/*
- * bepbqpalloc.c
+ * Copyright (C) 1995-2008 University of Karlsruhe. All right reserved.
*
- * Created on: Nov 11, 2009
- * Author: bersch
+ * This file is part of libFirm.
+ *
+ * This file may be distributed and/or modified under the terms of the
+ * GNU General Public License version 2 as published by the Free Software
+ * Foundation and appearing in the file LICENSE.GPL included in the
+ * packaging of this file.
+ *
+ * Licensees holding valid libFirm Professional Edition licenses may use
+ * this file in accordance with the libFirm Commercial License.
+ * Agreement provided with the Software.
+ *
+ * This file is provided AS IS with NO WARRANTY OF ANY KIND, INCLUDING THE
+ * WARRANTY OF DESIGN, MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ * PURPOSE.
+ */
+
+/**
+ * @file
+ * @brief PBQP based register allocation.
+ * @author Thomas Bersch
+ * @date 27.11.2009
+ * @version $Id: bechordal.c 26750 2009-11-27 09:37:43Z bersch $
*/
/* miscellaneous includes */
#include "beutil.h"
#include "plist.h"
#include "pqueue.h"
+#include "becopyopt.h"
/* pbqp includes */
#include "kaps.h"
#include "matrix.h"
#include "vector.h"
#include "vector_t.h"
-#include "heuristical.h"
+#include "heuristical_co.h"
+#include "heuristical_co_ld.h"
#include "pbqp_t.h"
#include "html_dumper.h"
#include "pbqp_node_t.h"
#include "pbqp_node.h"
+#define TIMER 1
+#define PRINT_RPEO 1
+
+
+static int use_exec_freq = true;
+static int use_late_decision = true;
typedef struct _be_pbqp_alloc_env_t {
- pbqp *pbqp_inst; /**< PBQP instance for register allocation */
- be_irg_t *birg; /**< Back-end IRG session. */
- ir_graph *irg; /**< The graph under examination. */
- const arch_register_class_t *cls; /**< Current processed register class */
+ pbqp *pbqp_inst; /**< PBQP instance for register allocation */
+ be_irg_t *birg; /**< Back-end IRG session. */
+ ir_graph *irg; /**< The graph under examination. */
+ const arch_register_class_t *cls; /**< Current processed register class */
be_lv_t *lv;
bitset_t *ignored_regs;
- pbqp_matrix *ife_matrix_dummy;
- pbqp_matrix *aff_matrix_dummy;
+ pbqp_matrix *ife_matrix_template;
+ pbqp_matrix *aff_matrix_template;
plist_t *rpeo;
unsigned *restr_nodes;
+ unsigned *ife_edge_num;
be_chordal_env_t *env;
} be_pbqp_alloc_env_t;
-#define is_Reg_Phi(irn) (is_Phi(irn) && mode_is_data(get_irn_mode(irn)))
-#define get_Perm_src(irn) (get_irn_n(get_Proj_pred(irn), get_Proj_proj(irn)))
-#define is_Perm_Proj(irn) (is_Proj(irn) && be_is_Perm(get_Proj_pred(irn)))
+#define is_Reg_Phi(irn) (is_Phi(irn) && mode_is_data(get_irn_mode(irn)))
+#define get_Perm_src(irn) (get_irn_n(get_Proj_pred(irn), get_Proj_proj(irn)))
+#define is_Perm_Proj(irn) (is_Proj(irn) && be_is_Perm(get_Proj_pred(irn)))
+#define insert_edge(pbqp, src_node, trg_node, template_matrix) (add_edge_costs(pbqp, get_irn_idx(src_node), get_irn_idx(trg_node), pbqp_matrix_copy(pbqp, template_matrix)))
+#define get_free_regs(restr_nodes, cls, irn) (arch_register_class_n_regs(cls) - restr_nodes[get_irn_idx(irn)])
static inline int is_2addr_code(const arch_register_req_t *req)
{
return (req->type & arch_register_req_type_should_be_same) != 0;
}
+static const lc_opt_table_entry_t options[] = {
+ LC_OPT_ENT_BOOL ("exec_freq", "use exec_freq", &use_exec_freq),
+ LC_OPT_ENT_BOOL ("late_decision", "use late decision for register allocation", &use_late_decision),
+ LC_OPT_LAST
+};
#if KAPS_DUMP
