X-Git-Url: http://nsz.repo.hu/git/?a=blobdiff_plain;f=ir%2Fbe%2Fbecopyheur4.c;h=4b903b20d34698af8d01c0da502592c0a05118b0;hb=6b1c4998aa018f971620bb1eb72898ee7b905371;hp=6818e3a331dba98391a798cd4752fd0a7cacf082;hpb=4f1f35735b0cad291e9fbf753636fa1bbb0b3009;p=libfirm

diff --git a/ir/be/becopyheur4.c b/ir/be/becopyheur4.c
index 6818e3a33..4b903b20d 100644
--- a/ir/be/becopyheur4.c
+++ b/ir/be/becopyheur4.c
@@ -1,113 +1,604 @@
-/**
- * This is the C implementation of the trivial mst algo
- * originally written in Java by Sebastian Hack.
- * Performs simple copy minimzation.
+/*
+ * Copyright (C) 1995-2008 University of Karlsruhe.  All right reserved.
+ *
+ * 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.
  *
- * @author Christian Wuerdig
- * @date   27.04.2007
- * @id     $Id$
+ * 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.
  */
-#if 0
 
-#ifdef HAVE_CONFIG_H
+/**
+ * @file
+ * @brief       Simple copy minimization heuristics.
+ * @author      Christian Wuerdig
+ * @date        27.04.2007
+ * @version     $Id$
+ *
+ * This is the C implementation of the mst algorithm
+ * originally written in Java by Sebastian Hack.
+ * (also known as "heur3" :)
+ * Performs simple copy minimization.
+ */
 #include "config.h"
-#endif /* HAVE_CONFIG_H */
+
+#define DISABLE_STATEV
+
+#include <float.h>
 
 #include "array.h"
-#include "irnode.h"
+#include "irnode_t.h"
 #include "bitset.h"
 #include "raw_bitset.h"
 #include "irphase_t.h"
+#include "pqueue.h"
+#include "xmalloc.h"
+#include "pdeq.h"
+#include "pset.h"
+#include "irprintf.h"
+#include "irbitset.h"
+#include "error.h"
+#include "list.h"
+#include "statev.h"
+
+#include "irbitset.h"
 
 #include "bearch.h"
 #include "beifg.h"
 #include "be_t.h"
+#include "becopyopt_t.h"
+#include "bemodule.h"
+
+
+#define COL_COST_INFEASIBLE       DBL_MAX
+#define AFF_NEIGHBOUR_FIX_BENEFIT 128.0
+#define NEIGHBOUR_CONSTR_COSTS    64.0
 
+
+#ifdef DEBUG_libfirm
+
+#define DBG_AFF_CHUNK(env, level, chunk) do { if (firm_dbg_get_mask(dbg) & (level)) dbg_aff_chunk((env), (chunk)); } while(0)
+#define DBG_COL_COST(env, level, cost)   do { if (firm_dbg_get_mask(dbg) & (level)) dbg_col_cost((env), (cost)); } while(0)
+
+static firm_dbg_module_t *dbg = NULL;
+
+#else
+
+#define DBG_AFF_CHUNK(env, level, chunk)
+#define DBG_COL_COST(env, level, cost)
+
+#endif
+
+typedef float real_t;
+#define REAL(C)   (C ## f)
+
+static unsigned last_chunk_id   = 0;
+static int recolor_limit        = 7;
+static real_t dislike_influence = REAL(0.1);
+
+typedef struct _col_cost_t {
+	int     col;
+	real_t  cost;
+} col_cost_t;
+
+/**
+ * An affinity chunk.
+ */
 typedef struct _aff_chunk_t {
-	ir_node **nodes;
-	int     weight;
+	const ir_node    **n;                   /**< An ARR_F containing all nodes of the chunk. */
+	const ir_node  **interfere;             /**< An ARR_F containing all inference. */
+	int              weight;                /**< Weight of this chunk */
+	unsigned         weight_consistent : 1; /**< Set if the weight is consistent. */
+	unsigned         deleted           : 1; /**< For debugging: Set if the was deleted. */
+	unsigned         id;                    /**< An id of this chunk. */
+	unsigned         visited;
+	col_cost_t       color_affinity[1];
 } aff_chunk_t;
 
+/**
+ * An affinity edge.
+ */
 typedef struct _aff_edge_t {
-	ir_node *src;
-	ir_node *tgt;
-	double  weight;
+	const ir_node *src;                   /**< Source node. */
+	const ir_node *tgt;                   /**< Target node. */
+	int           weight;                 /**< The weight of this edge. */
 } aff_edge_t;
 
-/* main coalescing environment*/
-typedef struct _co_mst_safe_env_t {
+/* main coalescing environment */
+typedef struct _co_mst_env_t {
 	int              n_regs;         /**< number of regs in class */
 	int              k;              /**< number of non-ignore registers in class */
 	bitset_t         *ignore_regs;   /**< set containing all global ignore registers */
 	ir_phase         ph;             /**< phase object holding data for nodes */
-	pqueue           *chunks;        /**< priority queue for chunks */
+	pqueue_t         *chunks;        /**< priority queue for chunks */
+	pset             *chunkset;      /**< set holding all chunks */
 	be_ifg_t         *ifg;           /**< the interference graph */
-	const arch_env_t *aenv;          /**< the arch environment */
-} co_mst_safe_env_t;
+	copy_opt_t       *co;            /**< the copy opt object */
+	unsigned         chunk_visited;
+	col_cost_t      **single_cols;
+} co_mst_env_t;
 
 /* stores coalescing related information for a node */
-typedef struct _co_mst_safe_irn_t {
-	ir_node     *irn;
-	aff_chunk_t *chunk;
-	bitset_t    *adm_colors;
-	int         init_col;
-	int         tmp_col;
-	int         int_neigh;
-	unsigned    fixed : 1;
-} co_mst_safe_irn_t;
-
-#define get_co_mst_safe_irn(mst_env, irn) (phase_get_or_set_irn_data(&(mst_env)->ph, (irn)))
-
-/* compares two affinity edges */
+typedef struct _co_mst_irn_t {
+	const ir_node    *irn;              /**< the irn this information belongs to */
+	aff_chunk_t      *chunk;            /**< the chunk this irn belongs to */
+	bitset_t         *adm_colors;       /**< set of admissible colors for this irn */
+	ir_node          **int_neighs;      /**< array of all interfering neighbours (cached for speed reasons) */
+	int              n_neighs;          /**< length of the interfering neighbours array. */
+	int              int_aff_neigh;     /**< number of interfering affinity neighbours */
+	int              col;               /**< color currently assigned */
+	int              init_col;          /**< the initial color */
+	int              tmp_col;           /**< a temporary assigned color */
+	unsigned         fixed     : 1;     /**< the color is fixed */
+	struct list_head list;              /**< Queue for coloring undo. */
+	real_t           constr_factor;
+} co_mst_irn_t;
+
+#define get_co_mst_irn(mst_env, irn) (phase_get_or_set_irn_data(&(mst_env)->ph, (irn)))
+
+typedef int decide_func_t(const co_mst_irn_t *node, int col);
+
+#ifdef DEBUG_libfirm
+
+/**
+ * Write a chunk to stderr for debugging.
+ */
+static void dbg_aff_chunk(const co_mst_env_t *env, const aff_chunk_t *c) {
+	int i, l;
+	(void) env;
+	if (c->weight_consistent)
+		ir_fprintf(stderr, " $%d ", c->weight);
+	ir_fprintf(stderr, "{");
+	for (i = 0, l = ARR_LEN(c->n); i < l; ++i) {
+		const ir_node *n = c->n[i];
+		ir_fprintf(stderr, " %+F,", n);
+	}
+	ir_fprintf(stderr, "}");
+}
+
+/**
+ * Dump all admissible colors to stderr.
+ */
+static void dbg_admissible_colors(const co_mst_env_t *env, const co_mst_irn_t *node) {
+	bitset_pos_t idx;
+	(void) env;
+
+	if (bitset_popcnt(node->adm_colors) < 1)
+		fprintf(stderr, "no admissible colors?!?");
+	else {
+		bitset_foreach(node->adm_colors, idx) {
+			fprintf(stderr, " %d", idx);
+		}
+	}
+}
+
+/**
+ * Dump color-cost pairs to stderr.
+ */
+static void dbg_col_cost(const co_mst_env_t *env, const col_cost_t *cost) {
+	int i;
+	for (i = 0; i < env->n_regs; ++i)
+		fprintf(stderr, " (%d, %.4f)", cost[i].col, cost[i].cost);
+}
+
+#endif /* DEBUG_libfirm */
+
+static inline int get_mst_irn_col(const co_mst_irn_t *node) {
+	return node->tmp_col >= 0 ? node->tmp_col : node->col;
+}
+
+/**
+ * @return 1 if node @p node has color @p col, 0 otherwise.
+ */
+static int decider_has_color(const co_mst_irn_t *node, int col) {
+	return get_mst_irn_col(node) == col;
+}
+
+/**
+ * @return 1 if node @p node has not color @p col, 0 otherwise.
+ */
+static int decider_hasnot_color(const co_mst_irn_t *node, int col) {
+	return get_mst_irn_col(node) != col;
+}
+
+/**
+ * Always returns true.
+ */
+static int decider_always_yes(const co_mst_irn_t *node, int col) {
+	(void) node;
+	(void) col;
+	return 1;
+}
+
+/** compares two affinity edges by its weight */
 static int cmp_aff_edge(const void *a, const void *b) {
-	const aff_edge_t * const *e1 = d1;
-	const aff_edge_t * const *e2 = d2;
+	const aff_edge_t *e1 = a;
+	const aff_edge_t *e2 = b;
 
