid, nuid; | |
827 | struct int_tree_map ielt, *nielt; | |
828 | struct name_to_copy_elt elt, *nelt; | |
4a8fb1a1 LC |
829 | name_to_copy_elt **slot; |
830 | int_tree_map **dslot; | |
5f40b3cb ZD |
831 | |
832 | if (TREE_CODE (name) != SSA_NAME) | |
833 | return name; | |
834 | ||
835 | idx = SSA_NAME_VERSION (name); | |
836 | elt.version = idx; | |
4a8fb1a1 LC |
837 | slot = name_copies.find_slot_with_hash (&elt, idx, |
838 | copy_name_p ? INSERT : NO_INSERT); | |
5f40b3cb | 839 | if (slot && *slot) |
4a8fb1a1 | 840 | return (*slot)->new_name; |
5f40b3cb | 841 | |
70b5e7dc RG |
842 | if (copy_name_p) |
843 | { | |
844 | copy = duplicate_ssa_name (name, NULL); | |
845 | nelt = XNEW (struct name_to_copy_elt); | |
846 | nelt->version = idx; | |
847 | nelt->new_name = copy; | |
848 | nelt->field = NULL_TREE; | |
849 | *slot = nelt; | |
850 | } | |
851 | else | |
852 | { | |
853 | gcc_assert (!slot); | |
854 | copy = name; | |
855 | } | |
856 | ||
5f40b3cb | 857 | var = SSA_NAME_VAR (name); |
70b5e7dc RG |
858 | if (!var) |
859 | return copy; | |
860 | ||
5f40b3cb ZD |
861 | uid = DECL_UID (var); |
862 | ielt.uid = uid; | |
4a8fb1a1 | 863 | dslot = decl_copies.find_slot_with_hash (&ielt, uid, INSERT); |
5f40b3cb ZD |
864 | if (!*dslot) |
865 | { | |
866 | var_copy = create_tmp_var (TREE_TYPE (var), get_name (var)); | |
36ad7922 | 867 | DECL_GIMPLE_REG_P (var_copy) = DECL_GIMPLE_REG_P (var); |
5f40b3cb ZD |
868 | nielt = XNEW (struct int_tree_map); |
869 | nielt->uid = uid; | |
870 | nielt->to = var_copy; | |
871 | *dslot = nielt; | |
872 | ||
873 | /* Ensure that when we meet this decl next time, we won't duplicate | |
a509ebb5 | 874 | it again. */ |
5f40b3cb ZD |
875 | nuid = DECL_UID (var_copy); |
876 | ielt.uid = nuid; | |
4a8fb1a1 | 877 | dslot = decl_copies.find_slot_with_hash (&ielt, nuid, INSERT); |
5f40b3cb ZD |
878 | gcc_assert (!*dslot); |
879 | nielt = XNEW (struct int_tree_map); | |
880 | nielt->uid = nuid; | |
881 | nielt->to = var_copy; | |
882 | *dslot = nielt; | |
883 | } | |
884 | else | |
885 | var_copy = ((struct int_tree_map *) *dslot)->to; | |
886 | ||
b2ec94d4 | 887 | replace_ssa_name_symbol (copy, var_copy); |
5f40b3cb ZD |
888 | return copy; |
889 | } | |
890 | ||
9f9f72aa AP |
891 | /* Finds the ssa names used in STMT that are defined outside the |
892 | region between ENTRY and EXIT and replaces such ssa names with | |
893 | their duplicates. The duplicates are stored to NAME_COPIES. Base | |
894 | decls of all ssa names used in STMT (including those defined in | |
895 | LOOP) are replaced with the new temporary variables; the | |
896 | replacement decls are stored in DECL_COPIES. */ | |
5f40b3cb ZD |
897 | |
898 | static void | |
726a989a | 899 | separate_decls_in_region_stmt (edge entry, edge exit, gimple stmt, |
4a8fb1a1 LC |
900 | name_to_copy_table_type name_copies, |
901 | int_tree_htab_type decl_copies) | |
5f40b3cb ZD |
902 | { |
903 | use_operand_p use; | |
904 | def_operand_p def; | |
905 | ssa_op_iter oi; | |
906 | tree name, copy; | |
907 | bool copy_name_p; | |
908 | ||
5f40b3cb | 909 | FOR_EACH_PHI_OR_STMT_DEF (def, stmt, oi, SSA_OP_DEF) |
a509ebb5 RL |
910 | { |
911 | name = DEF_FROM_PTR (def); | |
912 | gcc_assert (TREE_CODE (name) == SSA_NAME); | |
9f9f72aa AP |
913 | copy = separate_decls_in_region_name (name, name_copies, decl_copies, |
914 | false); | |
a509ebb5 RL |
915 | gcc_assert (copy == name); |
916 | } | |
5f40b3cb ZD |
917 | |
918 | FOR_EACH_PHI_OR_STMT_USE (use, stmt, oi, SSA_OP_USE) | |
a509ebb5 RL |
919 | { |
920 | name = USE_FROM_PTR (use); | |
921 | if (TREE_CODE (name) != SSA_NAME) | |
922 | continue; | |
923 | ||
9f9f72aa AP |
924 | copy_name_p = expr_invariant_in_region_p (entry, exit, name); |
925 | copy = separate_decls_in_region_name (name, name_copies, decl_copies, | |
926 | copy_name_p); | |
a509ebb5 RL |
927 | SET_USE (use, copy); |
928 | } | |
5f40b3cb ZD |
929 | } |
930 | ||
b5b8b0ac AO |
931 | /* Finds the ssa names used in STMT that are defined outside the |
932 | region between ENTRY and EXIT and replaces such ssa names with | |
933 | their duplicates. The duplicates are stored to NAME_COPIES. Base | |
934 | decls of all ssa names used in STMT (including those defined in | |
935 | LOOP) are replaced with the new temporary variables; the | |
936 | replacement decls are stored in DECL_COPIES. */ | |
937 | ||
938 | static bool | |
4a8fb1a1 LC |
939 | separate_decls_in_region_debug (gimple stmt, |
940 | name_to_copy_table_type name_copies, | |
941 | int_tree_htab_type decl_copies) | |
b5b8b0ac AO |
942 | { |
943 | use_operand_p use; | |
944 | ssa_op_iter oi; | |
945 | tree var, name; | |
946 | struct int_tree_map ielt; | |
947 | struct name_to_copy_elt elt; | |
4a8fb1a1 LC |
948 | name_to_copy_elt **slot; |
949 | int_tree_map **dslot; | |
b5b8b0ac | 950 | |
ddb555ed JJ |
951 | if (gimple_debug_bind_p (stmt)) |
952 | var = gimple_debug_bind_get_var (stmt); | |
953 | else if (gimple_debug_source_bind_p (stmt)) | |
954 | var = gimple_debug_source_bind_get_var (stmt); | |
955 | else | |
956 | return true; | |
598e67d7 | 957 | if (TREE_CODE (var) == DEBUG_EXPR_DECL || TREE_CODE (var) == LABEL_DECL) |
4f2a9af8 | 958 | return true; |
b5b8b0ac AO |
959 | gcc_assert (DECL_P (var) && SSA_VAR_P (var)); |
960 | ielt.uid = DECL_UID (var); | |
4a8fb1a1 | 961 | dslot = decl_copies.find_slot_with_hash (&ielt, ielt.uid, NO_INSERT); |
b5b8b0ac AO |
962 | if (!dslot) |
963 | return true; | |
ddb555ed JJ |
964 | if (gimple_debug_bind_p (stmt)) |
965 | gimple_debug_bind_set_var (stmt, ((struct int_tree_map *) *dslot)->to); | |
966 | else if (gimple_debug_source_bind_p (stmt)) | |
967 | gimple_debug_source_bind_set_var (stmt, ((struct int_tree_map *) *dslot)->to); | |
b5b8b0ac AO |
968 | |
969 | FOR_EACH_PHI_OR_STMT_USE (use, stmt, oi, SSA_OP_USE) | |
970 | { | |
971 | name = USE_FROM_PTR (use); | |
972 | if (TREE_CODE (name) != SSA_NAME) | |
973 | continue; | |
974 | ||
975 | elt.version = SSA_NAME_VERSION (name); | |
4a8fb1a1 | 976 | slot = name_copies.find_slot_with_hash (&elt, elt.version, NO_INSERT); |
b5b8b0ac AO |
977 | if (!slot) |
978 | { | |
979 | gimple_debug_bind_reset_value (stmt); | |
980 | update_stmt (stmt); | |
981 | break; | |
982 | } | |
983 | ||
4a8fb1a1 | 984 | SET_USE (use, (*slot)->new_name); |
b5b8b0ac AO |
985 | } |
986 | ||
987 | return false; | |
988 | } | |
989 | ||
0eb7e7aa RL |
990 | /* Callback for htab_traverse. Adds a field corresponding to the reduction |
991 | specified in SLOT. The type is passed in DATA. */ | |
992 | ||
4a8fb1a1 LC |
993 | int |
994 | add_field_for_reduction (reduction_info **slot, tree type) | |
a509ebb5 | 995 | { |
b8698a0f | 996 | |
4a8fb1a1 | 997 | struct reduction_info *const red = *slot; |
aa06a978 RB |
998 | tree var = gimple_assign_lhs (red->reduc_stmt); |
999 | tree field = build_decl (gimple_location (red->reduc_stmt), FIELD_DECL, | |
1000 | SSA_NAME_IDENTIFIER (var), TREE_TYPE (var)); | |
0eb7e7aa RL |
1001 | |
1002 | insert_field_into_struct (type, field); | |
1003 | ||
1004 | red->field = field; | |
1005 | ||
1006 | return 1; | |
1007 | } | |
a509ebb5 | 1008 | |
5f40b3cb | 1009 | /* Callback for htab_traverse. Adds a field corresponding to a ssa name |
b8698a0f | 1010 | described in SLOT. The type is passed in DATA. */ |
5f40b3cb | 1011 | |
4a8fb1a1 LC |
1012 | int |
1013 | add_field_for_name (name_to_copy_elt **slot, tree type) | |
5f40b3cb | 1014 | { |
4a8fb1a1 | 1015 | struct name_to_copy_elt *const elt = *slot; |
5f40b3cb | 1016 | tree name = ssa_name (elt->version); |
70b5e7dc RG |
1017 | tree field = build_decl (UNKNOWN_LOCATION, |
1018 | FIELD_DECL, SSA_NAME_IDENTIFIER (name), | |
1019 | TREE_TYPE (name)); | |
5f40b3cb ZD |
1020 | |
1021 | insert_field_into_struct (type, field); | |
1022 | elt->field = field; | |
a509ebb5 | 1023 | |
5f40b3cb ZD |
1024 | return 1; |
1025 | } | |
1026 | ||
b8698a0f L |