static FILE *my_open(const be_chordal_env_t *env, const char *prefix, const char *suffix)
ir_snprintf(buf, sizeof(buf), "%s%s_%F_%s%s", prefix, tu_name, env->irg, env->cls->name, suffix);
xfree(tu_name);
result = fopen(buf, "wt");
- if(result == NULL) {
+ if (result == NULL) {
panic("Couldn't open '%s' for writing.", buf);
}
#endif
-static unsigned create_pbqp_node(be_pbqp_alloc_env_t *pbqp_alloc_env, ir_node *irn) {
+static void create_pbqp_node(be_pbqp_alloc_env_t *pbqp_alloc_env, ir_node *irn)
+{
const arch_register_class_t *cls = pbqp_alloc_env->cls;
pbqp *pbqp_inst = pbqp_alloc_env->pbqp_inst;
bitset_t *ignored_regs = pbqp_alloc_env->ignored_regs;
/* set costs depending on register constrains */
unsigned idx;
- for(idx = 0; idx < colors_n; idx++) {
- if(bitset_is_set(ignored_regs, idx) || !arch_reg_out_is_allocatable(irn, arch_register_for_index(cls, idx))) {
+ for (idx = 0; idx < colors_n; idx++) {
+ if (bitset_is_set(ignored_regs, idx) || !arch_reg_out_is_allocatable(irn, arch_register_for_index(cls, idx))) {
vector_set(costs_vector, idx, INF_COSTS);
cntConstrains++;
}
/* add vector to pbqp node */
add_node_costs(pbqp_inst, get_irn_idx(irn), costs_vector);
+ pbqp_alloc_env->restr_nodes[get_irn_idx(irn)] = cntConstrains;
+}
+
+static void insert_ife_edge(be_pbqp_alloc_env_t *pbqp_alloc_env, ir_node *src_node, ir_node *trg_node)
+{
+ pbqp *pbqp = pbqp_alloc_env->pbqp_inst;
+ const arch_register_class_t *cls = pbqp_alloc_env->cls;
+ pbqp_matrix *ife_matrix_template = pbqp_alloc_env->ife_matrix_template;
+ unsigned *restr_nodes = pbqp_alloc_env->restr_nodes;
+
+ if (get_edge(pbqp, get_irn_idx(src_node), get_irn_idx(trg_node)) == NULL) {
+
+ /* increase ife edge counter */
+ pbqp_alloc_env->ife_edge_num[get_irn_idx(src_node)]++;
+ pbqp_alloc_env->ife_edge_num[get_irn_idx(trg_node)]++;
+
+ /* do useful optimization to speed up pbqp solving (we can do this because we know our matrix) */
+ if (get_free_regs(restr_nodes, cls, src_node) == 1 && get_free_regs(restr_nodes, cls, trg_node) == 1) {
+ unsigned src_idx = vector_get_min_index(get_node(pbqp, get_irn_idx(src_node))->costs);
+ unsigned trg_idx = vector_get_min_index(get_node(pbqp, get_irn_idx(trg_node))->costs);
+ assert(src_idx != trg_idx && "Interfering nodes could not have the same register!");
+ return;
+ }
+ if (get_free_regs(restr_nodes, cls, src_node) == 1 || get_free_regs(restr_nodes, cls, trg_node) == 1) {
+ if (get_free_regs(restr_nodes, cls, src_node) == 1) {
+ unsigned idx = vector_get_min_index(get_node(pbqp, get_irn_idx(src_node))->costs);
+ vector_set(get_node(pbqp, get_irn_idx(trg_node))->costs, idx, INF_COSTS);
+ }
+ else {
+ unsigned idx = vector_get_min_index(get_node(pbqp, get_irn_idx(trg_node))->costs);
+ vector_set(get_node(pbqp, get_irn_idx(src_node))->costs, idx, INF_COSTS);
+ }
+ return;
+ }
- /* return number of free selectable registers */
- return (colors_n - cntConstrains);
+ /* insert interference edge */
+ insert_edge(pbqp, src_node, trg_node, ife_matrix_template);
+ }
}
-static void build_graph_walker(ir_node *irn, void *env) {
- be_pbqp_alloc_env_t *pbqp_alloc_env = env;
- pbqp *pbqp_inst = pbqp_alloc_env->pbqp_inst;
- const arch_register_class_t *cls = pbqp_alloc_env->cls;
- const arch_register_req_t *req = arch_get_register_req_out(irn);
- unsigned pos, max;
+static void inser_afe_edge(be_pbqp_alloc_env_t *pbqp_alloc_env, ir_node *src_node, ir_node *trg_node, int pos)
+{
+ pbqp *pbqp = pbqp_alloc_env->pbqp_inst;