+	if (e2->weight == e1->weight) {
+		if (e2->src->node_idx == e1->src->node_idx)
+			return QSORT_CMP(e2->tgt->node_idx, e1->tgt->node_idx);
+		else
+			return QSORT_CMP(e2->src->node_idx, e1->src->node_idx);
+	}
 	/* sort in descending order */
-	return (*e1)->weight < (*e2)->weight ? 1 : -1;
+	return QSORT_CMP(e2->weight, e1->weight);
+}
+
+/** compares to color-cost pairs */
+static __attribute__((unused)) int cmp_col_cost_lt(const void *a, const void *b) {
+	const col_cost_t *c1 = a;
+	const col_cost_t *c2 = b;
+	real_t diff = c1->cost - c2->cost;
+	return (diff > 0) - (diff < 0);
+}
+
+static int cmp_col_cost_gt(const void *a, const void *b) {
+	const col_cost_t *c1 = a;
+	const col_cost_t *c2 = b;
+	real_t diff = c2->cost - c1->cost;
+	return (diff > 0) - (diff < 0);
+}
+
+/**
+ * Creates a new affinity chunk
+ */
+static inline aff_chunk_t *new_aff_chunk(co_mst_env_t *env) {
+	aff_chunk_t *c = XMALLOCF(aff_chunk_t, color_affinity, env->n_regs);
+	c->n                 = NEW_ARR_F(const ir_node *, 0);
+	c->interfere         = NEW_ARR_F(const ir_node *, 0);
+	c->weight            = -1;
+	c->weight_consistent = 0;
+	c->deleted           = 0;
+	c->id                = ++last_chunk_id;
+	c->visited           = 0;
+	pset_insert(env->chunkset, c, c->id);
+	return c;
+}
+
+/**
+ * Frees all memory allocated by an affinity chunk.
+ */
+static inline void delete_aff_chunk(co_mst_env_t *env, aff_chunk_t *c) {
+	pset_remove(env->chunkset, c, c->id);
+	DEL_ARR_F(c->interfere);
+	DEL_ARR_F(c->n);
+	c->deleted = 1;
+	free(c);
+}
+
+/**
+ * binary search of sorted nodes.
+ *
+ * @return the position where n is found in the array arr or ~pos
+ * if the nodes is not here.
+ */
+static inline int nodes_bsearch(const ir_node **arr, const ir_node *n) {
+	int hi = ARR_LEN(arr);
+	int lo = 0;
+
+	while (lo < hi) {
+		int md = lo + ((hi - lo) >> 1);
+
+		if (arr[md] == n)
+			return md;
+		if (arr[md] < n)
+			lo = md + 1;
+		else
+			hi = md;
+	}
+
+	return ~lo;
+}
+
+/** Check if a node n can be found inside arr. */
+static int node_contains(const ir_node **arr, const ir_node *n) {
+	int i = nodes_bsearch(arr, n);
+	return i >= 0;
+}
+
+/**
+ * Insert a node into the sorted nodes list.
+ *
+ * @return 1 if the node was inserted, 0 else
+ */
+static int nodes_insert(const ir_node ***arr, const ir_node *irn) {
+	int idx = nodes_bsearch(*arr, irn);
+
+	if (idx < 0) {
+		int i, n = ARR_LEN(*arr);
+		const ir_node **l;
+
+		ARR_APP1(const ir_node *, *arr, irn);
+
+		/* move it */
+		idx = ~idx;
+		l = *arr;
+		for (i = n - 1; i >= idx; --i)
+			l[i + 1] = l[i];
+		l[idx] = irn;
+		return 1;
+	}
+	return 0;
+}
+
+/**
+ * Adds a node to an affinity chunk
+ */
+static inline void aff_chunk_add_node(aff_chunk_t *c, co_mst_irn_t *node) {
+	int i;
+
+	if (! nodes_insert(&c->n, node->irn))
+		return;
+
+	c->weight_consistent = 0;
+	node->chunk          = c;
+
+	for (i = node->n_neighs - 1; i >= 0; --i) {
+		ir_node *neigh = node->int_neighs[i];
+		nodes_insert(&c->interfere, neigh);
+	}
 }
 
 /**
  * In case there is no phase information for irn, initialize it.
  */
-static void *co_mst_safe_irn_init(ir_phase *ph, ir_node *irn, void *old) {
-	co_mst_safe_irn_t *res = old ? old : phase_alloc(ph, sizeof(res[0]));
-	co_mst_safe_env_t *env = ph->priv;
+static void *co_mst_irn_init(ir_phase *ph, const ir_node *irn, void *old) {
+	co_mst_irn_t *res = old ? old : phase_alloc(ph, sizeof(res[0]));
+	co_mst_env_t *env = ph->priv;
 
-	if (res != old) {
-		void                      *neigh_it = be_ifg_neighbours_iter_alloca(env->ifg);
+	if (!old) {
 		const arch_register_req_t *req;
-		ir_node                   *m;
+		void     *nodes_it = be_ifg_nodes_iter_alloca(env->ifg);
+		ir_node  *neigh;
+		unsigned len;
+
+		res->irn           = irn;
+		res->chunk         = NULL;
+		res->fixed         = 0;
+		res->tmp_col       = -1;
+		res->int_neighs    = NULL;
+		res->int_aff_neigh = 0;
+		res->col           = arch_register_get_index(arch_get_irn_register(irn));
+		res->init_col      = res->col;
+		INIT_LIST_HEAD(&res->list);
 
-		res->irn       = irn;
-		res->chunk     = NULL;
-		res->fixed     = 0;
-		res->tmp_col   = -1;
-		res->int_neigh = 0;
-		res->init_col  = arch_register_get_index(arch_get_irn_register(env->aenv, irn));
+		DB((dbg, LEVEL_4, "Creating phase info for %+F\n", irn));
 