1027 | /* Callback for htab_traverse. A local result is the intermediate result |
1028 | computed by a single | |
fa10beec | 1029 | thread, or the initial value in case no iteration was executed. |
b8698a0f L |
1030 | This function creates a phi node reflecting these values. |
1031 | The phi's result will be stored in NEW_PHI field of the | |
1032 | reduction's data structure. */ | |
a509ebb5 | 1033 | |
4a8fb1a1 LC |
1034 | int |
1035 | create_phi_for_local_result (reduction_info **slot, struct loop *loop) | |
a509ebb5 | 1036 | { |
4a8fb1a1 | 1037 | struct reduction_info *const reduc = *slot; |
a509ebb5 | 1038 | edge e; |
726a989a | 1039 | gimple new_phi; |
a509ebb5 RL |
1040 | basic_block store_bb; |
1041 | tree local_res; | |
f5045c96 | 1042 | source_location locus; |
a509ebb5 | 1043 | |
b8698a0f L |
1044 | /* STORE_BB is the block where the phi |
1045 | should be stored. It is the destination of the loop exit. | |
726a989a | 1046 | (Find the fallthru edge from GIMPLE_OMP_CONTINUE). */ |
a509ebb5 RL |
1047 | store_bb = FALLTHRU_EDGE (loop->latch)->dest; |
1048 | ||
1049 | /* STORE_BB has two predecessors. One coming from the loop | |
1050 | (the reduction's result is computed at the loop), | |
b8698a0f L |
1051 | and another coming from a block preceding the loop, |
1052 | when no iterations | |
1053 | are executed (the initial value should be taken). */ | |
a509ebb5 RL |
1054 | if (EDGE_PRED (store_bb, 0) == FALLTHRU_EDGE (loop->latch)) |
1055 | e = EDGE_PRED (store_bb, 1); | |
1056 | else | |
1057 | e = EDGE_PRED (store_bb, 0); | |
6b4a85ad | 1058 | local_res = copy_ssa_name (gimple_assign_lhs (reduc->reduc_stmt), NULL); |
f5045c96 | 1059 | locus = gimple_location (reduc->reduc_stmt); |
a509ebb5 | 1060 | new_phi = create_phi_node (local_res, store_bb); |
9e227d60 | 1061 | add_phi_arg (new_phi, reduc->init, e, locus); |
726a989a | 1062 | add_phi_arg (new_phi, gimple_assign_lhs (reduc->reduc_stmt), |
9e227d60 | 1063 | FALLTHRU_EDGE (loop->latch), locus); |
a509ebb5 RL |
1064 | reduc->new_phi = new_phi; |
1065 | ||
1066 | return 1; | |
1067 | } | |
5f40b3cb ZD |
1068 | |
1069 | struct clsn_data | |
1070 | { | |
1071 | tree store; | |
1072 | tree load; | |
1073 | ||
1074 | basic_block store_bb; | |
1075 | basic_block load_bb; | |
1076 | }; | |
1077 | ||
a509ebb5 | 1078 | /* Callback for htab_traverse. Create an atomic instruction for the |
b8698a0f | 1079 | reduction described in SLOT. |
a509ebb5 RL |
1080 | DATA annotates the place in memory the atomic operation relates to, |
1081 | and the basic block it needs to be generated in. */ | |
1082 | ||
4a8fb1a1 LC |
1083 | int |
1084 | create_call_for_reduction_1 (reduction_info **slot, struct clsn_data *clsn_data) | |
a509ebb5 | 1085 | { |
4a8fb1a1 | 1086 | struct reduction_info *const reduc = *slot; |
726a989a | 1087 | gimple_stmt_iterator gsi; |
a509ebb5 | 1088 | tree type = TREE_TYPE (PHI_RESULT (reduc->reduc_phi)); |
a509ebb5 RL |
1089 | tree load_struct; |
1090 | basic_block bb; | |
1091 | basic_block new_bb; | |
1092 | edge e; | |
0f900dfa | 1093 | tree t, addr, ref, x; |
726a989a RB |
1094 | tree tmp_load, name; |
1095 | gimple load; | |
a509ebb5 | 1096 | |
70f34814 | 1097 | load_struct = build_simple_mem_ref (clsn_data->load); |
a509ebb5 | 1098 | t = build3 (COMPONENT_REF, type, load_struct, reduc->field, NULL_TREE); |
a509ebb5 RL |
1099 | |
1100 | addr = build_addr (t, current_function_decl); | |
1101 | ||
1102 | /* Create phi node. */ | |
1103 | bb = clsn_data->load_bb; | |
1104 | ||
1105 | e = split_block (bb, t); | |
1106 | new_bb = e->dest; | |
1107 | ||
1108 | tmp_load = create_tmp_var (TREE_TYPE (TREE_TYPE (addr)), NULL); | |
a509ebb5 | 1109 | tmp_load = make_ssa_name (tmp_load, NULL); |
726a989a | 1110 | load = gimple_build_omp_atomic_load (tmp_load, addr); |
a509ebb5 | 1111 | SSA_NAME_DEF_STMT (tmp_load) = load; |
726a989a RB |
1112 | gsi = gsi_start_bb (new_bb); |
1113 | gsi_insert_after (&gsi, load, GSI_NEW_STMT); | |
a509ebb5 RL |
1114 | |
1115 | e = split_block (new_bb, load); | |
1116 | new_bb = e->dest; | |
726a989a | 1117 | gsi = gsi_start_bb (new_bb); |
a509ebb5 | 1118 | ref = tmp_load; |
726a989a RB |
1119 | x = fold_build2 (reduc->reduction_code, |
1120 | TREE_TYPE (PHI_RESULT (reduc->new_phi)), ref, | |
1121 | PHI_RESULT (reduc->new_phi)); | |
a509ebb5 | 1122 | |
726a989a RB |
1123 | name = force_gimple_operand_gsi (&gsi, x, true, NULL_TREE, true, |
1124 | GSI_CONTINUE_LINKING); | |
a509ebb5 | 1125 | |
726a989a | 1126 | gsi_insert_after (&gsi, gimple_build_omp_atomic_store (name), GSI_NEW_STMT); |
a509ebb5 RL |
1127 | return 1; |
1128 | } | |
1129 | ||
b8698a0f L |
1130 | /* Create the atomic operation at the join point of the threads. |
1131 | REDUCTION_LIST describes the reductions in the LOOP. | |
1132 | LD_ST_DATA describes the shared data structure where | |
a509ebb5 RL |
1133 | shared data is stored in and loaded from. */ |
1134 | static void | |
4a8fb1a1 LC |
1135 | create_call_for_reduction (struct loop *loop, |
1136 | reduction_info_table_type reduction_list, | |
a509ebb5 RL |
1137 | struct clsn_data *ld_st_data) |
1138 | { | |
4a8fb1a1 | 1139 | reduction_list.traverse <struct loop *, create_phi_for_local_result> (loop); |
726a989a | 1140 | /* Find the fallthru edge from GIMPLE_OMP_CONTINUE. */ |
a509ebb5 | 1141 | ld_st_data->load_bb = FALLTHRU_EDGE (loop->latch)->dest; |
4a8fb1a1 LC |
1142 | reduction_list |
1143 | .traverse <struct clsn_data *, create_call_for_reduction_1> (ld_st_data); | |
a509ebb5 RL |
1144 | } |
1145 | ||
ae0bce62 RL |
1146 | /* Callback for htab_traverse. Loads the final reduction value at the |
1147 | join point of all threads, and inserts it in the right place. */ | |
a509ebb5 | 1148 | |
4a8fb1a1 LC |
1149 | int |
1150 | create_loads_for_reductions (reduction_info **slot, struct clsn_data *clsn_data) | |
a509ebb5 | 1151 | { |
4a8fb1a1 | 1152 | struct reduction_info *const red = *slot; |
726a989a RB |
1153 | gimple stmt; |
1154 | gimple_stmt_iterator gsi; | |
1155 | tree type = TREE_TYPE (gimple_assign_lhs (red->reduc_stmt)); | |
a509ebb5 | 1156 | tree load_struct; |
ae0bce62 | 1157 | tree name; |
a509ebb5 RL |
1158 | tree x; |
1159 | ||
726a989a | 1160 | gsi = gsi_after_labels (clsn_data->load_bb); |
70f34814 | 1161 | load_struct = build_simple_mem_ref (clsn_data->load); |
a509ebb5 RL |
1162 | load_struct = build3 (COMPONENT_REF, type, load_struct, red->field, |
1163 | NULL_TREE); | |
a509ebb5 | 1164 | |
ae0bce62 | 1165 | x = load_struct; |
a509ebb5 | 1166 | name = PHI_RESULT (red->keep_res); |
726a989a | 1167 | stmt = gimple_build_assign (name, x); |
a509ebb5 | 1168 | |
726a989a | 1169 | gsi_insert_after (&gsi, stmt, GSI_NEW_STMT); |
a509ebb5 | 1170 | |
726a989a RB |
1171 | for (gsi = gsi_start_phis (gimple_bb (red->keep_res)); |
1172 | !gsi_end_p (gsi); gsi_next (&gsi)) | |
1173 | if (gsi_stmt (gsi) == red->keep_res) | |
1174 | { | |
1175 | remove_phi_node (&gsi, false); | |
1176 | return 1; | |
1177 | } | |
1178 | gcc_unreachable (); | |
a509ebb5 RL |
1179 | } |
1180 | ||
b8698a0f | 1181 | /* Load the reduction result that was stored in LD_ST_DATA. |
a509ebb5 | 1182 | REDUCTION_LIST describes the list of reductions that the |
fa10beec | 1183 | loads should be generated for. */ |
a509ebb5 | 1184 | static void |
4a8fb1a1 | 1185 | create_final_loads_for_reduction (reduction_info_table_type reduction_list, |
a509ebb5 RL |
1186 | struct clsn_data *ld_st_data) |
1187 | { | |
726a989a | 1188 | gimple_stmt_iterator gsi; |
a509ebb5 | 1189 | tree t; |
726a989a | 1190 | gimple stmt; |
a509ebb5 | 1191 | |
726a989a | 1192 | gsi = gsi_after_labels (ld_st_data->load_bb); |
a509ebb5 | 1193 | t = build_fold_addr_expr (ld_st_data->store); |
726a989a | 1194 | stmt = gimple_build_assign (ld_st_data->load, t); |
a509ebb5 | 1195 | |
726a989a | 1196 | gsi_insert_before (&gsi, stmt, GSI_NEW_STMT); |
a509ebb5 | 1197 | |
4a8fb1a1 LC |
1198 | reduction_list |