+ const arch_register_class_t *cls = pbqp_alloc_env->cls;
+ unsigned *restr_nodes = pbqp_alloc_env->restr_nodes;
+ pbqp_matrix *afe_matrix = pbqp_matrix_alloc(pbqp, arch_register_class_n_regs(cls), arch_register_class_n_regs(cls));
+ unsigned colors_n = arch_register_class_n_regs(cls);
+
+ if (get_edge(pbqp, get_irn_idx(src_node), get_irn_idx(trg_node)) == NULL) {
+ if (use_exec_freq) {
+ /* get exec_freq for copy_block */
+ ir_node *root_bl = get_nodes_block(src_node);
+ ir_node *copy_bl = is_Phi(src_node) ? get_Block_cfgpred_block(root_bl, pos) : root_bl;
+ unsigned long res = get_block_execfreq_ulong(pbqp_alloc_env->birg->exec_freq, copy_bl);
+
+ /* create afe-matrix */
+ unsigned row, col;
+ for (row = 0; row < colors_n; row++) {
+ for (col = 0; col < colors_n; col++) {
+ if (row != col)
+ pbqp_matrix_set(afe_matrix, row, col, (num)res);
+ }
+ }
+ }
+ else {
+ afe_matrix = pbqp_alloc_env->aff_matrix_template;
+ }
- if (arch_irn_consider_in_reg_alloc(cls, irn))
- return;
+ /* do useful optimization to speed up pbqp solving */
+ if (get_free_regs(restr_nodes, cls, src_node) == 1 && get_free_regs(restr_nodes, cls, trg_node) == 1) {
+ return;
+ }
+ if (get_free_regs(restr_nodes, cls, src_node) == 1 || get_free_regs(restr_nodes, cls, trg_node) == 1) {
+ if (get_free_regs(restr_nodes, cls, src_node) == 1) {
+ unsigned regIdx = vector_get_min_index(get_node(pbqp, get_irn_idx(src_node))->costs);
+ vector_add_matrix_col(get_node(pbqp, get_irn_idx(trg_node))->costs, afe_matrix, regIdx);
+ }
+ else {
+ unsigned regIdx = vector_get_min_index(get_node(pbqp, get_irn_idx(trg_node))->costs);
+ vector_add_matrix_col(get_node(pbqp, get_irn_idx(src_node))->costs, afe_matrix, regIdx);
+ }
+ return;
+ }
+
+ /* insert interference edge */
+ insert_edge(pbqp, src_node, trg_node, afe_matrix);
+ }
+}
+
+static void create_affinity_edges(ir_node *irn, void *env)
+{
+ be_pbqp_alloc_env_t *pbqp_alloc_env = env;
+ const arch_register_class_t *cls = pbqp_alloc_env->cls;
+ const arch_register_req_t *req = arch_get_register_req_out(irn);
+ unsigned pos, max;
if (is_Reg_Phi(irn)) { /* Phis */
for (pos=0, max=get_irn_arity(irn); pos<max; ++pos) {
ir_node *arg = get_irn_n(irn, pos);
- //add_edges(co, irn, arg, co->get_costs(co, irn, arg, pos));
if (!arch_irn_consider_in_reg_alloc(cls, arg))
continue;
/* no edges to itself */
- if(irn == arg) {
+ if (irn == arg) {
continue;
}
- if(get_edge(pbqp_inst, get_irn_idx(irn), get_irn_idx(arg)) == NULL) {
- /* copy matrix */
- struct pbqp_matrix *matrix = pbqp_matrix_copy(pbqp_inst, pbqp_alloc_env->aff_matrix_dummy);
- /* add costs matrix to affinity edge */
- add_edge_costs(pbqp_inst, get_irn_idx(irn), get_irn_idx(arg) , matrix);
- }
+ inser_afe_edge(pbqp_alloc_env, irn, arg, pos);
}
}
else if (is_Perm_Proj(irn)) { /* Perms */
ir_node *arg = get_Perm_src(irn);
- //add_edges(co, irn, arg, co->get_costs(co, irn, arg, 0));
-
if (!arch_irn_consider_in_reg_alloc(cls, arg))
return;
- if(get_edge(pbqp_inst, get_irn_idx(irn), get_irn_idx(arg)) == NULL) {
- /* copy matrix */
- struct pbqp_matrix *matrix = pbqp_matrix_copy(pbqp_inst, pbqp_alloc_env->aff_matrix_dummy);
- /* add costs matrix to affinity edge */
- add_edge_costs(pbqp_inst, get_irn_idx(irn), get_irn_idx(arg) , matrix);
- }
+ inser_afe_edge(pbqp_alloc_env, irn, arg, -1);
}
else { /* 2-address code */
if (is_2addr_code(req)) {
for (i = 0; 1U << i <= other; ++i) {
if (other & (1U << i)) {