 		/* set admissible registers */
 		res->adm_colors = bitset_obstack_alloc(phase_obst(ph), env->n_regs);
 
 		/* Exclude colors not assignable to the irn */
-		req = arch_get_register_req(env->aenv, irn, -1);
+		req = arch_get_register_req_out(irn);
 		if (arch_register_req_is(req, limited))
 			rbitset_copy_to_bitset(req->limited, res->adm_colors);
+		else
+			bitset_set_all(res->adm_colors);
 
 		/* exclude global ignore registers as well */
 		bitset_andnot(res->adm_colors, env->ignore_regs);
 
-		/* calculate the number of interfering neigbours */
-		be_ifg_foreach_neighbour(env->ifg, neigh_it, irn, m) {
-			if (! arch_irn_is(env->aenv, m, ignore))
-				res->int_neigh++;
+		/* compute the constraint factor */
+		res->constr_factor = (real_t) (1 + env->n_regs - bitset_popcnt(res->adm_colors)) / env->n_regs;
+
+		/* set the number of interfering affinity neighbours to -1, they are calculated later */
+		res->int_aff_neigh = -1;
+
+		/* build list of interfering neighbours */
+		len = 0;
+		be_ifg_foreach_neighbour(env->ifg, nodes_it, irn, neigh) {
+			if (!arch_irn_is_ignore(neigh)) {
+				obstack_ptr_grow(phase_obst(ph), neigh);
+				++len;
+			}
 		}
+		res->int_neighs = obstack_finish(phase_obst(ph));
+		res->n_neighs   = len;
+	}
+	return res;
+}
+
+/**
+ * Check if affinity chunk @p chunk interferes with node @p irn.
+ */
+static inline int aff_chunk_interferes(const aff_chunk_t *chunk, const ir_node *irn) {
+	return node_contains(chunk->interfere, irn);
+}
+
+/**
+ * Check if there are interference edges from c1 to c2.
+ * @param c1    A chunk
+ * @param c2    Another chunk
+ * @return 1 if there are interferences between nodes of c1 and c2, 0 otherwise.
+ */
+static inline int aff_chunks_interfere(const aff_chunk_t *c1, const aff_chunk_t *c2) {
+	int i;
+
+	if (c1 == c2)
+		return 0;
+
+	/* check if there is a node in c2 having an interfering neighbor in c1 */
+	for (i = ARR_LEN(c2->n) - 1; i >= 0; --i) {
+		const ir_node *irn = c2->n[i];
+
+		if (node_contains(c1->interfere, irn))
+			return 1;
+	}
+	return 0;
+}
+
+/**
+ * Returns the affinity chunk of @p irn or creates a new
+ * one with @p irn as element if there is none assigned.
+ */
+static inline aff_chunk_t *get_aff_chunk(co_mst_env_t *env, const ir_node *irn) {
+	co_mst_irn_t *node = get_co_mst_irn(env, irn);
+	return node->chunk;
+}
+
+/**
+ * Let chunk(src) absorb the nodes of chunk(tgt) (only possible when there
+ * are no interference edges from chunk(src) to chunk(tgt)).
+ * @return 1 if successful, 0 if not possible
+ */
+static int aff_chunk_absorb(co_mst_env_t *env, const ir_node *src, const ir_node *tgt) {
+	aff_chunk_t *c1 = get_aff_chunk(env, src);
+	aff_chunk_t *c2 = get_aff_chunk(env, tgt);
+
+#ifdef DEBUG_libfirm
+		DB((dbg, LEVEL_4, "Attempt to let c1 (id %u): ", c1 ? c1->id : 0));
+		if (c1) {
+			DBG_AFF_CHUNK(env, LEVEL_4, c1);
+		} else {
+			DB((dbg, LEVEL_4, "{%+F}", src));
+		}
+		DB((dbg, LEVEL_4, "\n\tabsorb c2 (id %u): ", c2 ? c2->id : 0));
+		if (c2) {
+			DBG_AFF_CHUNK(env, LEVEL_4, c2);
+		} else {
+			DB((dbg, LEVEL_4, "{%+F}", tgt));
+		}
+		DB((dbg, LEVEL_4, "\n"));
+#endif
+
+	if (c1 == NULL) {
+		if (c2 == NULL) {
+			/* no chunk exists */
+			co_mst_irn_t *mirn = get_co_mst_irn(env, src);
+			int i;
+
+			for (i = mirn->n_neighs - 1; i >= 0; --i) {
+				if (mirn->int_neighs[i] == tgt)
+					break;
+			}
+			if (i < 0) {
+				/* create one containing both nodes */
+				c1 = new_aff_chunk(env);
+				aff_chunk_add_node(c1, get_co_mst_irn(env, src));
+				aff_chunk_add_node(c1, get_co_mst_irn(env, tgt));
+				goto absorbed;
+			}
+		} else {
+			/* c2 already exists */
+			if (! aff_chunk_interferes(c2, src)) {
+				aff_chunk_add_node(c2, get_co_mst_irn(env, src));
+				goto absorbed;
+			}
+		}
+	} else if (c2 == NULL) {
+		/* c1 already exists */
+		if (! aff_chunk_interferes(c1, tgt)) {
+			aff_chunk_add_node(c1, get_co_mst_irn(env, tgt));
+			goto absorbed;
+		}
+	} else if (c1 != c2 && ! aff_chunks_interfere(c1, c2)) {
+		int idx, len;
+
+		for (idx = 0, len = ARR_LEN(c2->n); idx < len; ++idx)
+			aff_chunk_add_node(c1, get_co_mst_irn(env, c2->n[idx]));
+
+		for (idx = 0, len = ARR_LEN(c2->interfere); idx < len; ++idx) {
+			const ir_node *irn = c2->interfere[idx];
+			nodes_insert(&c1->interfere, irn);
+		}
+
+		c1->weight_consistent = 0;
+
+		delete_aff_chunk(env, c2);
+		goto absorbed;
+	}
+	DB((dbg, LEVEL_4, " ... c1 interferes with c2, skipped\n"));
+	return 0;
+
+absorbed:
+	DB((dbg, LEVEL_4, " ... absorbed\n"));
+	return 1;
+}
+
+/**
+ * Assures that the weight of the given chunk is consistent.
+ */
+static void aff_chunk_assure_weight(co_mst_env_t *env, aff_chunk_t *c) {
+	if (! c->weight_consistent) {
+		int w = 0;
+		int idx, len, i;
+
+		for (i = 0; i < env->n_regs; ++i) {
+			c->color_affinity[i].col = i;
+			c->color_affinity[i].cost = REAL(0.0);
+		}
+
+		for (idx = 0, len = ARR_LEN(c->n); idx < len; ++idx) {
+			const ir_node         *n       = c->n[idx];
+			const affinity_node_t *an      = get_affinity_info(env->co, n);
+			co_mst_irn_t          *node    = get_co_mst_irn(env, n);
+
+			node->chunk = c;
+			if (node->constr_factor > REAL(0.0)) {
+				bitset_pos_t col;
+				bitset_foreach (node->adm_colors, col)
+					c->color_affinity[col].cost += node->constr_factor;
+			}
+
+			if (an != NULL) {
+				neighb_t *neigh;
+				co_gs_foreach_neighb(an, neigh) {
+					const ir_node *m = neigh->irn;
+
+					if (arch_irn_is_ignore(m))
+						continue;
+
+					w += node_contains(c->n, m) ? neigh->costs : 0;
+				}
+			}
+		}
+
+		for (i = 0; i < env->n_regs; ++i)
+			c->color_affinity[i].cost *= (REAL(1.0) / ARR_LEN(c->n));
 