1199 | .traverse <struct clsn_data *, create_loads_for_reductions> (ld_st_data); | |
a509ebb5 RL |
1200 | |
1201 | } | |
1202 | ||
0eb7e7aa RL |
1203 | /* Callback for htab_traverse. Store the neutral value for the |
1204 | particular reduction's operation, e.g. 0 for PLUS_EXPR, | |
1205 | 1 for MULT_EXPR, etc. into the reduction field. | |
b8698a0f L |
1206 | The reduction is specified in SLOT. The store information is |
1207 | passed in DATA. */ | |
0eb7e7aa | 1208 | |
4a8fb1a1 LC |
1209 | int |
1210 | create_stores_for_reduction (reduction_info **slot, struct clsn_data *clsn_data) | |
0eb7e7aa | 1211 | { |
4a8fb1a1 | 1212 | struct reduction_info *const red = *slot; |
726a989a RB |
1213 | tree t; |
1214 | gimple stmt; | |
1215 | gimple_stmt_iterator gsi; | |
1216 | tree type = TREE_TYPE (gimple_assign_lhs (red->reduc_stmt)); | |
1217 | ||
1218 | gsi = gsi_last_bb (clsn_data->store_bb); | |
1219 | t = build3 (COMPONENT_REF, type, clsn_data->store, red->field, NULL_TREE); | |
1220 | stmt = gimple_build_assign (t, red->initial_value); | |
726a989a | 1221 | gsi_insert_after (&gsi, stmt, GSI_NEW_STMT); |
0eb7e7aa RL |
1222 | |
1223 | return 1; | |
1224 | } | |
1225 | ||
a509ebb5 RL |
1226 | /* Callback for htab_traverse. Creates loads to a field of LOAD in LOAD_BB and |
1227 | store to a field of STORE in STORE_BB for the ssa name and its duplicate | |
1228 | specified in SLOT. */ | |
1229 | ||
4a8fb1a1 LC |
1230 | int |
1231 | create_loads_and_stores_for_name (name_to_copy_elt **slot, | |
1232 | struct clsn_data *clsn_data) | |
5f40b3cb | 1233 | { |
4a8fb1a1 | 1234 | struct name_to_copy_elt *const elt = *slot; |
726a989a RB |
1235 | tree t; |
1236 | gimple stmt; | |
1237 | gimple_stmt_iterator gsi; | |
5f40b3cb | 1238 | tree type = TREE_TYPE (elt->new_name); |
5f40b3cb ZD |
1239 | tree load_struct; |
1240 | ||
726a989a RB |
1241 | gsi = gsi_last_bb (clsn_data->store_bb); |
1242 | t = build3 (COMPONENT_REF, type, clsn_data->store, elt->field, NULL_TREE); | |
1243 | stmt = gimple_build_assign (t, ssa_name (elt->version)); | |
726a989a | 1244 | gsi_insert_after (&gsi, stmt, GSI_NEW_STMT); |
5f40b3cb | 1245 | |
726a989a | 1246 | gsi = gsi_last_bb (clsn_data->load_bb); |
70f34814 | 1247 | load_struct = build_simple_mem_ref (clsn_data->load); |
726a989a RB |
1248 | t = build3 (COMPONENT_REF, type, load_struct, elt->field, NULL_TREE); |
1249 | stmt = gimple_build_assign (elt->new_name, t); | |
726a989a | 1250 | gsi_insert_after (&gsi, stmt, GSI_NEW_STMT); |
5f40b3cb ZD |
1251 | |
1252 | return 1; | |
1253 | } | |
1254 | ||
1255 | /* Moves all the variables used in LOOP and defined outside of it (including | |
1256 | the initial values of loop phi nodes, and *PER_THREAD if it is a ssa | |
1257 | name) to a structure created for this purpose. The code | |
b8698a0f | 1258 | |
5f40b3cb ZD |
1259 | while (1) |
1260 | { | |
1261 | use (a); | |
1262 | use (b); | |
1263 | } | |
1264 | ||
1265 | is transformed this way: | |
1266 | ||
1267 | bb0: | |
1268 | old.a = a; | |
1269 | old.b = b; | |
1270 | ||
1271 | bb1: | |
1272 | a' = new->a; | |
1273 | b' = new->b; | |
1274 | while (1) | |
1275 | { | |
1276 | use (a'); | |
1277 | use (b'); | |
1278 | } | |
1279 | ||
1280 | `old' is stored to *ARG_STRUCT and `new' is stored to NEW_ARG_STRUCT. The | |
1281 | pointer `new' is intentionally not initialized (the loop will be split to a | |
1282 | separate function later, and `new' will be initialized from its arguments). | |
a509ebb5 | 1283 | LD_ST_DATA holds information about the shared data structure used to pass |
b8698a0f L |
1284 | information among the threads. It is initialized here, and |
1285 | gen_parallel_loop will pass it to create_call_for_reduction that | |
1286 | needs this information. REDUCTION_LIST describes the reductions | |
a509ebb5 | 1287 | in LOOP. */ |
5f40b3cb ZD |
1288 | |
1289 | static void | |
4a8fb1a1 LC |
1290 | separate_decls_in_region (edge entry, edge exit, |
1291 | reduction_info_table_type reduction_list, | |
b8698a0f | 1292 | tree *arg_struct, tree *new_arg_struct, |
9f9f72aa | 1293 | struct clsn_data *ld_st_data) |
a509ebb5 | 1294 | |
5f40b3cb | 1295 | { |
9f9f72aa | 1296 | basic_block bb1 = split_edge (entry); |
5f40b3cb | 1297 | basic_block bb0 = single_pred (bb1); |
4a8fb1a1 LC |
1298 | name_to_copy_table_type name_copies; |
1299 | name_copies.create (10); | |
1300 | int_tree_htab_type decl_copies; | |
1301 | decl_copies.create (10); | |
5f40b3cb | 1302 | unsigned i; |
726a989a RB |
1303 | tree type, type_name, nvar; |
1304 | gimple_stmt_iterator gsi; | |
5f40b3cb | 1305 | struct clsn_data clsn_data; |
00f96dc9 | 1306 | auto_vec<basic_block, 3> body; |
9f9f72aa AP |
1307 | basic_block bb; |
1308 | basic_block entry_bb = bb1; | |
1309 | basic_block exit_bb = exit->dest; | |
b5b8b0ac | 1310 | bool has_debug_stmt = false; |
5f40b3cb | 1311 | |
726a989a | 1312 | entry = single_succ_edge (entry_bb); |
9f9f72aa | 1313 | gather_blocks_in_sese_region (entry_bb, exit_bb, &body); |
5f40b3cb | 1314 | |
9771b263 | 1315 | FOR_EACH_VEC_ELT (body, i, bb) |
9f9f72aa | 1316 | { |
b8698a0f | 1317 | if (bb != entry_bb && bb != exit_bb) |
9f9f72aa | 1318 | { |
726a989a RB |
1319 | for (gsi = gsi_start_phis (bb); !gsi_end_p (gsi); gsi_next (&gsi)) |
1320 | separate_decls_in_region_stmt (entry, exit, gsi_stmt (gsi), | |
1321 | name_copies, decl_copies); | |
1322 | ||
1323 | for (gsi = gsi_start_bb (bb); !gsi_end_p (gsi); gsi_next (&gsi)) | |
b5b8b0ac AO |
1324 | { |
1325 | gimple stmt = gsi_stmt (gsi); | |
1326 | ||
1327 | if (is_gimple_debug (stmt)) | |
1328 | has_debug_stmt = true; | |
1329 | else | |
1330 | separate_decls_in_region_stmt (entry, exit, stmt, | |
1331 | name_copies, decl_copies); | |
1332 | } | |
9f9f72aa | 1333 | } |
5f40b3cb | 1334 | } |
9f9f72aa | 1335 | |
b5b8b0ac AO |
1336 | /* Now process debug bind stmts. We must not create decls while |
1337 | processing debug stmts, so we defer their processing so as to | |
1338 | make sure we will have debug info for as many variables as | |
1339 | possible (all of those that were dealt with in the loop above), | |
1340 | and discard those for which we know there's nothing we can | |
1341 | do. */ | |
1342 | if (has_debug_stmt) | |
9771b263 | 1343 | FOR_EACH_VEC_ELT (body, i, bb) |
b5b8b0ac AO |
1344 | if (bb != entry_bb && bb != exit_bb) |
1345 | { | |
1346 | for (gsi = gsi_start_bb (bb); !gsi_end_p (gsi);) | |
1347 | { | |
1348 | gimple stmt = gsi_stmt (gsi); | |
1349 | ||
ddb555ed | 1350 | if (is_gimple_debug (stmt)) |
b5b8b0ac | 1351 | { |
ddb555ed JJ |
1352 | if (separate_decls_in_region_debug (stmt, name_copies, |
1353 | decl_copies)) | |
b5b8b0ac AO |
1354 | { |
1355 | gsi_remove (&gsi, true); | |
1356 | continue; | |
1357 | } | |
1358 | } | |
1359 | ||
1360 | gsi_next (&gsi); | |
1361 | } | |
1362 | } | |
1363 | ||
4a8fb1a1 | 1364 | if (name_copies.elements () == 0 && reduction_list.elements () == 0) |
5f40b3cb ZD |
1365 | { |
1366 | /* It may happen that there is nothing to copy (if there are only | |
a509ebb5 | 1367 | loop carried and external variables in the loop). */ |
5f40b3cb ZD |
1368 | *arg_struct = NULL; |
1369 | *new_arg_struct = NULL; | |
1370 | } | |
1371 | else | |
1372 | { | |
1373 | /* Create the type for the structure to store the ssa names to. */ | |
1374 | type = lang_hooks.types.make_type (RECORD_TYPE); | |
9ff70652 | 1375 | type_name = build_decl (UNKNOWN_LOCATION, |
c2255bc4 | 1376 | TYPE_DECL, create_tmp_var_name (".paral_data"), |
5f40b3cb ZD |
1377 | type); |
1378 | TYPE_NAME (type) = type_name; | |
1379 | ||
4a8fb1a1 LC |
1380 | name_copies.traverse <tree, add_field_for_name> (type); |
1381 | if (reduction_list.is_created () && reduction_list.elements () > 0) | |
0eb7e7aa RL |
1382 | { |
1383 | /* Create the fields for reductions. */ | |
4a8fb1a1 | 1384 | reduction_list.traverse <tree, add_field_for_reduction> (type); |
0eb7e7aa | 1385 | } |
5f40b3cb | 1386 | layout_type (type); |
b8698a0f | 1387 | |