ir_node *other = get_irn_n(skip_Proj(irn), i);
-// if (!arch_irn_is_ignore(other)) {
- //add_edges(co, irn, other, co->get_costs(co, irn, other, 0));
- if (!arch_irn_consider_in_reg_alloc(cls, other))
- continue;
-
- /* no edges to itself */
- if(irn == other) {
- continue;
- }
-
- if(get_edge(pbqp_inst, get_irn_idx(irn), get_irn_idx(other)) == NULL) {
- /* copy matrix */
- struct pbqp_matrix *matrix = pbqp_matrix_copy(pbqp_inst, pbqp_alloc_env->aff_matrix_dummy);
- /* add costs matrix to affinity edge */
- add_edge_costs(pbqp_inst, get_irn_idx(irn), get_irn_idx(other) , matrix);
- }
-// }
+ if (!arch_irn_consider_in_reg_alloc(cls, other))
+ continue;
+
+ /* no edges to itself */
+ if (irn == other) {
+ continue;
+ }
+
+ inser_afe_edge(pbqp_alloc_env, irn, other, i);
}
}
}
}
}
-static void create_pbqp_coloring_inst(ir_node *block, void *data) {
+static void create_pbqp_coloring_instance(ir_node *block, void *data)
+{
be_pbqp_alloc_env_t *pbqp_alloc_env = data;
be_lv_t *lv = pbqp_alloc_env->lv;
const arch_register_class_t *cls = pbqp_alloc_env->cls;
plist_t *rpeo = pbqp_alloc_env->rpeo;
pbqp *pbqp_inst = pbqp_alloc_env->pbqp_inst;
unsigned *restr_nodes = pbqp_alloc_env->restr_nodes;
- pbqp_matrix *ife_matrix_dummy = pbqp_alloc_env->ife_matrix_dummy;
pqueue_t *queue = new_pqueue();
pqueue_t *restr_nodes_queue = new_pqueue();
plist_t *temp_list = plist_new();
+ plist_t *sorted_list = plist_new();
ir_node *irn;
ir_nodeset_t live_nodes;
+ plist_element_t *el;
+ ir_node *last_element = NULL;
/* first, determine the pressure */
/* (this is only for compatibility with copymin optimization, it's not needed for pbqp coloring) */
- pressure(block, pbqp_alloc_env->env);
+ create_borders(block, pbqp_alloc_env->env);
/* calculate living nodes for the first step */
ir_nodeset_init(&live_nodes);
/* create pbqp nodes, interference edges and reverse perfect elimination order */
sched_foreach_reverse(block, irn) {
- ir_node *live, *if_live;
- ir_nodeset_iterator_t iter, iter2;
-
- /* create nodes and interference edges */
- foreach_ir_nodeset(&live_nodes, live, iter) {
- /* create pbqp source node if it dosn't exist */
- if(get_node(pbqp_inst, get_irn_idx(live)) == NULL) {
- restr_nodes[get_irn_idx(live)] = create_pbqp_node(pbqp_alloc_env, live);
- }
+ ir_node *live;
+ ir_nodeset_iterator_t iter;
+
+ if (get_irn_mode(irn) == mode_T) {
+ const ir_edge_t *edge;
+ foreach_out_edge(irn, edge) {
+ ir_node *proj = get_edge_src_irn(edge);
+ if (!arch_irn_consider_in_reg_alloc(cls, proj))
+ continue;
- iter2 = iter;
- for(if_live = ir_nodeset_iterator_next(&iter2); if_live != NULL; if_live = ir_nodeset_iterator_next(&iter2)) {
- /* create pbqp target node if it dosn't exist */
- if(get_node(pbqp_inst, get_irn_idx(if_live)) == NULL) {
- restr_nodes[get_irn_idx(if_live)] = create_pbqp_node(pbqp_alloc_env, if_live);
+ /* create pbqp source node if it dosn't exist */
+ if (get_node(pbqp_inst, get_irn_idx(proj)) == NULL) {
+ create_pbqp_node(pbqp_alloc_env, proj);
}
- else {
+
+ /* create nodes and interference edges */
+ foreach_ir_nodeset(&live_nodes, live, iter) {
+ /* create pbqp source node if it dosn't exist */
+ if (get_node(pbqp_inst, get_irn_idx(live)) == NULL) {
+ create_pbqp_node(pbqp_alloc_env, live);
+ }
+
/* no edges to itself */
- if(live == if_live)
+ if (proj == live) {
continue;
- /* only one interference edge between two nodes */
- if(get_edge(pbqp_inst, get_irn_idx(live), get_irn_idx(if_live)))