+		c->weight            = w;
+		// c->weight            = bitset_popcnt(c->nodes);
+		c->weight_consistent = 1;
 	}
+}
+
+/**
+ * Count the number of interfering affinity neighbours
+ */
+static int count_interfering_aff_neighs(co_mst_env_t *env, const affinity_node_t *an) {
+	const neighb_t     *neigh;
+	const ir_node      *irn  = an->irn;
+	const co_mst_irn_t *node = get_co_mst_irn(env, irn);
+	int                res   = 0;
+
+	co_gs_foreach_neighb(an, neigh) {
+		const ir_node *n = neigh->irn;
+		int           i;
 
+		if (arch_irn_is_ignore(n))
+			continue;
+
+		/* check if the affinity neighbour interfere */
+		for (i = 0; i < node->n_neighs; ++i) {
+			if (node->int_neighs[i] == n) {
+				++res;
+				break;
+			}
+		}
+	}
 	return res;
 }
 
+
 /**
  * Build chunks of nodes connected by affinity edges.
  * We start at the heaviest affinity edge.
@@ -115,91 +606,889 @@ static void *co_mst_safe_irn_init(ir_phase *ph, ir_node *irn, void *old) {
  * merged if there are no interference edges from one
  * chunk to the other.
  */
-static void build_affinity_chunks(co_mst_safe_env_t *env) {
-	void       *nodes_it = be_ifg_nodes_iter_alloca(env->ifg);
-	aff_edge_t **edges   = NEW_ARR_F(aff_edge_t *, 0);
-	int        i;
+static void build_affinity_chunks(co_mst_env_t *env) {
+	void        *nodes_it = be_ifg_nodes_iter_alloca(env->ifg);
+	aff_edge_t  *edges    = NEW_ARR_F(aff_edge_t, 0);
+	ir_node     *n;
+	int         i, len;
+	aff_chunk_t *curr_chunk;
 
 	/* at first we create the affinity edge objects */
 	be_ifg_foreach_node(env->ifg, nodes_it, n) {
-		int               n_idx = get_irn_idx(n);
-		co_mst_safe_irn_t *n1;
-		affinity_node_t   *an;
+		int             n_idx = get_irn_idx(n);
+		co_mst_irn_t    *n1;
+		affinity_node_t *an;
 
-		if (arch_irn_is(env->aenv, n, ignore))
+		if (arch_irn_is_ignore(n))
 			continue;
 
-		n1 = get_co_mst_safe_irn(env, n);
-		an = get_affinity_info(co, n);
+		n1 = get_co_mst_irn(env, n);
+		an = get_affinity_info(env->co, n);
 
 		if (an != NULL) {
 			neighb_t *neigh;
-			co_gs_foreach_neighb(an, neigh) {
-				ir_node           *m    = neigh->irn;
-				int               m_idx = get_irn_idx(m);
-				co_mst_safe_irn_t *n2;
 
-				if (arch_irn_is(env->aenv, m, ignore))
-					continue;
+			if (n1->int_aff_neigh < 0)
+				n1->int_aff_neigh = count_interfering_aff_neighs(env, an);
 
-				n2 = get_co_mst_safe_irn(env, m);
+			/* build the affinity edges */
+			co_gs_foreach_neighb(an, neigh) {
+				const ir_node *m     = neigh->irn;
+				int            m_idx = get_irn_idx(m);
 
-				/* record the edge in only one direction and only to non-ignore nodes */
+				/* record the edge in only one direction */
 				if (n_idx < m_idx) {
-					aff_edge_t *edge = phase_alloc(&env->ph, sizeof(*edge));
+					co_mst_irn_t *n2;
+					aff_edge_t   edge;
+
+					/* skip ignore nodes */
+					if (arch_irn_is_ignore(m))
+						continue;
+
+					edge.src = n;
+					edge.tgt = m;
 
-					edge->src    = n;
-					edge->tgt    = m;
-					edge->weight = (double)neigh->cost / (double)(1 + n1->int_neigh + n2->int_neigh);
-					ARR_APP1(aff_edge_t *, edges, edge);
+					n2 = get_co_mst_irn(env, m);
+					if (n2->int_aff_neigh < 0) {
+						affinity_node_t *am = get_affinity_info(env->co, m);
+						n2->int_aff_neigh = count_interfering_aff_neighs(env, am);
+					}
+					/*
+					 * these weights are pure hackery ;-).
+					 * It's not chriswue's fault but mine.
+					 */
+					edge.weight = neigh->costs;
+					ARR_APP1(aff_edge_t, edges, edge);
 				}
 			}
 		}
 	}
 
-	/* now: sort edges and build the chunks */
-	qsort(edges, ARR_LEN(edges), sizeof(edges[0]), cmp_aff_edge);
-	for (i = 0; i < ARR_LEN(edges); ++i) {
+	/* now: sort edges and build the affinity chunks */
+	len = ARR_LEN(edges);
+	qsort(edges, len, sizeof(edges[0]), cmp_aff_edge);
+	for (i = 0; i < len; ++i) {
+		DBG((dbg, LEVEL_1, "edge (%u,%u) %f\n", edges[i].src->node_idx, edges[i].tgt->node_idx, edges[i].weight));
+
+		(void)aff_chunk_absorb(env, edges[i].src, edges[i].tgt);
+	}
+
+	/* now insert all chunks into a priority queue */
+	foreach_pset(env->chunkset, curr_chunk) {
+		aff_chunk_assure_weight(env, curr_chunk);
+
+		DBG((dbg, LEVEL_1, "entry #%u", curr_chunk->id));
+		DBG_AFF_CHUNK(env, LEVEL_1, curr_chunk);
+		DBG((dbg, LEVEL_1, "\n"));
+
+		pqueue_put(env->chunks, curr_chunk, curr_chunk->weight);
+	}
+
+	foreach_phase_irn(&env->ph, n) {
+		co_mst_irn_t *mirn = get_co_mst_irn(env, n);
+
+		if (mirn->chunk == NULL) {
+			/* no chunk is allocated so far, do it now */
+			aff_chunk_t *curr_chunk = new_aff_chunk(env);
+			aff_chunk_add_node(curr_chunk, mirn);
+
+			aff_chunk_assure_weight(env, curr_chunk);
 
+			DBG((dbg, LEVEL_1, "entry #%u", curr_chunk->id));
+			DBG_AFF_CHUNK(env, LEVEL_1, curr_chunk);
+			DBG((dbg, LEVEL_1, "\n"));
+
+			pqueue_put(env->chunks, curr_chunk, curr_chunk->weight);
+		}
 	}
 