5f40b3cb ZD |
1388 | /* Create the loads and stores. */ |
1389 | *arg_struct = create_tmp_var (type, ".paral_data_store"); | |
5f40b3cb | 1390 | nvar = create_tmp_var (build_pointer_type (type), ".paral_data_load"); |
726a989a | 1391 | *new_arg_struct = make_ssa_name (nvar, NULL); |
5f40b3cb | 1392 | |
a509ebb5 RL |
1393 | ld_st_data->store = *arg_struct; |
1394 | ld_st_data->load = *new_arg_struct; | |
1395 | ld_st_data->store_bb = bb0; | |
1396 | ld_st_data->load_bb = bb1; | |
0eb7e7aa | 1397 | |
4a8fb1a1 LC |
1398 | name_copies |
1399 | .traverse <struct clsn_data *, create_loads_and_stores_for_name> | |
1400 | (ld_st_data); | |
a509ebb5 | 1401 | |
ae0bce62 RL |
1402 | /* Load the calculation from memory (after the join of the threads). */ |
1403 | ||
4a8fb1a1 | 1404 | if (reduction_list.is_created () && reduction_list.elements () > 0) |
a509ebb5 | 1405 | { |
4a8fb1a1 LC |
1406 | reduction_list |
1407 | .traverse <struct clsn_data *, create_stores_for_reduction> | |
1408 | (ld_st_data); | |
726a989a | 1409 | clsn_data.load = make_ssa_name (nvar, NULL); |
9f9f72aa | 1410 | clsn_data.load_bb = exit->dest; |
a509ebb5 RL |
1411 | clsn_data.store = ld_st_data->store; |
1412 | create_final_loads_for_reduction (reduction_list, &clsn_data); | |
1413 | } | |
5f40b3cb ZD |
1414 | } |
1415 | ||
4a8fb1a1 LC |
1416 | decl_copies.dispose (); |
1417 | name_copies.dispose (); | |
5f40b3cb ZD |
1418 | } |
1419 | ||
1420 | /* Bitmap containing uids of functions created by parallelization. We cannot | |
1421 | allocate it from the default obstack, as it must live across compilation | |
1422 | of several functions; we make it gc allocated instead. */ | |
1423 | ||
1424 | static GTY(()) bitmap parallelized_functions; | |
1425 | ||
1426 | /* Returns true if FN was created by create_loop_fn. */ | |
1427 | ||
62e0a1ed | 1428 | bool |
5f40b3cb ZD |
1429 | parallelized_function_p (tree fn) |
1430 | { | |
1431 | if (!parallelized_functions || !DECL_ARTIFICIAL (fn)) | |
1432 | return false; | |
1433 | ||
1434 | return bitmap_bit_p (parallelized_functions, DECL_UID (fn)); | |
1435 | } | |
1436 | ||
1437 | /* Creates and returns an empty function that will receive the body of | |
1438 | a parallelized loop. */ | |
1439 | ||
1440 | static tree | |
9ff70652 | 1441 | create_loop_fn (location_t loc) |
5f40b3cb ZD |
1442 | { |
1443 | char buf[100]; | |
1444 | char *tname; | |
1445 | tree decl, type, name, t; | |
1446 | struct function *act_cfun = cfun; | |
1447 | static unsigned loopfn_num; | |
1448 | ||
5368224f | 1449 | loc = LOCATION_LOCUS (loc); |
5f40b3cb ZD |
1450 | snprintf (buf, 100, "%s.$loopfn", current_function_name ()); |
1451 | ASM_FORMAT_PRIVATE_NAME (tname, buf, loopfn_num++); | |
1452 | clean_symbol_name (tname); | |
1453 | name = get_identifier (tname); | |
1454 | type = build_function_type_list (void_type_node, ptr_type_node, NULL_TREE); | |
1455 | ||
9ff70652 | 1456 | decl = build_decl (loc, FUNCTION_DECL, name, type); |
5f40b3cb ZD |
1457 | if (!parallelized_functions) |
1458 | parallelized_functions = BITMAP_GGC_ALLOC (); | |
1459 | bitmap_set_bit (parallelized_functions, DECL_UID (decl)); | |
1460 | ||
1461 | TREE_STATIC (decl) = 1; | |
1462 | TREE_USED (decl) = 1; | |
1463 | DECL_ARTIFICIAL (decl) = 1; | |
1464 | DECL_IGNORED_P (decl) = 0; | |
1465 | TREE_PUBLIC (decl) = 0; | |
1466 | DECL_UNINLINABLE (decl) = 1; | |
1467 | DECL_EXTERNAL (decl) = 0; | |
1468 | DECL_CONTEXT (decl) = NULL_TREE; | |
1469 | DECL_INITIAL (decl) = make_node (BLOCK); | |
1470 | ||
9ff70652 | 1471 | t = build_decl (loc, RESULT_DECL, NULL_TREE, void_type_node); |
5f40b3cb ZD |
1472 | DECL_ARTIFICIAL (t) = 1; |
1473 | DECL_IGNORED_P (t) = 1; | |
1474 | DECL_RESULT (decl) = t; | |
1475 | ||
9ff70652 | 1476 | t = build_decl (loc, PARM_DECL, get_identifier (".paral_data_param"), |
5f40b3cb ZD |
1477 | ptr_type_node); |
1478 | DECL_ARTIFICIAL (t) = 1; | |
1479 | DECL_ARG_TYPE (t) = ptr_type_node; | |
1480 | DECL_CONTEXT (t) = decl; | |
1481 | TREE_USED (t) = 1; | |
1482 | DECL_ARGUMENTS (decl) = t; | |
1483 | ||
182e0d71 | 1484 | allocate_struct_function (decl, false); |
5f40b3cb ZD |
1485 | |
1486 | /* The call to allocate_struct_function clobbers CFUN, so we need to restore | |
1487 | it. */ | |
5576d6f2 | 1488 | set_cfun (act_cfun); |
5f40b3cb ZD |
1489 | |
1490 | return decl; | |
1491 | } | |
1492 | ||
5f40b3cb ZD |
1493 | /* Moves the exit condition of LOOP to the beginning of its header, and |
1494 | duplicates the part of the last iteration that gets disabled to the | |
1495 | exit of the loop. NIT is the number of iterations of the loop | |
1496 | (used to initialize the variables in the duplicated part). | |
b8698a0f | 1497 | |
fa10beec | 1498 | TODO: the common case is that latch of the loop is empty and immediately |
5f40b3cb ZD |
1499 | follows the loop exit. In this case, it would be better not to copy the |
1500 | body of the loop, but only move the entry of the loop directly before the | |
1501 | exit check and increase the number of iterations of the loop by one. | |
b8698a0f | 1502 | This may need some additional preconditioning in case NIT = ~0. |
a509ebb5 | 1503 | REDUCTION_LIST describes the reductions in LOOP. */ |
5f40b3cb ZD |
1504 | |
1505 | static void | |
4a8fb1a1 LC |
1506 | transform_to_exit_first_loop (struct loop *loop, |
1507 | reduction_info_table_type reduction_list, | |
1508 | tree nit) | |
5f40b3cb ZD |
1509 | { |
1510 | basic_block *bbs, *nbbs, ex_bb, orig_header; | |
1511 | unsigned n; | |
1512 | bool ok; | |
1513 | edge exit = single_dom_exit (loop), hpred; | |
726a989a | 1514 | tree control, control_name, res, t; |
48710229 | 1515 | gimple phi, nphi, cond_stmt, stmt, cond_nit; |
726a989a | 1516 | gimple_stmt_iterator gsi; |
48710229 | 1517 | tree nit_1; |
5f40b3cb ZD |
1518 | |
1519 | split_block_after_labels (loop->header); | |
1520 | orig_header = single_succ (loop->header); | |
1521 | hpred = single_succ_edge (loop->header); | |
1522 | ||
1523 | cond_stmt = last_stmt (exit->src); | |
726a989a RB |
1524 | control = gimple_cond_lhs (cond_stmt); |
1525 | gcc_assert (gimple_cond_rhs (cond_stmt) == nit); | |
5f40b3cb ZD |
1526 | |
1527 | /* Make sure that we have phi nodes on exit for all loop header phis | |
1528 | (create_parallel_loop requires that). */ | |
726a989a | 1529 | for (gsi = gsi_start_phis (loop->header); !gsi_end_p (gsi); gsi_next (&gsi)) |
5f40b3cb | 1530 | { |
726a989a | 1531 | phi = gsi_stmt (gsi); |
5f40b3cb | 1532 | res = PHI_RESULT (phi); |
070ecdfd | 1533 | t = copy_ssa_name (res, phi); |
5f40b3cb | 1534 | SET_PHI_RESULT (phi, t); |
5f40b3cb | 1535 | nphi = create_phi_node (res, orig_header); |
9e227d60 | 1536 | add_phi_arg (nphi, t, hpred, UNKNOWN_LOCATION); |
5f40b3cb ZD |
1537 | |
1538 | if (res == control) | |
1539 | { | |
726a989a | 1540 | gimple_cond_set_lhs (cond_stmt, t); |
5f40b3cb ZD |
1541 | update_stmt (cond_stmt); |
1542 | control = t; | |
1543 | } | |
1544 | } | |
12037899 | 1545 | |
5f40b3cb | 1546 | bbs = get_loop_body_in_dom_order (loop); |
48710229 | 1547 | |
69958396 RL |
1548 | for (n = 0; bbs[n] != exit->src; n++) |
1549 | continue; | |
5f40b3cb | 1550 | nbbs = XNEWVEC (basic_block, n); |
726a989a RB |
1551 | ok = gimple_duplicate_sese_tail (single_succ_edge (loop->header), exit, |
1552 | bbs + 1, n, nbbs); | |
5f40b3cb ZD |
1553 | gcc_assert (ok); |
1554 | free (bbs); | |
1555 | ex_bb = nbbs[0]; | |
1556 | free (nbbs); | |
1557 | ||
b8698a0f | 1558 | /* Other than reductions, the only gimple reg that should be copied |
726a989a | 1559 | out of the loop is the control variable. */ |
69958396 | 1560 | exit = single_dom_exit (loop); |
5f40b3cb | 1561 | control_name = NULL_TREE; |
726a989a | 1562 | for (gsi = gsi_start_phis (ex_bb); !gsi_end_p (gsi); ) |
5f40b3cb | 1563 | { |
726a989a | 1564 | phi = gsi_stmt (gsi); |
5f40b3cb | 1565 | res = PHI_RESULT (phi); |
ea057359 | 1566 | if (virtual_operand_p (res)) |