- continue;
- }
+ }
- /* do useful optimization to improve pbqp solving (we can do this because we know our matrix) */
- if(restr_nodes[get_irn_idx(live)] == 1 && restr_nodes[get_irn_idx(if_live)] == 1) {
- unsigned src_idx = vector_get_min_index(get_node(pbqp_inst, get_irn_idx(live))->costs);
- unsigned trg_idx = vector_get_min_index(get_node(pbqp_inst, get_irn_idx(if_live))->costs);
- assert(src_idx != trg_idx && "Interfering nodes could not have the same register!");
- continue;
+ insert_ife_edge(pbqp_alloc_env, proj, live);
}
- if(restr_nodes[get_irn_idx(live)] == 1 || restr_nodes[get_irn_idx(if_live)] == 1) {
- if(restr_nodes[get_irn_idx(live)] == 1) {
- unsigned idx = vector_get_min_index(get_node(pbqp_inst, get_irn_idx(live))->costs);
- vector_set(get_node(pbqp_inst, get_irn_idx(if_live))->costs, idx, INF_COSTS);
+ }
+ }
+ else {
+ if (arch_irn_consider_in_reg_alloc(cls, irn)) {
+ /* create pbqp source node if it dosn't exist */
+ if (get_node(pbqp_inst, get_irn_idx(irn)) == NULL) {
+ create_pbqp_node(pbqp_alloc_env, irn);
+ }
+
+ /* create nodes and interference edges */
+ foreach_ir_nodeset(&live_nodes, live, iter) {
+ /* create pbqp source node if it dosn't exist */
+ if (get_node(pbqp_inst, get_irn_idx(live)) == NULL) {
+ create_pbqp_node(pbqp_alloc_env, live);
}
- else {
- unsigned idx = vector_get_min_index(get_node(pbqp_inst, get_irn_idx(if_live))->costs);
- vector_set(get_node(pbqp_inst, get_irn_idx(live))->costs, idx, INF_COSTS);
+
+ /* no edges to itself */
+ if (irn == live) {
+ continue;
}
- continue;
- }
- /* copy matrix */
- struct pbqp_matrix *matrix = pbqp_matrix_copy(pbqp_inst, ife_matrix_dummy);
- /* add costs matrix to interference edge */
- add_edge_costs(pbqp_inst, get_irn_idx(live), get_irn_idx(if_live) , matrix);
+ /* insert interference edge */
+ insert_ife_edge(pbqp_alloc_env, irn, live);
+ }
}
}
+ /* get living nodes for next step */
+ if (!is_Phi(irn)) {
+ be_liveness_transfer(cls, irn, &live_nodes);
+ }
+
/* order nodes for perfect elimination order */
if (get_irn_mode(irn) == mode_T) {
- plist_element_t *first = plist_first(temp_list);
- const ir_edge_t *edge;
+ bool allHaveIFEdges = true;
+ const ir_edge_t *edge;
foreach_out_edge(irn, edge) {
ir_node *proj = get_edge_src_irn(edge);
if (!arch_irn_consider_in_reg_alloc(cls, proj))
continue;
- // insert proj node into priority queue (descending by the number of interference edges)
- if(restr_nodes[get_irn_idx(proj)] <= 4/*bitset_is_set(restr_nodes, get_irn_idx(proj))*/) {
- pqueue_put(restr_nodes_queue, proj, pbqp_node_get_degree(get_node(pbqp_inst, get_irn_idx(proj))));
+ /* insert proj node into priority queue (descending by the number of interference edges) */
+ if (get_free_regs(restr_nodes, cls, proj) <= 4) {
+ pqueue_put(restr_nodes_queue, proj, pbqp_alloc_env->ife_edge_num[get_irn_idx(proj)]);
}
else {
- pqueue_put(queue,proj, pbqp_node_get_degree(get_node(pbqp_inst, get_irn_idx(proj))));
+ pqueue_put(queue, proj, pbqp_alloc_env->ife_edge_num[get_irn_idx(proj)]);
}
+ /* skip last step if there is no last_element */
+ if(last_element == NULL)
+ continue;
+
+ /* check if proj has an if edge to last_element (at this time pbqp contains only if edges) */
+ if(get_edge(pbqp_inst, proj->node_idx, last_element->node_idx) == NULL && get_edge(pbqp_inst, last_element->node_idx, proj->node_idx) == NULL) {
+ allHaveIFEdges = false; /* there is no if edge between proj and last_element */