 	DEL_ARR_F(edges);
 }
 
+static __attribute__((unused)) void chunk_order_nodes(co_mst_env_t *env, aff_chunk_t *chunk)
+{
+	pqueue_t *grow = new_pqueue();
+	const ir_node *max_node = NULL;
+	int max_weight = 0;
+	int i;
+
+	for (i = ARR_LEN(chunk->n) - 1; i >= 0; i--) {
+		const ir_node   *irn = chunk->n[i];
+		affinity_node_t *an  = get_affinity_info(env->co, irn);
+		int w = 0;
+		neighb_t *neigh;
+
+		if (arch_irn_is_ignore(irn))
+			continue;
+
+		if (an) {
+			co_gs_foreach_neighb(an, neigh)
+				w += neigh->costs;
+
+			if (w > max_weight) {
+				max_weight = w;
+				max_node   = irn;
+			}
+		}
+	}
+
+	if (max_node) {
+		bitset_t *visited = bitset_irg_malloc(env->co->irg);
+
+		for (i = ARR_LEN(chunk->n) - 1; i >= 0; --i)
+			bitset_add_irn(visited, chunk->n[i]);
+
+		pqueue_put(grow, (void *) max_node, max_weight);
+		bitset_remv_irn(visited, max_node);
+		i = 0;
+		while (!pqueue_empty(grow)) {
+			ir_node *irn = pqueue_pop_front(grow);
+			affinity_node_t *an = get_affinity_info(env->co, irn);
+			neighb_t        *neigh;
+
+			if (arch_irn_is_ignore(irn))
+				continue;
+
+			assert(i <= ARR_LEN(chunk->n));
+			chunk->n[i++] = irn;
+
+			assert(an);
+
+			/* build the affinity edges */
+			co_gs_foreach_neighb(an, neigh) {
+				co_mst_irn_t *node = get_co_mst_irn(env, neigh->irn);
+
+				if (bitset_contains_irn(visited, node->irn)) {
+					pqueue_put(grow, (void *) neigh->irn, neigh->costs);
+					bitset_remv_irn(visited, node->irn);
+				}
+			}
+		}
+
+		del_pqueue(grow);
+		bitset_free(visited);
+	}
+}
+
 /**
- * Main driver for mst safe coalescing algorithm.
+ * Greedy collect affinity neighbours into thew new chunk @p chunk starting at node @p node.
+ */
+static void expand_chunk_from(co_mst_env_t *env, co_mst_irn_t *node, bitset_t *visited,
+	aff_chunk_t *chunk, aff_chunk_t *orig_chunk, decide_func_t *decider, int col)
+{
+	waitq *nodes = new_waitq();
+
+	DBG((dbg, LEVEL_1, "\n\tExpanding new chunk (#%u) from %+F, color %d:", chunk->id, node->irn, col));
+
+	/* init queue and chunk */
+	waitq_put(nodes, node);
+	bitset_set(visited, get_irn_idx(node->irn));
+	aff_chunk_add_node(chunk, node);
+	DB((dbg, LEVEL_1, " %+F", node->irn));
+
+	/* as long as there are nodes in the queue */
+	while (! waitq_empty(nodes)) {
+		co_mst_irn_t    *n  = waitq_get(nodes);
+		affinity_node_t *an = get_affinity_info(env->co, n->irn);
+
+		/* check all affinity neighbors */
+		if (an != NULL) {
+			neighb_t *neigh;
+			co_gs_foreach_neighb(an, neigh) {
+				const ir_node *m    = neigh->irn;
+				int            m_idx = get_irn_idx(m);
+				co_mst_irn_t *n2;
+
+				if (arch_irn_is_ignore(m))
+					continue;
+
+				n2 = get_co_mst_irn(env, m);
+
+				if (! bitset_is_set(visited, m_idx)  &&
+					decider(n2, col)                 &&
+					! n2->fixed                      &&
+					! aff_chunk_interferes(chunk, m) &&
+					node_contains(orig_chunk->n, m))
+				{
+					/*
+						following conditions are met:
+						- neighbour is not visited
+						- neighbour likes the color
+						- neighbour has not yet a fixed color
+						- the new chunk doesn't interfere with the neighbour
+						- neighbour belongs or belonged once to the original chunk
+					*/
+					bitset_set(visited, m_idx);
+					aff_chunk_add_node(chunk, n2);
+					DB((dbg, LEVEL_1, " %+F", n2->irn));
+					/* enqueue for further search */
+					waitq_put(nodes, n2);
+				}
+			}
+		}
+	}
+
+	DB((dbg, LEVEL_1, "\n"));
+
+	del_waitq(nodes);
+}
+
+/**
+ * Fragment the given chunk into chunks having given color and not having given color.
+ */
+static aff_chunk_t *fragment_chunk(co_mst_env_t *env, int col, aff_chunk_t *c, waitq *tmp) {
+	bitset_t    *visited = bitset_irg_malloc(env->co->irg);
+	int         idx, len;
+	aff_chunk_t *best = NULL;
+
+	for (idx = 0, len = ARR_LEN(c->n); idx < len; ++idx) {
+		const ir_node *irn;
+		co_mst_irn_t  *node;
+		aff_chunk_t   *tmp_chunk;
+		decide_func_t *decider;
+		int           check_for_best;
+
+		irn = c->n[idx];
+		if (bitset_is_set(visited, get_irn_idx(irn)))
+			continue;
+
+		node = get_co_mst_irn(env, irn);
+
+		if (get_mst_irn_col(node) == col) {
+			decider        = decider_has_color;
+			check_for_best = 1;
+			DBG((dbg, LEVEL_4, "\tcolor %d wanted\n", col));
+		}
+		else {
+			decider        = decider_hasnot_color;
+			check_for_best = 0;
+			DBG((dbg, LEVEL_4, "\tcolor %d forbidden\n", col));
+		}
+
+		/* create a new chunk starting at current node */
+		tmp_chunk = new_aff_chunk(env);
+		waitq_put(tmp, tmp_chunk);
+		expand_chunk_from(env, node, visited, tmp_chunk, c, decider, col);
+		assert(ARR_LEN(tmp_chunk->n) > 0 && "No nodes added to chunk");
+
+		/* remember the local best */
+		aff_chunk_assure_weight(env, tmp_chunk);
+		if (check_for_best && (! best || best->weight < tmp_chunk->weight))
+			best = tmp_chunk;
+	}
+
+	assert(best && "No chunk found?");
+	bitset_free(visited);
+	return best;
+}
+
+/**
+ * Resets the temporary fixed color of all nodes within wait queue @p nodes.
+ * ATTENTION: the queue is empty after calling this function!
  */
-int co_solve_heuristic_mst_safe(copy_opt_t *co)
+static inline void reject_coloring(struct list_head *nodes) {
+	co_mst_irn_t *n, *temp;
+	DB((dbg, LEVEL_4, "\treject coloring for"));
+	list_for_each_entry_safe(co_mst_irn_t, n, temp, nodes, list) {
+		DB((dbg, LEVEL_4, " %+F", n->irn));
+		assert(n->tmp_col >= 0);
+		n->tmp_col = -1;
+		list_del_init(&n->list);
+	}
+	DB((dbg, LEVEL_4, "\n"));
+}
+
+static inline void materialize_coloring(struct list_head *nodes) {
+	co_mst_irn_t *n, *temp;
+	list_for_each_entry_safe(co_mst_irn_t, n, temp, nodes, list) {
+		assert(n->tmp_col >= 0);
+		n->col     = n->tmp_col;
+		n->tmp_col = -1;
+		list_del_init(&n->list);
+	}
+}
+
+static inline void set_temp_color(co_mst_irn_t *node, int col, struct list_head *changed)
+{
+	assert(col >= 0);
+	assert(!node->fixed);
+	assert(node->tmp_col < 0);
+	assert(node->list.next == &node->list && node->list.prev == &node->list);
+	assert(bitset_is_set(node->adm_colors, col));
+
+	list_add_tail(&node->list, changed);
+	node->tmp_col = col;
+}
+
+static inline int is_loose(co_mst_irn_t *node)
 {
-	be_ifg_t *ifg         = co->cenv->ifg;
-	unsigned n_regs       = co->cenv->cls->n_regs;
-	bitset_t *ignore_regs = bitset_alloca(n_regs);
-	unsigned k;
-	co_mst_safe_env_t mst_env;
+	return !node->fixed && node->tmp_col < 0;
+}
+
+/**
+ * Determines the costs for each color if it would be assigned to node @p node.
+ */
+static void determine_color_costs(co_mst_env_t *env, co_mst_irn_t *node, col_cost_t *costs)
+{
+	int   *neigh_cols = ALLOCAN(int, env->n_regs);
+	int    n_loose    = 0;
+	real_t coeff;
+	int    i;
+
+	for (i = 0; i < env->n_regs; ++i) {
+		neigh_cols[i] = 0;
+		costs[i].col = i;
+		costs[i].cost = bitset_is_set(node->adm_colors, i) ? node->constr_factor : REAL(0.0);
+	}
+
+	for (i = 0; i < node->n_neighs; ++i) {
+		co_mst_irn_t *n = get_co_mst_irn(env, node->int_neighs[i]);
+		int col = get_mst_irn_col(n);
+		if (is_loose(n)) {
+			++n_loose;
+			++neigh_cols[col];
+		} else
+			costs[col].cost = REAL(0.0);