726a989a RB |
1567 | { |
1568 | gsi_next (&gsi); | |
1569 | continue; | |
1570 | } | |
5f40b3cb | 1571 | |
a509ebb5 | 1572 | /* Check if it is a part of reduction. If it is, |
b8698a0f L |
1573 | keep the phi at the reduction's keep_res field. The |
1574 | PHI_RESULT of this phi is the resulting value of the reduction | |
a509ebb5 RL |
1575 | variable when exiting the loop. */ |
1576 | ||
4a8fb1a1 | 1577 | if (reduction_list.elements () > 0) |
a509ebb5 RL |
1578 | { |
1579 | struct reduction_info *red; | |
1580 | ||
1581 | tree val = PHI_ARG_DEF_FROM_EDGE (phi, exit); | |
a509ebb5 RL |
1582 | red = reduction_phi (reduction_list, SSA_NAME_DEF_STMT (val)); |
1583 | if (red) | |
726a989a RB |
1584 | { |
1585 | red->keep_res = phi; | |
1586 | gsi_next (&gsi); | |
1587 | continue; | |
1588 | } | |
a509ebb5 | 1589 | } |
726a989a RB |
1590 | gcc_assert (control_name == NULL_TREE |
1591 | && SSA_NAME_VAR (res) == SSA_NAME_VAR (control)); | |
5f40b3cb | 1592 | control_name = res; |
726a989a | 1593 | remove_phi_node (&gsi, false); |
5f40b3cb ZD |
1594 | } |
1595 | gcc_assert (control_name != NULL_TREE); | |
5f40b3cb | 1596 | |
b8698a0f | 1597 | /* Initialize the control variable to number of iterations |
48710229 | 1598 | according to the rhs of the exit condition. */ |
726a989a | 1599 | gsi = gsi_after_labels (ex_bb); |
b8698a0f | 1600 | cond_nit = last_stmt (exit->src); |
48710229 RL |
1601 | nit_1 = gimple_cond_rhs (cond_nit); |
1602 | nit_1 = force_gimple_operand_gsi (&gsi, | |
1603 | fold_convert (TREE_TYPE (control_name), nit_1), | |
726a989a | 1604 | false, NULL_TREE, false, GSI_SAME_STMT); |
48710229 | 1605 | stmt = gimple_build_assign (control_name, nit_1); |
726a989a | 1606 | gsi_insert_before (&gsi, stmt, GSI_NEW_STMT); |
5f40b3cb ZD |
1607 | } |
1608 | ||
1609 | /* Create the parallel constructs for LOOP as described in gen_parallel_loop. | |
726a989a | 1610 | LOOP_FN and DATA are the arguments of GIMPLE_OMP_PARALLEL. |
5f40b3cb ZD |
1611 | NEW_DATA is the variable that should be initialized from the argument |
1612 | of LOOP_FN. N_THREADS is the requested number of threads. Returns the | |
726a989a | 1613 | basic block containing GIMPLE_OMP_PARALLEL tree. */ |
5f40b3cb ZD |
1614 | |
1615 | static basic_block | |
1616 | create_parallel_loop (struct loop *loop, tree loop_fn, tree data, | |
9ff70652 | 1617 | tree new_data, unsigned n_threads, location_t loc) |
5f40b3cb | 1618 | { |
726a989a | 1619 | gimple_stmt_iterator gsi; |
5f40b3cb | 1620 | basic_block bb, paral_bb, for_bb, ex_bb; |
0f900dfa | 1621 | tree t, param; |
726a989a RB |
1622 | gimple stmt, for_stmt, phi, cond_stmt; |
1623 | tree cvar, cvar_init, initvar, cvar_next, cvar_base, type; | |
5f40b3cb ZD |
1624 | edge exit, nexit, guard, end, e; |
1625 | ||
726a989a | 1626 | /* Prepare the GIMPLE_OMP_PARALLEL statement. */ |
5f40b3cb ZD |
1627 | bb = loop_preheader_edge (loop)->src; |
1628 | paral_bb = single_pred (bb); | |
726a989a | 1629 | gsi = gsi_last_bb (paral_bb); |
5f40b3cb | 1630 | |
9ff70652 | 1631 | t = build_omp_clause (loc, OMP_CLAUSE_NUM_THREADS); |
5f40b3cb | 1632 | OMP_CLAUSE_NUM_THREADS_EXPR (t) |
a509ebb5 | 1633 | = build_int_cst (integer_type_node, n_threads); |
726a989a | 1634 | stmt = gimple_build_omp_parallel (NULL, t, loop_fn, data); |
9ff70652 | 1635 | gimple_set_location (stmt, loc); |
5f40b3cb | 1636 | |
726a989a | 1637 | gsi_insert_after (&gsi, stmt, GSI_NEW_STMT); |
5f40b3cb ZD |
1638 | |
1639 | /* Initialize NEW_DATA. */ | |
1640 | if (data) | |
1641 | { | |
726a989a RB |
1642 | gsi = gsi_after_labels (bb); |
1643 | ||
1644 | param = make_ssa_name (DECL_ARGUMENTS (loop_fn), NULL); | |
1645 | stmt = gimple_build_assign (param, build_fold_addr_expr (data)); | |
1646 | gsi_insert_before (&gsi, stmt, GSI_SAME_STMT); | |
726a989a RB |
1647 | |
1648 | stmt = gimple_build_assign (new_data, | |
1649 | fold_convert (TREE_TYPE (new_data), param)); | |
1650 | gsi_insert_before (&gsi, stmt, GSI_SAME_STMT); | |
5f40b3cb ZD |
1651 | } |
1652 | ||
726a989a | 1653 | /* Emit GIMPLE_OMP_RETURN for GIMPLE_OMP_PARALLEL. */ |
5f40b3cb | 1654 | bb = split_loop_exit_edge (single_dom_exit (loop)); |
726a989a | 1655 | gsi = gsi_last_bb (bb); |
9ff70652 JJ |
1656 | stmt = gimple_build_omp_return (false); |
1657 | gimple_set_location (stmt, loc); | |
1658 | gsi_insert_after (&gsi, stmt, GSI_NEW_STMT); | |
5f40b3cb | 1659 | |
726a989a | 1660 | /* Extract data for GIMPLE_OMP_FOR. */ |
5f40b3cb | 1661 | gcc_assert (loop->header == single_dom_exit (loop)->src); |
726a989a | 1662 | cond_stmt = last_stmt (loop->header); |
5f40b3cb | 1663 | |
726a989a | 1664 | cvar = gimple_cond_lhs (cond_stmt); |
5f40b3cb ZD |
1665 | cvar_base = SSA_NAME_VAR (cvar); |
1666 | phi = SSA_NAME_DEF_STMT (cvar); | |
1667 | cvar_init = PHI_ARG_DEF_FROM_EDGE (phi, loop_preheader_edge (loop)); | |
070ecdfd | 1668 | initvar = copy_ssa_name (cvar, NULL); |
5f40b3cb ZD |
1669 | SET_USE (PHI_ARG_DEF_PTR_FROM_EDGE (phi, loop_preheader_edge (loop)), |
1670 | initvar); | |
1671 | cvar_next = PHI_ARG_DEF_FROM_EDGE (phi, loop_latch_edge (loop)); | |
1672 | ||
1dff453d | 1673 | gsi = gsi_last_nondebug_bb (loop->latch); |
726a989a RB |
1674 | gcc_assert (gsi_stmt (gsi) == SSA_NAME_DEF_STMT (cvar_next)); |
1675 | gsi_remove (&gsi, true); | |
5f40b3cb ZD |
1676 | |
1677 | /* Prepare cfg. */ | |
1678 | for_bb = split_edge (loop_preheader_edge (loop)); | |
1679 | ex_bb = split_loop_exit_edge (single_dom_exit (loop)); | |
1680 | extract_true_false_edges_from_block (loop->header, &nexit, &exit); | |
1681 | gcc_assert (exit == single_dom_exit (loop)); | |
1682 | ||
1683 | guard = make_edge (for_bb, ex_bb, 0); | |
1684 | single_succ_edge (loop->latch)->flags = 0; | |
1685 | end = make_edge (loop->latch, ex_bb, EDGE_FALLTHRU); | |
726a989a | 1686 | for (gsi = gsi_start_phis (ex_bb); !gsi_end_p (gsi); gsi_next (&gsi)) |
5f40b3cb | 1687 | { |
f5045c96 AM |
1688 | source_location locus; |
1689 | tree def; | |
726a989a | 1690 | phi = gsi_stmt (gsi); |
726a989a | 1691 | stmt = SSA_NAME_DEF_STMT (PHI_ARG_DEF_FROM_EDGE (phi, exit)); |
f5045c96 AM |
1692 | |
1693 | def = PHI_ARG_DEF_FROM_EDGE (stmt, loop_preheader_edge (loop)); | |
b8698a0f | 1694 | locus = gimple_phi_arg_location_from_edge (stmt, |
f5045c96 | 1695 | loop_preheader_edge (loop)); |
9e227d60 | 1696 | add_phi_arg (phi, def, guard, locus); |
f5045c96 AM |
1697 | |
1698 | def = PHI_ARG_DEF_FROM_EDGE (stmt, loop_latch_edge (loop)); | |
1699 | locus = gimple_phi_arg_location_from_edge (stmt, loop_latch_edge (loop)); | |
9e227d60 | 1700 | add_phi_arg (phi, def, end, locus); |
5f40b3cb ZD |
1701 | } |
1702 | e = redirect_edge_and_branch (exit, nexit->dest); | |
1703 | PENDING_STMT (e) = NULL; | |
1704 | ||
726a989a RB |
1705 | /* Emit GIMPLE_OMP_FOR. */ |
1706 | gimple_cond_set_lhs (cond_stmt, cvar_base); | |
5f40b3cb | 1707 | type = TREE_TYPE (cvar); |
9ff70652 | 1708 | t = build_omp_clause (loc, OMP_CLAUSE_SCHEDULE); |
5f40b3cb ZD |
1709 | OMP_CLAUSE_SCHEDULE_KIND (t) = OMP_CLAUSE_SCHEDULE_STATIC; |
1710 | ||
74bf76ed | 1711 | for_stmt = gimple_build_omp_for (NULL, GF_OMP_FOR_KIND_FOR, t, 1, NULL); |
9ff70652 | 1712 | gimple_set_location (for_stmt, loc); |
726a989a RB |
1713 | gimple_omp_for_set_index (for_stmt, 0, initvar); |
1714 | gimple_omp_for_set_initial (for_stmt, 0, cvar_init); | |
1715 | gimple_omp_for_set_final (for_stmt, 0, gimple_cond_rhs (cond_stmt)); | |
1716 | gimple_omp_for_set_cond (for_stmt, 0, gimple_cond_code (cond_stmt)); | |
1717 | gimple_omp_for_set_incr (for_stmt, 0, build2 (PLUS_EXPR, type, | |
1718 | cvar_base, | |
1719 | build_int_cst (type, 1))); | |
1720 | ||
1721 | gsi = gsi_last_bb (for_bb); | |
1722 | gsi_insert_after (&gsi, for_stmt, GSI_NEW_STMT); | |
5f40b3cb ZD |
1723 | SSA_NAME_DEF_STMT (initvar) = for_stmt; |
1724 | ||
726a989a RB |
1725 | /* Emit GIMPLE_OMP_CONTINUE. */ |
1726 | gsi = gsi_last_bb (loop->latch); | |