+ }
}
- /* first insert all restricted nodes */
- while(!pqueue_empty(restr_nodes_queue)) {
- if(first == NULL) {
- plist_insert_back(temp_list, get_node(pbqp_inst, get_irn_idx(pqueue_pop_front(restr_nodes_queue))));
- first = plist_first(temp_list);
- } else {
- plist_insert_before(temp_list, first, get_node(pbqp_inst, get_irn_idx(pqueue_pop_front(restr_nodes_queue))));
+ if(last_element != NULL && allHaveIFEdges) {
+ if (get_free_regs(restr_nodes, cls, last_element) <= 4) {
+ pqueue_put(restr_nodes_queue, last_element, pbqp_alloc_env->ife_edge_num[get_irn_idx(last_element)]);
+ }
+ else {
+ pqueue_put(queue, last_element, pbqp_alloc_env->ife_edge_num[get_irn_idx(last_element)]);
}
+ plist_erase(temp_list, plist_find_value(temp_list, get_node(pbqp_inst, last_element->node_idx)));
+ last_element = NULL;
+ }
+
+ /* first insert all restricted proj nodes */
+ while (!pqueue_empty(restr_nodes_queue)) {
+ plist_insert_front(sorted_list, get_node(pbqp_inst, get_irn_idx(pqueue_pop_front(restr_nodes_queue))));
}
/* insert proj nodes descending by their number of interference edges */
- while(!pqueue_empty(queue)) {
- if(first == NULL) {
- plist_insert_back(temp_list, get_node(pbqp_inst, get_irn_idx(pqueue_pop_front(queue))));
- first = plist_first(temp_list);
- } else {
- plist_insert_before(temp_list, first, get_node(pbqp_inst, get_irn_idx(pqueue_pop_front(queue))));
- }
+ while (!pqueue_empty(queue)) {
+ plist_insert_front(sorted_list, get_node(pbqp_inst, get_irn_idx(pqueue_pop_front(queue))));
+ }
+
+ /* invert sorted list */
+ foreach_plist(sorted_list, el) {
+ plist_insert_front(temp_list, el->data);
}
+
+ plist_clear(sorted_list);
+
}
else {
if (arch_irn_consider_in_reg_alloc(cls, irn)) {
+ // remember last colorable node
+ last_element = irn;
plist_insert_front(temp_list, get_node(pbqp_inst, get_irn_idx(irn)));
}
- }
-
- /* get living nodes for next step */
- if (!is_Phi(irn)) {
- be_liveness_transfer(cls, irn, &live_nodes);
+ else {
+ // node not colorable, so ignore it
+ last_element = NULL;
+ }
}
}
/* insert nodes into reverse perfect elimination order */
- plist_element_t *el;
foreach_plist(temp_list, el) {
plist_insert_back(rpeo, el->data);
}
/* free reserved memory */
ir_nodeset_destroy(&live_nodes);
plist_free(temp_list);
+ plist_free(sorted_list);
del_pqueue(queue);
del_pqueue(restr_nodes_queue);
}
-static void insert_perms(ir_node *block, void *data) {
+static void insert_perms(ir_node *block, void *data)
+{
/*
* Start silent in the start block.
* The silence remains until the first barrier is seen.
}
}
-
-void be_pbqp_coloring(be_chordal_env_t *env) {
- ir_graph *irg = env->irg;
- be_irg_t *birg = env->birg;
- const arch_register_class_t *cls = env->cls;
- unsigned colors_n = arch_register_class_n_regs(cls);
- be_pbqp_alloc_env_t pbqp_alloc_env;
- unsigned idx, row, col;
- be_lv_t *lv;
-
-// ir_timer_t *t_ra_pbqp_alloc_create = ir_timer_register("be_pbqp_alloc_create", "pbqp alloc create");
-// ir_timer_t *t_ra_pbqp_alloc_solve = ir_timer_register("be_pbqp_alloc_solve", "pbqp alloc solve");
-// ir_timer_t *t_ra_pbqp_alloc_create_aff = ir_timer_register("be_pbqp_alloc_create_aff", "pbqp alloc create aff");
-
+static void be_pbqp_coloring(be_chordal_env_t *env)
+{
+ ir_graph *irg = env->irg;
+ be_irg_t *birg = env->birg;
+ const arch_register_class_t *cls = env->cls;
+ be_lv_t *lv = NULL;
+ plist_element_t *element = NULL;