+	}
+
+	if (n_loose > 0) {
+		coeff = REAL(1.0) / n_loose;
+		for (i = 0; i < env->n_regs; ++i)
+			costs[i].cost *= REAL(1.0) - coeff * neigh_cols[i];
+	}
+}
+
+/* need forward declaration due to recursive call */
+static int recolor_nodes(co_mst_env_t *env, co_mst_irn_t *node, col_cost_t *costs, struct list_head *changed_ones, int depth, int *max_depth, int *trip);
+
+/**
+ * Tries to change node to a color but @p explude_col.
+ * @return 1 if succeeded, 0 otherwise.
+ */
+static int change_node_color_excluded(co_mst_env_t *env, co_mst_irn_t *node, int exclude_col, struct list_head *changed, int depth, int *max_depth, int *trip) {
+	int col = get_mst_irn_col(node);
+	int res = 0;
+
+	/* neighbours has already a different color -> good, temporary fix it */
+	if (col != exclude_col) {
+		if (is_loose(node))
+			set_temp_color(node, col, changed);
+		return 1;
+	}
+
+	/* The node has the color it should not have _and_ has not been visited yet. */
+	if (is_loose(node)) {
+		col_cost_t *costs = ALLOCAN(col_cost_t, env->n_regs);
+
+		/* Get the costs for giving the node a specific color. */
+		determine_color_costs(env, node, costs);
+
+		/* Since the node must not have the not_col, set the costs for that color to "infinity" */
+		costs[exclude_col].cost = REAL(0.0);
+
+		/* sort the colors according costs, cheapest first. */
+		qsort(costs, env->n_regs, sizeof(costs[0]), cmp_col_cost_gt);
+
+		/* Try recoloring the node using the color list. */
+		res = recolor_nodes(env, node, costs, changed, depth + 1, max_depth, trip);
+	}
+
+	return res;
+}
+
+/**
+ * Tries to bring node @p node to cheapest color and color all interfering neighbours with other colors.
+ * ATTENTION: Expect @p costs already sorted by increasing costs.
+ * @return 1 if coloring could be applied, 0 otherwise.
+ */
+static int recolor_nodes(co_mst_env_t *env, co_mst_irn_t *node, col_cost_t *costs, struct list_head *changed, int depth, int *max_depth, int *trip) {
+	int   i;
+	struct list_head local_changed;
+
+	++*trip;
+	if (depth > *max_depth)
+		*max_depth = depth;
+
+	DBG((dbg, LEVEL_4, "\tRecoloring %+F with color-costs", node->irn));
+	DBG_COL_COST(env, LEVEL_4, costs);
+	DB((dbg, LEVEL_4, "\n"));
+
+	if (depth >= recolor_limit) {
+		DBG((dbg, LEVEL_4, "\tHit recolor limit\n"));
+		return 0;
+	}
+
+	for (i = 0; i < env->n_regs; ++i) {
+		int tgt_col  = costs[i].col;
+		int neigh_ok = 1;
+		int j;
+
+		/* If the costs for that color (and all successive) are infinite, bail out we won't make it anyway. */
+		if (costs[i].cost == REAL(0.0)) {
+			DBG((dbg, LEVEL_4, "\tAll further colors forbidden\n"));
+			return 0;
+		}
+
+		/* Set the new color of the node and mark the node as temporarily fixed. */
+		assert(node->tmp_col < 0 && "Node must not have been temporary fixed.");
+		INIT_LIST_HEAD(&local_changed);
+		set_temp_color(node, tgt_col, &local_changed);
+		DBG((dbg, LEVEL_4, "\tTemporary setting %+F to color %d\n", node->irn, tgt_col));
+
+		/* try to color all interfering neighbours with current color forbidden */
+		for (j = 0; j < node->n_neighs; ++j) {
+			co_mst_irn_t *nn;
+			ir_node      *neigh;
+
+			neigh = node->int_neighs[j];
+
+			if (arch_irn_is_ignore(neigh))
+				continue;
+
+			nn = get_co_mst_irn(env, neigh);
+			DB((dbg, LEVEL_4, "\tHandling neighbour %+F, at position %d (fixed: %d, tmp_col: %d, col: %d)\n",
+				neigh, j, nn->fixed, nn->tmp_col, nn->col));
+
+			/*
+				Try to change the color of the neighbor and record all nodes which
+				get changed in the tmp list. Add this list to the "changed" list for
+				that color. If we did not succeed to change the color of the neighbor,
+				we bail out and try the next color.
+			*/
+			if (get_mst_irn_col(nn) == tgt_col) {
+				/* try to color neighbour with tgt_col forbidden */
+				neigh_ok = change_node_color_excluded(env, nn, tgt_col, &local_changed, depth + 1, max_depth, trip);
+
+				if (!neigh_ok)
+					break;
+			}
+		}
+
+		/*
+			We managed to assign the target color to all neighbors, so from the perspective
+			of the current node, every thing was ok and we can return safely.
+		*/
+		if (neigh_ok) {
+			/* append the local_changed ones to global ones */
+			list_splice(&local_changed, changed);
+			return 1;
+		}
+		else {
+			/* coloring of neighbours failed, so we try next color */
+			reject_coloring(&local_changed);
+		}
+	}
+
+	DBG((dbg, LEVEL_4, "\tAll colors failed\n"));
+	return 0;
+}
+
+/**
+ * Tries to bring node @p node and all it's neighbours to color @p tgt_col.
+ * @return 1 if color @p col could be applied, 0 otherwise
+ */
+static int change_node_color(co_mst_env_t *env, co_mst_irn_t *node, int tgt_col, struct list_head *changed) {
+	int col = get_mst_irn_col(node);
+
+	/* if node already has the target color -> good, temporary fix it */
+	if (col == tgt_col) {
+		DBG((dbg, LEVEL_4, "\t\tCNC: %+F has already color %d, fix temporary\n", node->irn, tgt_col));
+		if (is_loose(node))
+			set_temp_color(node, tgt_col, changed);
+		return 1;
+	}
+
+	/*
+		Node has not yet a fixed color and target color is admissible
+		-> try to recolor node and it's affinity neighbours
+	*/
+	if (is_loose(node) && bitset_is_set(node->adm_colors, tgt_col)) {
+		col_cost_t *costs = env->single_cols[tgt_col];
+		int res, max_depth, trip;
+
+		max_depth = 0;
+		trip      = 0;
+
+		DBG((dbg, LEVEL_4, "\t\tCNC: Attempt to recolor %+F ===>>\n", node->irn));
+		res = recolor_nodes(env, node, costs, changed, 0, &max_depth, &trip);
+		DBG((dbg, LEVEL_4, "\t\tCNC: <<=== Recoloring of %+F %s\n", node->irn, res ? "succeeded" : "failed"));
+		stat_ev_int("heur4_recolor_depth_max", max_depth);
+		stat_ev_int("heur4_recolor_trip", trip);
+
+
+		return res;
+	}
+
+#ifdef DEBUG_libfirm
+		if (firm_dbg_get_mask(dbg) & LEVEL_4) {
+			if (!is_loose(node))
+				DB((dbg, LEVEL_4, "\t\tCNC: %+F has already fixed color %d\n", node->irn, col));
+			else {
+				DB((dbg, LEVEL_4, "\t\tCNC: color %d not admissible for %+F (", tgt_col, node->irn));
+				dbg_admissible_colors(env, node);
+				DB((dbg, LEVEL_4, ")\n"));
+			}
+		}
+#endif
+
+	return 0;
+}
+
+/**
+ * Tries to color an affinity chunk (or at least a part of it).
+ * Inserts uncolored parts of the chunk as a new chunk into the priority queue.
+ */
+static void color_aff_chunk(co_mst_env_t *env, aff_chunk_t *c) {
+	aff_chunk_t *best_chunk   = NULL;
+	int         n_nodes       = ARR_LEN(c->n);
+	int         best_color    = -1;
+	int         n_int_chunks  = 0;
+	waitq       *tmp_chunks   = new_waitq();
+	waitq       *best_starts  = NULL;
+	col_cost_t  *order        = ALLOCANZ(col_cost_t, env->n_regs);
+	bitset_t    *visited;
+	int         idx, len, i, nidx, pos;
+	struct list_head changed;
+
+	DB((dbg, LEVEL_2, "fragmentizing chunk #%u", c->id));
+	DBG_AFF_CHUNK(env, LEVEL_2, c);
+	DB((dbg, LEVEL_2, "\n"));
+
+	stat_ev_ctx_push_fmt("heur4_color_chunk", "%u", c->id);
+
+	++env->chunk_visited;
+
+	/* compute color preference */
+	for (pos = 0, len = ARR_LEN(c->interfere); pos < len; ++pos) {
+		const ir_node *n = c->interfere[pos];