1727 | stmt = gimple_build_omp_continue (cvar_next, cvar); | |
9ff70652 | 1728 | gimple_set_location (stmt, loc); |
726a989a RB |
1729 | gsi_insert_after (&gsi, stmt, GSI_NEW_STMT); |
1730 | SSA_NAME_DEF_STMT (cvar_next) = stmt; | |
5f40b3cb | 1731 | |
726a989a RB |
1732 | /* Emit GIMPLE_OMP_RETURN for GIMPLE_OMP_FOR. */ |
1733 | gsi = gsi_last_bb (ex_bb); | |
9ff70652 JJ |
1734 | stmt = gimple_build_omp_return (true); |
1735 | gimple_set_location (stmt, loc); | |
1736 | gsi_insert_after (&gsi, stmt, GSI_NEW_STMT); | |
5f40b3cb | 1737 | |
cd7d9fd7 RG |
1738 | /* After the above dom info is hosed. Re-compute it. */ |
1739 | free_dominance_info (CDI_DOMINATORS); | |
1740 | calculate_dominance_info (CDI_DOMINATORS); | |
1741 | ||
5f40b3cb ZD |
1742 | return paral_bb; |
1743 | } | |
1744 | ||
08dab97a RL |
1745 | /* Generates code to execute the iterations of LOOP in N_THREADS |
1746 | threads in parallel. | |
1747 | ||
1748 | NITER describes number of iterations of LOOP. | |
fa10beec | 1749 | REDUCTION_LIST describes the reductions existent in the LOOP. */ |
5f40b3cb ZD |
1750 | |
1751 | static void | |
4a8fb1a1 | 1752 | gen_parallel_loop (struct loop *loop, reduction_info_table_type reduction_list, |
a509ebb5 | 1753 | unsigned n_threads, struct tree_niter_desc *niter) |
5f40b3cb | 1754 | { |
5f40b3cb | 1755 | tree many_iterations_cond, type, nit; |
726a989a RB |
1756 | tree arg_struct, new_arg_struct; |
1757 | gimple_seq stmts; | |
5f40b3cb | 1758 | basic_block parallel_head; |
9f9f72aa | 1759 | edge entry, exit; |
a509ebb5 | 1760 | struct clsn_data clsn_data; |
5f40b3cb | 1761 | unsigned prob; |
9ff70652 JJ |
1762 | location_t loc; |
1763 | gimple cond_stmt; | |
768da0da | 1764 | unsigned int m_p_thread=2; |
5f40b3cb ZD |
1765 | |
1766 | /* From | |
1767 | ||
1768 | --------------------------------------------------------------------- | |
1769 | loop | |
1770 | { | |
1771 | IV = phi (INIT, IV + STEP) | |
1772 | BODY1; | |
1773 | if (COND) | |
1774 | break; | |
1775 | BODY2; | |
1776 | } | |
1777 | --------------------------------------------------------------------- | |
1778 | ||
1779 | with # of iterations NITER (possibly with MAY_BE_ZERO assumption), | |
1780 | we generate the following code: | |
1781 | ||
1782 | --------------------------------------------------------------------- | |
1783 | ||
1784 | if (MAY_BE_ZERO | |
a509ebb5 RL |
1785 | || NITER < MIN_PER_THREAD * N_THREADS) |
1786 | goto original; | |
5f40b3cb ZD |
1787 | |
1788 | BODY1; | |
1789 | store all local loop-invariant variables used in body of the loop to DATA. | |
726a989a | 1790 | GIMPLE_OMP_PARALLEL (OMP_CLAUSE_NUM_THREADS (N_THREADS), LOOPFN, DATA); |
5f40b3cb | 1791 | load the variables from DATA. |
726a989a | 1792 | GIMPLE_OMP_FOR (IV = INIT; COND; IV += STEP) (OMP_CLAUSE_SCHEDULE (static)) |
5f40b3cb ZD |
1793 | BODY2; |
1794 | BODY1; | |
726a989a RB |
1795 | GIMPLE_OMP_CONTINUE; |
1796 | GIMPLE_OMP_RETURN -- GIMPLE_OMP_FOR | |
1797 | GIMPLE_OMP_RETURN -- GIMPLE_OMP_PARALLEL | |
5f40b3cb ZD |
1798 | goto end; |
1799 | ||
1800 | original: | |
1801 | loop | |
1802 | { | |
1803 | IV = phi (INIT, IV + STEP) | |
1804 | BODY1; | |
1805 | if (COND) | |
1806 | break; | |
1807 | BODY2; | |
1808 | } | |
1809 | ||
1810 | end: | |
1811 | ||
1812 | */ | |
1813 | ||
1814 | /* Create two versions of the loop -- in the old one, we know that the | |
1815 | number of iterations is large enough, and we will transform it into the | |
1816 | loop that will be split to loop_fn, the new one will be used for the | |
1817 | remaining iterations. */ | |
a509ebb5 | 1818 | |
768da0da RL |
1819 | /* We should compute a better number-of-iterations value for outer loops. |
1820 | That is, if we have | |
1821 | ||
1822 | for (i = 0; i < n; ++i) | |
1823 | for (j = 0; j < m; ++j) | |
1824 | ... | |
1825 | ||
1826 | we should compute nit = n * m, not nit = n. | |
1827 | Also may_be_zero handling would need to be adjusted. */ | |
1828 | ||
5f40b3cb ZD |
1829 | type = TREE_TYPE (niter->niter); |
1830 | nit = force_gimple_operand (unshare_expr (niter->niter), &stmts, true, | |
1831 | NULL_TREE); | |
1832 | if (stmts) | |
726a989a | 1833 | gsi_insert_seq_on_edge_immediate (loop_preheader_edge (loop), stmts); |
5f40b3cb | 1834 | |
768da0da RL |
1835 | if (loop->inner) |
1836 | m_p_thread=2; | |
1837 | else | |
1838 | m_p_thread=MIN_PER_THREAD; | |
1839 | ||
1840 | many_iterations_cond = | |
1841 | fold_build2 (GE_EXPR, boolean_type_node, | |
1842 | nit, build_int_cst (type, m_p_thread * n_threads)); | |
1843 | ||
5f40b3cb | 1844 | many_iterations_cond |
a509ebb5 RL |
1845 | = fold_build2 (TRUTH_AND_EXPR, boolean_type_node, |
1846 | invert_truthvalue (unshare_expr (niter->may_be_zero)), | |
1847 | many_iterations_cond); | |
5f40b3cb | 1848 | many_iterations_cond |
a509ebb5 | 1849 | = force_gimple_operand (many_iterations_cond, &stmts, false, NULL_TREE); |
5f40b3cb | 1850 | if (stmts) |
726a989a | 1851 | gsi_insert_seq_on_edge_immediate (loop_preheader_edge (loop), stmts); |
5f40b3cb ZD |
1852 | if (!is_gimple_condexpr (many_iterations_cond)) |
1853 | { | |
1854 | many_iterations_cond | |
a509ebb5 RL |
1855 | = force_gimple_operand (many_iterations_cond, &stmts, |
1856 | true, NULL_TREE); | |
5f40b3cb | 1857 | if (stmts) |
726a989a | 1858 | gsi_insert_seq_on_edge_immediate (loop_preheader_edge (loop), stmts); |
5f40b3cb ZD |
1859 | } |
1860 | ||
1861 | initialize_original_copy_tables (); | |
1862 | ||
1863 | /* We assume that the loop usually iterates a lot. */ | |
1864 | prob = 4 * REG_BR_PROB_BASE / 5; | |
0f900dfa JJ |
1865 | loop_version (loop, many_iterations_cond, NULL, |
1866 | prob, prob, REG_BR_PROB_BASE - prob, true); | |
5f40b3cb ZD |
1867 | update_ssa (TODO_update_ssa); |
1868 | free_original_copy_tables (); | |
1869 | ||
1870 | /* Base all the induction variables in LOOP on a single control one. */ | |
c80a5403 | 1871 | canonicalize_loop_ivs (loop, &nit, true); |
5f40b3cb ZD |
1872 | |
1873 | /* Ensure that the exit condition is the first statement in the loop. */ | |
a509ebb5 RL |
1874 | transform_to_exit_first_loop (loop, reduction_list, nit); |
1875 | ||
fa10beec | 1876 | /* Generate initializations for reductions. */ |
4a8fb1a1 LC |
1877 | if (reduction_list.elements () > 0) |
1878 | reduction_list.traverse <struct loop *, initialize_reductions> (loop); | |
5f40b3cb ZD |
1879 | |
1880 | /* Eliminate the references to local variables from the loop. */ | |
9f9f72aa AP |
1881 | gcc_assert (single_exit (loop)); |
1882 | entry = loop_preheader_edge (loop); | |
1883 | exit = single_dom_exit (loop); | |
5f40b3cb | 1884 | |
9f9f72aa | 1885 | eliminate_local_variables (entry, exit); |
5f40b3cb ZD |
1886 | /* In the old loop, move all variables non-local to the loop to a structure |
1887 | and back, and create separate decls for the variables used in loop. */ | |
b8698a0f | 1888 | separate_decls_in_region (entry, exit, reduction_list, &arg_struct, |
9f9f72aa | 1889 | &new_arg_struct, &clsn_data); |
5f40b3cb ZD |
1890 | |
1891 | /* Create the parallel constructs. */ | |
9ff70652 JJ |
1892 | loc = UNKNOWN_LOCATION; |
1893 | cond_stmt = last_stmt (loop->header); | |
1894 | if (cond_stmt) | |
1895 | loc = gimple_location (cond_stmt); | |
1896 | parallel_head = create_parallel_loop (loop, create_loop_fn (loc), arg_struct, | |
1897 | new_arg_struct, n_threads, loc); | |
4a8fb1a1 | 1898 | if (reduction_list.elements () > 0) |
a509ebb5 | 1899 | create_call_for_reduction (loop, reduction_list, &clsn_data); |
5f40b3cb ZD |
1900 | |
1901 | scev_reset (); | |
1902 | ||
1903 | /* Cancel the loop (it is simpler to do it here rather than to teach the | |
1904 | expander to do it). */ | |
1905 | cancel_loop_tree (loop); | |
1906 | ||
92a6bdbd SP |
1907 | /* Free loop bound estimations that could contain references to |
1908 | removed statements. */ | |
f0bd40b1 | 1909 | FOR_EACH_LOOP (loop, 0) |
92a6bdbd SP |
1910 | free_numbers_of_iterations_estimates_loop (loop); |