+ unsigned colors_n = arch_register_class_n_regs(cls);
+ be_pbqp_alloc_env_t pbqp_alloc_env;
+ unsigned row, col;
+
+
+#if TIMER
+ ir_timer_t *t_ra_pbqp_alloc_create = ir_timer_new();
+ ir_timer_t *t_ra_pbqp_alloc_solve = ir_timer_new();
+ ir_timer_t *t_ra_pbqp_alloc_create_aff = ir_timer_new();
+
+ printf("#### ----- === Allocating registers of %s (%s) ===\n", cls->name, get_entity_name(get_irg_entity(irg)));
+#endif
lv = be_assure_liveness(birg);
be_liveness_assure_sets(lv);
be_liveness_assure_chk(lv);
-// printf("#### ----- === Allocating registers of %s (%s) ===\n", cls->name, get_entity_name(get_irg_entity(irg)));
-
/* insert perms */
assure_doms(irg);
dom_tree_walk_irg(irg, insert_perms, NULL, env);
be_dump(irg, buf, dump_ir_block_graph_sched);
}
+
/* initialize pbqp allocation data structure */
pbqp_alloc_env.pbqp_inst = alloc_pbqp(get_irg_last_idx(irg)); /* initialize pbqp instance */
pbqp_alloc_env.birg = birg;
pbqp_alloc_env.ignored_regs = bitset_malloc(colors_n);
pbqp_alloc_env.rpeo = plist_new();
pbqp_alloc_env.restr_nodes = XMALLOCNZ(unsigned, get_irg_last_idx(irg));
+ pbqp_alloc_env.ife_edge_num = XMALLOCNZ(unsigned, get_irg_last_idx(irg));
pbqp_alloc_env.env = env;
be_put_ignore_regs(birg, cls, pbqp_alloc_env.ignored_regs); /* get ignored registers */
- /* create costs matrix for interference edges */
+
+ /* create costs matrix template for interference edges */
struct pbqp_matrix *ife_matrix = pbqp_matrix_alloc(pbqp_alloc_env.pbqp_inst, colors_n, colors_n);
/* set costs */
- for(row = 0, col=0; row < colors_n; row++, col++)
+ for (row = 0, col=0; row < colors_n; row++, col++)
pbqp_matrix_set(ife_matrix, row, col, INF_COSTS);
- pbqp_alloc_env.ife_matrix_dummy = ife_matrix;
+ pbqp_alloc_env.ife_matrix_template = ife_matrix;
- /* create costs matrix for affinity edges */
- struct pbqp_matrix *afe_matrix = pbqp_matrix_alloc(pbqp_alloc_env.pbqp_inst, colors_n, colors_n);
- /* set costs */
- for(row = 0; row < colors_n; row++) {
- for(col = 0; col < colors_n; col++) {
- if(row != col)
- pbqp_matrix_set(afe_matrix, row, col, 2);
+
+ if (!use_exec_freq) {
+ /* create costs matrix template for affinity edges */
+ struct pbqp_matrix *afe_matrix = pbqp_matrix_alloc(pbqp_alloc_env.pbqp_inst, colors_n, colors_n);
+ /* set costs */
+ for (row = 0; row < colors_n; row++) {
+ for (col = 0; col < colors_n; col++) {
+ if (row != col)
+ pbqp_matrix_set(afe_matrix, row, col, 2);
+ }
}
+ pbqp_alloc_env.aff_matrix_template = afe_matrix;
}
- pbqp_alloc_env.aff_matrix_dummy = afe_matrix;
/* create pbqp instance */
-// ir_timer_reset_and_start(t_ra_pbqp_alloc_create);
+#if TIMER
+ ir_timer_reset_and_start(t_ra_pbqp_alloc_create);
+#endif
assure_doms(irg);
- dom_tree_walk_irg(irg, create_pbqp_coloring_inst , NULL, &pbqp_alloc_env);
-// ir_timer_stop(t_ra_pbqp_alloc_create);
+ dom_tree_walk_irg(irg, create_pbqp_coloring_instance , NULL, &pbqp_alloc_env);
+#if TIMER
+ ir_timer_stop(t_ra_pbqp_alloc_create);
+#endif
+
/* set up affinity edges */
-// ir_timer_reset_and_start(t_ra_pbqp_alloc_create_aff);
- irg_walk_graph(irg, build_graph_walker, NULL, &pbqp_alloc_env);
-// ir_timer_stop(t_ra_pbqp_alloc_create_aff);
+#if TIMER
+ ir_timer_reset_and_start(t_ra_pbqp_alloc_create_aff);
+#endif
+ foreach_plist(pbqp_alloc_env.rpeo, element) {
+ pbqp_node *node = element->data;
+ ir_node *irn = get_idx_irn(irg, node->index);
+
+ create_affinity_edges(irn, &pbqp_alloc_env);
+ }
+#if TIMER