+		co_mst_irn_t *node = get_co_mst_irn(env, n);
+		aff_chunk_t *chunk = node->chunk;
+
+		if (is_loose(node) && chunk && chunk->visited < env->chunk_visited) {
+			assert(!chunk->deleted);
+			chunk->visited = env->chunk_visited;
+			++n_int_chunks;
+
+			aff_chunk_assure_weight(env, chunk);
+			for (i = 0; i < env->n_regs; ++i)
+				order[i].cost += chunk->color_affinity[i].cost;
+		}
+	}
+
+	for (i = 0; i < env->n_regs; ++i) {
+		real_t dislike = n_int_chunks > 0 ? REAL(1.0) - order[i].cost / n_int_chunks : REAL(0.0);
+		order[i].col  = i;
+		order[i].cost = (REAL(1.0) - dislike_influence) * c->color_affinity[i].cost + dislike_influence * dislike;
+	}
+
+	qsort(order, env->n_regs, sizeof(order[0]), cmp_col_cost_gt);
+
+	DBG_COL_COST(env, LEVEL_2, order);
+	DB((dbg, LEVEL_2, "\n"));
+
+	/* check which color is the "best" for the given chunk.
+	 * if we found a color which was ok for all nodes, we take it
+	 * and do not look further. (see did_all flag usage below.)
+	 * If we have many colors which fit all nodes it is hard to decide
+	 * which one to take anyway.
+	 * TODO Sebastian: Perhaps we should at all nodes and figure out
+	 * a suitable color using costs as done above (determine_color_costs).
+	 */
+	for (i = 0; i < env->k; ++i) {
+		int         col = order[i].col;
+		waitq       *good_starts = new_waitq();
+		aff_chunk_t *local_best;
+		int          n_succeeded;
+
+		/* skip ignore colors */
+		if (bitset_is_set(env->ignore_regs, col))
+			continue;
+
+		DB((dbg, LEVEL_2, "\ttrying color %d\n", col));
+
+		n_succeeded = 0;
+
+		/* try to bring all nodes of given chunk to the current color. */
+		for (idx = 0, len = ARR_LEN(c->n); idx < len; ++idx) {
+			const ir_node   *irn  = c->n[idx];
+			co_mst_irn_t    *node = get_co_mst_irn(env, irn);
+			int              good;
+
+			assert(! node->fixed && "Node must not have a fixed color.");
+			DB((dbg, LEVEL_4, "\t\tBringing %+F from color %d to color %d ...\n", irn, node->col, col));
+
+			/*
+				The order of the colored nodes is important, so we record the successfully
+				colored ones in the order they appeared.
+			*/
+			INIT_LIST_HEAD(&changed);
+			stat_ev_tim_push();
+			good = change_node_color(env, node, col, &changed);
+			stat_ev_tim_pop("heur4_recolor");
+			if (good) {
+				waitq_put(good_starts, node);
+				materialize_coloring(&changed);
+				node->fixed = 1;
+			}
+
+			else
+				reject_coloring(&changed);
+
+			n_succeeded += good;
+			DB((dbg, LEVEL_4, "\t\t... %+F attempt from %d to %d %s\n", irn, node->col, col, good ? "succeeded" : "failed"));
+		}
+
+		/* unfix all nodes */
+		for (idx = 0, len = ARR_LEN(c->n); idx < len; ++idx) {
+			co_mst_irn_t *node = get_co_mst_irn(env, c->n[idx]);
+			node->fixed = 0;
+		}
+
+		/* try next color when failed */
+		if (n_succeeded == 0)
+			continue;
+
+		/* fragment the chunk according to the coloring */
+		local_best = fragment_chunk(env, col, c, tmp_chunks);
+
+		/* search the best of the good list
+		   and make it the new best if it is better than the current */
+		if (local_best) {
+			aff_chunk_assure_weight(env, local_best);
+
+			DB((dbg, LEVEL_3, "\t\tlocal best chunk (id %u) for color %d: ", local_best->id, col));
+			DBG_AFF_CHUNK(env, LEVEL_3, local_best);
+
+			if (! best_chunk || best_chunk->weight < local_best->weight) {
+				best_chunk = local_best;
+				best_color = col;
+				if (best_starts)
+					del_waitq(best_starts);
+				best_starts = good_starts;
+				DB((dbg, LEVEL_3, "\n\t\t... setting global best chunk (id %u), color %d\n", best_chunk->id, best_color));
+			} else {
+				DB((dbg, LEVEL_3, "\n\t\t... omitting, global best is better\n"));
+				del_waitq(good_starts);
+			}
+		}
+		else {
+			del_waitq(good_starts);
+		}
+
+		/* if all nodes were recolored, bail out */
+		if (n_succeeded == n_nodes)
+			break;
+	}
+
+	stat_ev_int("heur4_colors_tried", i);
+
+	/* free all intermediate created chunks except best one */
+	while (! waitq_empty(tmp_chunks)) {
+		aff_chunk_t *tmp = waitq_get(tmp_chunks);
+		if (tmp != best_chunk)
+			delete_aff_chunk(env, tmp);
+	}
+	del_waitq(tmp_chunks);
+
+	/* return if coloring failed */
+	if (! best_chunk) {
+		if (best_starts)
+			del_waitq(best_starts);
+		return;
+	}
+
+	DB((dbg, LEVEL_2, "\tbest chunk #%u ", best_chunk->id));
+	DBG_AFF_CHUNK(env, LEVEL_2, best_chunk);
+	DB((dbg, LEVEL_2, "using color %d\n", best_color));
+
+	for (idx = 0, len = ARR_LEN(best_chunk->n); idx < len; ++idx) {
+		const ir_node *irn  = best_chunk->n[idx];
+		co_mst_irn_t  *node = get_co_mst_irn(env, irn);
+		int res;
+
+		/* bring the node to the color. */
+		DB((dbg, LEVEL_4, "\tManifesting color %d for %+F, chunk #%u\n", best_color, node->irn, best_chunk->id));
+		INIT_LIST_HEAD(&changed);
+		stat_ev_tim_push();
+		res = change_node_color(env, node, best_color, &changed);
+		stat_ev_tim_pop("heur4_recolor");
+		if (res) {
+			materialize_coloring(&changed);
+			node->fixed = 1;
+		}
+		assert(list_empty(&changed));
+	}
+
+	/* remove the nodes in best chunk from original chunk */
+	len = ARR_LEN(best_chunk->n);
+	for (idx = 0; idx < len; ++idx) {
+		const ir_node *irn = best_chunk->n[idx];
+		int pos = nodes_bsearch(c->n, irn);
+
+		if (pos > 0)
+			c->n[pos] = NULL;
+	}
+	len = ARR_LEN(c->n);
+	for (idx = nidx = 0; idx < len; ++idx) {
+		const ir_node *irn = c->n[idx];
+
+		if (irn != NULL) {
+			c->n[nidx++] = irn;
+		}
+	}
+	ARR_SHRINKLEN(c->n, nidx);
+
+
+	/* we have to get the nodes back into the original chunk because they are scattered over temporary chunks */
+	for (idx = 0, len = ARR_LEN(c->n); idx < len; ++idx) {
+		const ir_node *n  = c->n[idx];
+		co_mst_irn_t  *nn = get_co_mst_irn(env, n);
+		nn->chunk = c;
+	}
+
+	/* fragment the remaining chunk */
+	visited = bitset_irg_malloc(env->co->irg);
+	for (idx = 0, len = ARR_LEN(best_chunk->n); idx < len; ++idx)
+		bitset_set(visited, get_irn_idx(best_chunk->n[idx]));
+
+	for (idx = 0, len = ARR_LEN(c->n); idx < len; ++idx) {
+		const ir_node *irn = c->n[idx];
+		if (! bitset_is_set(visited, get_irn_idx(irn))) {
+			aff_chunk_t  *new_chunk = new_aff_chunk(env);
+			co_mst_irn_t *node      = get_co_mst_irn(env, irn);
+
+			expand_chunk_from(env, node, visited, new_chunk, c, decider_always_yes, 0);
+			aff_chunk_assure_weight(env, new_chunk);
+			pqueue_put(env->chunks, new_chunk, new_chunk->weight);
+		}
+	}
+
+	for (idx = 0, len = ARR_LEN(best_chunk->n); idx < len; ++idx) {
+		const ir_node *n  = best_chunk->n[idx];
+		co_mst_irn_t  *nn = get_co_mst_irn(env, n);
+		nn->chunk = NULL;
+	}
+
+	/* clear obsolete chunks and free some memory */
+	delete_aff_chunk(env, best_chunk);
+	bitset_free(visited);
+	if (best_starts)
+		del_waitq(best_starts);
+
+	stat_ev_ctx_pop("heur4_color_chunk");
+}
+
+/**
+ * Main driver for mst safe coalescing algorithm.
+ */
+static int co_solve_heuristic_mst(copy_opt_t *co) {
+	unsigned     n_regs       = co->cls->n_regs;
+	bitset_t     *ignore_regs = bitset_alloca(n_regs);
+	unsigned     i, j, k;
+	ir_node      *irn;
+	co_mst_env_t mst_env;
+
+	last_chunk_id = 0;
+
+	stat_ev_tim_push();
 