1911 | ||
5f40b3cb ZD |
1912 | /* Expand the parallel constructs. We do it directly here instead of running |
1913 | a separate expand_omp pass, since it is more efficient, and less likely to | |
1914 | cause troubles with further analyses not being able to deal with the | |
1915 | OMP trees. */ | |
a509ebb5 | 1916 | |
5f40b3cb ZD |
1917 | omp_expand_local (parallel_head); |
1918 | } | |
1919 | ||
9857228c SP |
1920 | /* Returns true when LOOP contains vector phi nodes. */ |
1921 | ||
1922 | static bool | |
726a989a | 1923 | loop_has_vector_phi_nodes (struct loop *loop ATTRIBUTE_UNUSED) |
9857228c SP |
1924 | { |
1925 | unsigned i; | |
1926 | basic_block *bbs = get_loop_body_in_dom_order (loop); | |
726a989a | 1927 | gimple_stmt_iterator gsi; |
9857228c | 1928 | bool res = true; |
9857228c SP |
1929 | |
1930 | for (i = 0; i < loop->num_nodes; i++) | |
726a989a RB |
1931 | for (gsi = gsi_start_phis (bbs[i]); !gsi_end_p (gsi); gsi_next (&gsi)) |
1932 | if (TREE_CODE (TREE_TYPE (PHI_RESULT (gsi_stmt (gsi)))) == VECTOR_TYPE) | |
9857228c SP |
1933 | goto end; |
1934 | ||
1935 | res = false; | |
1936 | end: | |
1937 | free (bbs); | |
1938 | return res; | |
1939 | } | |
1940 | ||
08dab97a RL |
1941 | /* Create a reduction_info struct, initialize it with REDUC_STMT |
1942 | and PHI, insert it to the REDUCTION_LIST. */ | |
1943 | ||
1944 | static void | |
4a8fb1a1 LC |
1945 | build_new_reduction (reduction_info_table_type reduction_list, |
1946 | gimple reduc_stmt, gimple phi) | |
08dab97a | 1947 | { |
4a8fb1a1 | 1948 | reduction_info **slot; |
08dab97a RL |
1949 | struct reduction_info *new_reduction; |
1950 | ||
1951 | gcc_assert (reduc_stmt); | |
b8698a0f | 1952 | |
08dab97a RL |
1953 | if (dump_file && (dump_flags & TDF_DETAILS)) |
1954 | { | |
1955 | fprintf (dump_file, | |
1956 | "Detected reduction. reduction stmt is: \n"); | |
1957 | print_gimple_stmt (dump_file, reduc_stmt, 0, 0); | |
1958 | fprintf (dump_file, "\n"); | |
1959 | } | |
b8698a0f | 1960 | |
08dab97a | 1961 | new_reduction = XCNEW (struct reduction_info); |
b8698a0f | 1962 | |
08dab97a RL |
1963 | new_reduction->reduc_stmt = reduc_stmt; |
1964 | new_reduction->reduc_phi = phi; | |
5d1fd1de | 1965 | new_reduction->reduc_version = SSA_NAME_VERSION (gimple_phi_result (phi)); |
08dab97a | 1966 | new_reduction->reduction_code = gimple_assign_rhs_code (reduc_stmt); |
4a8fb1a1 | 1967 | slot = reduction_list.find_slot (new_reduction, INSERT); |
08dab97a RL |
1968 | *slot = new_reduction; |
1969 | } | |
1970 | ||
5d1fd1de JJ |
1971 | /* Callback for htab_traverse. Sets gimple_uid of reduc_phi stmts. */ |
1972 | ||
4a8fb1a1 LC |
1973 | int |
1974 | set_reduc_phi_uids (reduction_info **slot, void *data ATTRIBUTE_UNUSED) | |
5d1fd1de | 1975 | { |
4a8fb1a1 | 1976 | struct reduction_info *const red = *slot; |
5d1fd1de JJ |
1977 | gimple_set_uid (red->reduc_phi, red->reduc_version); |
1978 | return 1; | |
1979 | } | |
1980 | ||
08dab97a RL |
1981 | /* Detect all reductions in the LOOP, insert them into REDUCTION_LIST. */ |
1982 | ||
1983 | static void | |
4a8fb1a1 | 1984 | gather_scalar_reductions (loop_p loop, reduction_info_table_type reduction_list) |
08dab97a RL |
1985 | { |
1986 | gimple_stmt_iterator gsi; | |
1987 | loop_vec_info simple_loop_info; | |
1988 | ||
08dab97a RL |
1989 | simple_loop_info = vect_analyze_loop_form (loop); |
1990 | ||
1991 | for (gsi = gsi_start_phis (loop->header); !gsi_end_p (gsi); gsi_next (&gsi)) | |
1992 | { | |
1993 | gimple phi = gsi_stmt (gsi); | |
1994 | affine_iv iv; | |
1995 | tree res = PHI_RESULT (phi); | |
1996 | bool double_reduc; | |
1997 | ||
ea057359 | 1998 | if (virtual_operand_p (res)) |
08dab97a RL |
1999 | continue; |
2000 | ||
2001 | if (!simple_iv (loop, loop, res, &iv, true) | |
2002 | && simple_loop_info) | |
2003 | { | |
8a9ecffd MM |
2004 | gimple reduc_stmt = vect_force_simple_reduction (simple_loop_info, |
2005 | phi, true, | |
2006 | &double_reduc); | |
48710229 | 2007 | if (reduc_stmt && !double_reduc) |
08dab97a RL |
2008 | build_new_reduction (reduction_list, reduc_stmt, phi); |
2009 | } | |
2010 | } | |
5d1fd1de JJ |
2011 | destroy_loop_vec_info (simple_loop_info, true); |
2012 | ||
2013 | /* As gimple_uid is used by the vectorizer in between vect_analyze_loop_form | |
2014 | and destroy_loop_vec_info, we can set gimple_uid of reduc_phi stmts | |
2015 | only now. */ | |
4a8fb1a1 | 2016 | reduction_list.traverse <void *, set_reduc_phi_uids> (NULL); |
08dab97a RL |
2017 | } |
2018 | ||
2019 | /* Try to initialize NITER for code generation part. */ | |
2020 | ||
2021 | static bool | |
2022 | try_get_loop_niter (loop_p loop, struct tree_niter_desc *niter) | |
2023 | { | |
2024 | edge exit = single_dom_exit (loop); | |
2025 | ||
2026 | gcc_assert (exit); | |
2027 | ||
2028 | /* We need to know # of iterations, and there should be no uses of values | |
2029 | defined inside loop outside of it, unless the values are invariants of | |
2030 | the loop. */ | |
2031 | if (!number_of_iterations_exit (loop, exit, niter, false)) | |
2032 | { | |
2033 | if (dump_file && (dump_flags & TDF_DETAILS)) | |
2034 | fprintf (dump_file, " FAILED: number of iterations not known\n"); | |
2035 | return false; | |
2036 | } | |
2037 | ||
2038 | return true; | |
2039 | } | |
2040 | ||
2041 | /* Try to initialize REDUCTION_LIST for code generation part. | |
2042 | REDUCTION_LIST describes the reductions. */ | |
2043 | ||
2044 | static bool | |
4a8fb1a1 LC |
2045 | try_create_reduction_list (loop_p loop, |
2046 | reduction_info_table_type reduction_list) | |
08dab97a RL |
2047 | { |
2048 | edge exit = single_dom_exit (loop); | |
2049 | gimple_stmt_iterator gsi; | |
2050 | ||
2051 | gcc_assert (exit); | |
2052 | ||
2053 | gather_scalar_reductions (loop, reduction_list); | |
2054 | ||
b8698a0f | 2055 | |
08dab97a RL |
2056 | for (gsi = gsi_start_phis (exit->dest); !gsi_end_p (gsi); gsi_next (&gsi)) |
2057 | { | |
2058 | gimple phi = gsi_stmt (gsi); | |
2059 | struct reduction_info *red; | |
2060 | imm_use_iterator imm_iter; | |
2061 | use_operand_p use_p; | |
2062 | gimple reduc_phi; | |
2063 | tree val = PHI_ARG_DEF_FROM_EDGE (phi, exit); | |
2064 | ||
ea057359 | 2065 | if (!virtual_operand_p (val)) |
08dab97a RL |
2066 | { |
2067 | if (dump_file && (dump_flags & TDF_DETAILS)) | |
2068 | { | |
2069 | fprintf (dump_file, "phi is "); | |
2070 | print_gimple_stmt (dump_file, phi, 0, 0); | |
2071 | fprintf (dump_file, "arg of phi to exit: value "); | |
2072 | print_generic_expr (dump_file, val, 0); | |
2073 | fprintf (dump_file, " used outside loop\n"); | |
2074 | fprintf (dump_file, | |
2075 | " checking if it a part of reduction pattern: \n"); | |
2076 | } | |
4a8fb1a1 | 2077 | if (reduction_list.elements () == 0) |
08dab97a RL |
2078 | { |
2079 | if (dump_file && (dump_flags & TDF_DETAILS)) | |
2080 | fprintf (dump_file, | |
2081 | " FAILED: it is not a part of reduction.\n"); | |
2082 | return false; | |
2083 | } | |
2084 | reduc_phi = NULL; | |
2085 | FOR_EACH_IMM_USE_FAST (use_p, imm_iter, val) | |
2086 | { | |
4942af9b JJ |
2087 | if (!gimple_debug_bind_p (USE_STMT (use_p)) |
2088 | && flow_bb_inside_loop_p (loop, gimple_bb (USE_STMT (use_p)))) | |
08dab97a RL |
2089 | { |
2090 | reduc_phi = USE_STMT (use_p); | |
2091 | break; | |
2092 | } | |
2093 | } | |
2094 | red = reduction_phi (reduction_list, reduc_phi); | |
2095 | if (red == NULL) | |
2096 | { | |
2097 | if (dump_file && (dump_flags & TDF_DETAILS)) | |
2098 | fprintf (dump_file, | |
2099 | " FAILED: it is not a part of reduction.\n"); | |
2100 | return false; | |
2101 | } | |
2102 | if (dump_file && (dump_flags & TDF_DETAILS)) | |
2103 | { | |
2104 | fprintf (dump_file, "reduction phi is "); | |
2105 | print_gimple_stmt (dump_file, red->reduc_phi, 0, 0); | |
2106 | fprintf (dump_file, "reduction stmt is "); | |
2107 | print_gimple_stmt (dump_file, red->reduc_stmt, 0, 0); | |