+ ir_timer_stop(t_ra_pbqp_alloc_create_aff);
+#endif
+
#if KAPS_DUMP
// dump graph before solving pbqp
set_dumpfile(pbqp_alloc_env.pbqp_inst, file_before);
#endif
+ /* print out reverse perfect eleminiation order */
+#if PRINT_RPEO
+ plist_element_t *elements;
+ foreach_plist(pbqp_alloc_env.rpeo, elements) {
+ pbqp_node *node = elements->data;
+ printf(" %d(%lu);", node->index, get_idx_irn(irg, node->index)->node_nr);
+ }
+ printf("\n");
+#endif
+
+
/* solve pbqp instance */
-// ir_timer_reset_and_start(t_ra_pbqp_alloc_solve);
- solve_pbqp_heuristical_co(pbqp_alloc_env.pbqp_inst,pbqp_alloc_env.rpeo);
-// ir_timer_stop(t_ra_pbqp_alloc_solve);
+#if TIMER
+ ir_timer_reset_and_start(t_ra_pbqp_alloc_solve);
+#endif
+ if(use_late_decision) {
+ solve_pbqp_heuristical_co_ld(pbqp_alloc_env.pbqp_inst,pbqp_alloc_env.rpeo);
+ }
+ else {
+ solve_pbqp_heuristical_co(pbqp_alloc_env.pbqp_inst,pbqp_alloc_env.rpeo);
+ }
+#if TIMER
+ ir_timer_stop(t_ra_pbqp_alloc_solve);
+#endif
num solution = get_solution(pbqp_alloc_env.pbqp_inst);
assert(solution != INF_COSTS && "No PBQP solution found");
- plist_element_t *element;
+
+ /* assign colors */
foreach_plist(pbqp_alloc_env.rpeo, element) {
- pbqp_node *node = element->data;
- idx = node->index;
- ir_node *irn = get_idx_irn(irg, idx);
- num color = get_node_solution(pbqp_alloc_env.pbqp_inst, idx);
- const arch_register_t *reg = arch_register_for_index(cls, color);
+ pbqp_node *node = element->data;
+ ir_node *irn = get_idx_irn(irg, node->index);
+ num color = get_node_solution(pbqp_alloc_env.pbqp_inst, node->index);
+ const arch_register_t *reg = arch_register_for_index(cls, color);
arch_set_irn_register(irn, reg);
}
-// printf("%-20s: %8.3lf msec\n" , ir_timer_get_description(t_ra_pbqp_alloc_create), (double)ir_timer_elapsed_usec(t_ra_pbqp_alloc_create) / 1000.0);
-// printf("%-20s: %8.3lf msec\n" , ir_timer_get_description(t_ra_pbqp_alloc_solve), (double)ir_timer_elapsed_usec(t_ra_pbqp_alloc_solve) / 1000.0);
-// printf("%-20s: %8.3lf msec\n" , ir_timer_get_description(t_ra_pbqp_alloc_create_aff), (double)ir_timer_elapsed_usec(t_ra_pbqp_alloc_create_aff) / 1000.0);
+
+#if TIMER
+ printf("%-20s: %8.3lf msec\n", "pbqp alloc create",
+ (double)ir_timer_elapsed_usec(t_ra_pbqp_alloc_create) / 1000.0);
+ printf("%-20s: %8.3lf msec\n", "pbqp alloc solve",
+ (double)ir_timer_elapsed_usec(t_ra_pbqp_alloc_solve) / 1000.0);
+ printf("%-20s: %8.3lf msec\n", "pbqp alloc create aff",
+ (double)ir_timer_elapsed_usec(t_ra_pbqp_alloc_create_aff) / 1000.0);
+#endif
/* free reserved memory */
free_pbqp(pbqp_alloc_env.pbqp_inst);
plist_free(pbqp_alloc_env.rpeo);
xfree(pbqp_alloc_env.restr_nodes);
+ xfree(pbqp_alloc_env.ife_edge_num);
}
/**
* Initializes this module.
*/
-void be_init_pbqp_coloring(void) {
+BE_REGISTER_MODULE_CONSTRUCTOR(be_init_pbqp_coloring);
+void be_init_pbqp_coloring(void)
+{
+ lc_opt_entry_t *be_grp = lc_opt_get_grp(firm_opt_get_root(), "be");
+ lc_opt_entry_t *ra_grp = lc_opt_get_grp(be_grp, "ra");
+ lc_opt_entry_t *chordal_grp = lc_opt_get_grp(ra_grp, "chordal");
+ lc_opt_entry_t *coloring_grp = lc_opt_get_grp(chordal_grp, "coloring");
+ lc_opt_entry_t *pbqp_grp = lc_opt_get_grp(coloring_grp, "pbqp");
static be_ra_chordal_coloring_t coloring = {
be_pbqp_coloring
};
+ lc_opt_add_table(pbqp_grp, options);
be_register_chordal_coloring("pbqp", &coloring);
}
-BE_REGISTER_MODULE_CONSTRUCTOR(be_pbqp_alloc);
-
#endif
projects / libfirm / blobdiff