 	/* init phase */
-	phase_init(&mst_env.ph, "co_mst_safe", co->irg, PHASE_DEFAULT_GROWTH, co_mst_safe_irn_init, &mst_env);
+	phase_init(&mst_env.ph, "co_mst", co->irg, PHASE_DEFAULT_GROWTH, co_mst_irn_init, &mst_env);
 
-	k = be_put_ignore_regs(co->cenv->birg, co->cenv->cls, ignore_regs);
+	k = be_put_ignore_regs(co->cenv->birg, co->cls, ignore_regs);
 	k = n_regs - k;
 
-	mst_env.n_regs = n_regs;
-	mst_env.k      = k;
-	mst_env.chunks = new_pqueue();
-	mst_env.co     = co;
+	mst_env.n_regs        = n_regs;
+	mst_env.k             = k;
+	mst_env.chunks        = new_pqueue();
+	mst_env.co            = co;
+	mst_env.ignore_regs   = ignore_regs;
+	mst_env.ifg           = co->cenv->ifg;
+	mst_env.chunkset      = pset_new_ptr(512);
+	mst_env.chunk_visited = 0;
+	mst_env.single_cols   = phase_alloc(&mst_env.ph, sizeof(*mst_env.single_cols) * n_regs);
+
+	for (i = 0; i < n_regs; ++i) {
+		col_cost_t *vec = phase_alloc(&mst_env.ph, sizeof(*vec) * n_regs);
+
+		mst_env.single_cols[i] = vec;
+		for (j = 0; j < n_regs; ++j) {
+			vec[j].col  = j;
+			vec[j].cost = REAL(0.0);
+		}
+		vec[i].col  = 0;
+		vec[0].col  = i;
+		vec[0].cost = REAL(1.0);
+	}
+
+	DBG((dbg, LEVEL_1, "==== Coloring %+F, class %s ====\n", co->irg, co->cls->name));
 
 	/* build affinity chunks */
+	stat_ev_tim_push();
 	build_affinity_chunks(&mst_env);
+	stat_ev_tim_pop("heur4_initial_chunk");
 
 	/* color chunks as long as there are some */
 	while (! pqueue_empty(mst_env.chunks)) {
-		aff_chunk_t *chunk = pqueue_get(mst_env.chunks);
+		aff_chunk_t *chunk = pqueue_pop_front(mst_env.chunks);
+
 		color_aff_chunk(&mst_env, chunk);
+		DB((dbg, LEVEL_4, "<<<====== Coloring chunk (%u) done\n", chunk->id));
+		delete_aff_chunk(&mst_env, chunk);
+	}
+
+	/* apply coloring */
+	foreach_phase_irn(&mst_env.ph, irn) {
+		co_mst_irn_t          *mirn;
+		const arch_register_t *reg;
+
+		if (arch_irn_is_ignore(irn))
+			continue;
+
+		mirn = get_co_mst_irn(&mst_env, irn);
+		// assert(mirn->fixed && "Node should have fixed color");
+
+		/* skip nodes where color hasn't changed */
+		if (mirn->init_col == mirn->col)
+			continue;
+
+		reg = arch_register_for_index(co->cls, mirn->col);
+		arch_set_irn_register(irn, reg);
+		DB((dbg, LEVEL_1, "%+F set color from %d to %d\n", irn, mirn->init_col, mirn->col));
 	}
 
 	/* free allocated memory */
 	del_pqueue(mst_env.chunks);
 	phase_free(&mst_env.ph);
+	del_pset(mst_env.chunkset);
+
+	stat_ev_tim_pop("heur4_total");
+
+	return 0;
 }
 
-#endif
+static const lc_opt_table_entry_t options[] = {
+	LC_OPT_ENT_INT      ("limit", "limit recoloring",  &recolor_limit),
+	LC_OPT_ENT_DBL      ("di",    "dislike influence", &dislike_influence),
+	LC_OPT_LAST
+};
+
+
+void be_init_copyheur4(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 *co_grp = lc_opt_get_grp(chordal_grp, "co");
+	lc_opt_entry_t *heur4_grp = lc_opt_get_grp(co_grp, "heur4");
+
+	static co_algo_info copyheur = {
+		co_solve_heuristic_mst, 0
+	};
+
+	lc_opt_add_table(heur4_grp, options);
+	be_register_copyopt("heur4", &copyheur);
+
+	FIRM_DBG_REGISTER(dbg, "firm.be.co.heur4");
+}
+
+BE_REGISTER_MODULE_CONSTRUCTOR(be_init_copyheur4);