2108 | } | |
2109 | } | |
2110 | } | |
2111 | ||
2112 | /* The iterations of the loop may communicate only through bivs whose | |
2113 | iteration space can be distributed efficiently. */ | |
2114 | for (gsi = gsi_start_phis (loop->header); !gsi_end_p (gsi); gsi_next (&gsi)) | |
2115 | { | |
2116 | gimple phi = gsi_stmt (gsi); | |
2117 | tree def = PHI_RESULT (phi); | |
2118 | affine_iv iv; | |
2119 | ||
ea057359 | 2120 | if (!virtual_operand_p (def) && !simple_iv (loop, loop, def, &iv, true)) |
08dab97a RL |
2121 | { |
2122 | struct reduction_info *red; | |
2123 | ||
2124 | red = reduction_phi (reduction_list, phi); | |
2125 | if (red == NULL) | |
2126 | { | |
2127 | if (dump_file && (dump_flags & TDF_DETAILS)) | |
2128 | fprintf (dump_file, | |
2129 | " FAILED: scalar dependency between iterations\n"); | |
2130 | return false; | |
2131 | } | |
2132 | } | |
2133 | } | |
2134 | ||
2135 | ||
2136 | return true; | |
2137 | } | |
2138 | ||
5f40b3cb ZD |
2139 | /* Detect parallel loops and generate parallel code using libgomp |
2140 | primitives. Returns true if some loop was parallelized, false | |
2141 | otherwise. */ | |
2142 | ||
2143 | bool | |
2144 | parallelize_loops (void) | |
2145 | { | |
2146 | unsigned n_threads = flag_tree_parallelize_loops; | |
2147 | bool changed = false; | |
2148 | struct loop *loop; | |
2149 | struct tree_niter_desc niter_desc; | |
4a8fb1a1 | 2150 | reduction_info_table_type reduction_list; |
f873b205 | 2151 | struct obstack parloop_obstack; |
8adfe01d | 2152 | HOST_WIDE_INT estimated; |
b05e0233 | 2153 | source_location loop_loc; |
f873b205 | 2154 | |
5f40b3cb ZD |
2155 | /* Do not parallelize loops in the functions created by parallelization. */ |
2156 | if (parallelized_function_p (cfun->decl)) | |
2157 | return false; | |
8adfe01d RL |
2158 | if (cfun->has_nonlocal_label) |
2159 | return false; | |
5f40b3cb | 2160 | |
f873b205 | 2161 | gcc_obstack_init (&parloop_obstack); |
4a8fb1a1 | 2162 | reduction_list.create (10); |
726a989a | 2163 | init_stmt_vec_info_vec (); |
a509ebb5 | 2164 | |
f0bd40b1 | 2165 | FOR_EACH_LOOP (loop, 0) |
5f40b3cb | 2166 | { |
4a8fb1a1 | 2167 | reduction_list.empty (); |
48710229 RL |
2168 | if (dump_file && (dump_flags & TDF_DETAILS)) |
2169 | { | |
2170 | fprintf (dump_file, "Trying loop %d as candidate\n",loop->num); | |
2171 | if (loop->inner) | |
2172 | fprintf (dump_file, "loop %d is not innermost\n",loop->num); | |
2173 | else | |
2174 | fprintf (dump_file, "loop %d is innermost\n",loop->num); | |
2175 | } | |
b8698a0f | 2176 | |
48710229 | 2177 | /* If we use autopar in graphite pass, we use its marked dependency |
87d4d0ee SP |
2178 | checking results. */ |
2179 | if (flag_loop_parallelize_all && !loop->can_be_parallel) | |
48710229 RL |
2180 | { |
2181 | if (dump_file && (dump_flags & TDF_DETAILS)) | |
2182 | fprintf (dump_file, "loop is not parallel according to graphite\n"); | |
87d4d0ee | 2183 | continue; |
48710229 | 2184 | } |
87d4d0ee | 2185 | |
48710229 RL |
2186 | if (!single_dom_exit (loop)) |
2187 | { | |
b8698a0f | 2188 | |
48710229 RL |
2189 | if (dump_file && (dump_flags & TDF_DETAILS)) |
2190 | fprintf (dump_file, "loop is !single_dom_exit\n"); | |
b8698a0f | 2191 | |
08dab97a | 2192 | continue; |
48710229 | 2193 | } |
08dab97a RL |
2194 | |
2195 | if (/* And of course, the loop must be parallelizable. */ | |
2196 | !can_duplicate_loop_p (loop) | |
1d4af1e8 | 2197 | || loop_has_blocks_with_irreducible_flag (loop) |
8adfe01d | 2198 | || (loop_preheader_edge (loop)->src->flags & BB_IRREDUCIBLE_LOOP) |
9857228c | 2199 | /* FIXME: the check for vector phi nodes could be removed. */ |
69958396 | 2200 | || loop_has_vector_phi_nodes (loop)) |
08dab97a | 2201 | continue; |
e5b332cd | 2202 | |
652c4c71 | 2203 | estimated = estimated_stmt_executions_int (loop); |
e5b332cd RG |
2204 | if (estimated == -1) |
2205 | estimated = max_stmt_executions_int (loop); | |
87d4d0ee | 2206 | /* FIXME: Bypass this check as graphite doesn't update the |
e5b332cd | 2207 | count and frequency correctly now. */ |
87d4d0ee | 2208 | if (!flag_loop_parallelize_all |
e5b332cd RG |
2209 | && ((estimated != -1 |
2210 | && estimated <= (HOST_WIDE_INT) n_threads * MIN_PER_THREAD) | |
87d4d0ee SP |
2211 | /* Do not bother with loops in cold areas. */ |
2212 | || optimize_loop_nest_for_size_p (loop))) | |
08dab97a | 2213 | continue; |
b8698a0f | 2214 | |
08dab97a RL |
2215 | if (!try_get_loop_niter (loop, &niter_desc)) |
2216 | continue; | |
2217 | ||
2218 | if (!try_create_reduction_list (loop, reduction_list)) | |
2219 | continue; | |
2220 | ||
f873b205 LB |
2221 | if (!flag_loop_parallelize_all |
2222 | && !loop_parallel_p (loop, &parloop_obstack)) | |
5f40b3cb ZD |
2223 | continue; |
2224 | ||
2225 | changed = true; | |
48710229 RL |
2226 | if (dump_file && (dump_flags & TDF_DETAILS)) |
2227 | { | |
48710229 | 2228 | if (loop->inner) |
8adfe01d | 2229 | fprintf (dump_file, "parallelizing outer loop %d\n",loop->header->index); |
48710229 | 2230 | else |
8adfe01d RL |
2231 | fprintf (dump_file, "parallelizing inner loop %d\n",loop->header->index); |
2232 | loop_loc = find_loop_location (loop); | |
b05e0233 | 2233 | if (loop_loc != UNKNOWN_LOCATION) |
8adfe01d | 2234 | fprintf (dump_file, "\nloop at %s:%d: ", |
b05e0233 | 2235 | LOCATION_FILE (loop_loc), LOCATION_LINE (loop_loc)); |
b8698a0f L |
2236 | } |
2237 | gen_parallel_loop (loop, reduction_list, | |
08dab97a | 2238 | n_threads, &niter_desc); |
5f40b3cb ZD |
2239 | } |
2240 | ||
726a989a | 2241 | free_stmt_vec_info_vec (); |
4a8fb1a1 | 2242 | reduction_list.dispose (); |
f873b205 | 2243 | obstack_free (&parloop_obstack, NULL); |
6b8ed145 RG |
2244 | |
2245 | /* Parallelization will cause new function calls to be inserted through | |
d086d311 RG |
2246 | which local variables will escape. Reset the points-to solution |
2247 | for ESCAPED. */ | |
6b8ed145 | 2248 | if (changed) |
d086d311 | 2249 | pt_solution_reset (&cfun->gimple_df->escaped); |
6b8ed145 | 2250 | |
5f40b3cb ZD |
2251 | return changed; |
2252 | } | |
2253 | ||
c1bf2a39 AM |
2254 | /* Parallelization. */ |
2255 | ||
2256 | static bool | |
2257 | gate_tree_parallelize_loops (void) | |
2258 | { | |
2259 | return flag_tree_parallelize_loops > 1; | |
2260 | } | |
2261 | ||
2262 | static unsigned | |
2263 | tree_parallelize_loops (void) | |
2264 | { | |
2265 | if (number_of_loops (cfun) <= 1) | |
2266 | return 0; | |
2267 | ||
2268 | if (parallelize_loops ()) | |
2269 | return TODO_cleanup_cfg | TODO_rebuild_alias; | |
2270 | return 0; | |
2271 | } | |
2272 | ||
2273 | namespace { | |
2274 | ||
2275 | const pass_data pass_data_parallelize_loops = | |
2276 | { | |
2277 | GIMPLE_PASS, /* type */ | |
2278 | "parloops", /* name */ | |
2279 | OPTGROUP_LOOP, /* optinfo_flags */ | |
2280 | true, /* has_gate */ | |
2281 | true, /* has_execute */ | |
2282 | TV_TREE_PARALLELIZE_LOOPS, /* tv_id */ | |
2283 | ( PROP_cfg | PROP_ssa ), /* properties_required */ | |
2284 | 0, /* properties_provided */ | |
2285 | 0, /* properties_destroyed */ | |
2286 | 0, /* todo_flags_start */ | |
2287 | TODO_verify_flow, /* todo_flags_finish */ | |
2288 | }; | |
2289 | ||
2290 | class pass_parallelize_loops : public gimple_opt_pass | |
2291 | { | |
2292 | public: | |
2293 | pass_parallelize_loops (gcc::context *ctxt) | |
2294 | : gimple_opt_pass (pass_data_parallelize_loops, ctxt) | |
2295 | {} | |
2296 | ||
2297 | /* opt_pass methods: */ | |
2298 | bool gate () { return gate_tree_parallelize_loops (); } | |
2299 | unsigned int execute () { return tree_parallelize_loops (); } | |
2300 | ||
2301 | }; // class pass_parallelize_loops | |
2302 | ||
2303 | } // anon namespace | |
2304 | ||
2305 | gimple_opt_pass * | |
2306 | make_pass_parallelize_loops (gcc::context *ctxt) | |
2307 | { | |
2308 | return new pass_parallelize_loops (ctxt); | |
2309 | } | |
2310 | ||
2311 | ||
5f40b3cb | 2312 | #include "gt-tree-parloops.h" |