#include <isl_schedule_private.h>
#include <isl_band_private.h>
#include <isl_list_private.h>
+#include <isl_options_private.h>
/*
* The scheduling algorithm implemented in this file was inspired by
* n is the number of nodes
* node is the list of nodes
* maxvar is the maximal number of variables over all nodes
+ * max_row is the allocated number of rows in the schedule
* n_row is the current (maximal) number of linearly independent
* rows in the node schedules
* n_total_row is the current number of rows in the node schedules
struct isl_sched_node *node;
int n;
int maxvar;
+ int max_row;
int n_row;
int *sorted;
isl_basic_set_free(graph->lp);
}
+/* For each "set" on which this function is called, increment
+ * graph->n by one and update graph->maxvar.
+ */
+static int init_n_maxvar(__isl_take isl_set *set, void *user)
+{
+ struct isl_sched_graph *graph = user;
+ int nvar = isl_set_dim(set, isl_dim_set);
+
+ graph->n++;
+ if (nvar > graph->maxvar)
+ graph->maxvar = nvar;
+
+ isl_set_free(set);
+
+ return 0;
+}
+
+/* Compute the number of rows that should be allocated for the schedule.
+ * The graph can be split at most "n - 1" times, there can be at most
+ * two rows for each dimension in the iteration domains (in particular,
+ * we usually have one row, but it may be split by split_parallel),
+ * and there can be one extra row for ordering the statements.
+ * Note that if we have actually split "n - 1" times, then no ordering
+ * is needed, so in principle we could use "graph->n + 2 * graph->maxvar - 1".
+ */
+static int compute_max_row(struct isl_sched_graph *graph,
+ __isl_keep isl_union_set *domain)
+{
+ graph->n = 0;
+ graph->maxvar = 0;
+ if (isl_union_set_foreach_set(domain, &init_n_maxvar, graph) < 0)
+ return -1;
+ graph->max_row = graph->n + 2 * graph->maxvar;
+
+ return 0;
+}
+
/* Add a new node to the graph representing the given set.
*/
static int extract_node(__isl_take isl_set *set, void *user)
graph->node[graph->n].nparam = nparam;
graph->node[graph->n].sched = sched;
graph->node[graph->n].sched_map = NULL;
- band = isl_alloc_array(ctx, int, graph->n_edge + nvar);
+ band = isl_alloc_array(ctx, int, graph->max_row);
graph->node[graph->n].band = band;
- band_id = isl_calloc_array(ctx, int, graph->n_edge + nvar);
+ band_id = isl_calloc_array(ctx, int, graph->max_row);
graph->node[graph->n].band_id = band_id;
- zero = isl_calloc_array(ctx, int, graph->n_edge + nvar);
+ zero = isl_calloc_array(ctx, int, graph->max_row);
graph->node[graph->n].zero = zero;
graph->n++;
int parametric;
int param_pos;
int n_eq, n_ineq;
+ int max_constant_term;
+
+ max_constant_term = ctx->opt->schedule_max_constant_term;
parametric = ctx->opt->schedule_parametric;
nparam = isl_space_dim(graph->node[0].dim, isl_dim_param);
dim = isl_space_set_alloc(ctx, 0, total);
isl_basic_set_free(graph->lp);
n_eq += 2 + parametric + force_zero;
+ if (max_constant_term != -1)
+ n_ineq += graph->n;
+
graph->lp = isl_basic_set_alloc_space(dim, 0, n_eq, n_ineq);
k = isl_basic_set_alloc_equality(graph->lp);
isl_int_set_si(graph->lp->eq[k][pos + j], 1);
}
+ if (max_constant_term != -1)
+ for (i = 0; i < graph->n; ++i) {
+ struct isl_sched_node *node = &graph->node[i];
+ k = isl_basic_set_alloc_inequality(graph->lp);
+ if (k < 0)
+ return -1;
+ isl_seq_clr(graph->lp->ineq[k], 1 + total);
+ isl_int_set_si(graph->lp->ineq[k][1 + node->start], -1);
+ isl_int_set_si(graph->lp->ineq[k][0], max_constant_term);
+ }
+
if (add_all_validity_constraints(graph) < 0)
return -1;
if (add_all_proximity_constraints(graph) < 0)
/* Copy non-empty edges that satisfy edge_pred from the src dependence graph
* to the dst dependence graph.
+ * If the source or destination node of the edge is not in the destination
+ * graph, then it must be a backward proximity edge and it should simply
+ * be ignored.
*/
static int copy_edges(isl_ctx *ctx, struct isl_sched_graph *dst,
struct isl_sched_graph *src,
for (i = 0; i < src->n_edge; ++i) {
struct isl_sched_edge *edge = &src->edge[i];
isl_map *map;
+ struct isl_sched_node *dst_src, *dst_dst;
if (!edge_pred(edge, data))
continue;
if (isl_map_plain_is_empty(edge->map))
continue;
+ dst_src = graph_find_node(ctx, dst, edge->src->dim);
+ dst_dst = graph_find_node(ctx, dst, edge->dst->dim);
+ if (!dst_src || !dst_dst) {
+ if (edge->validity)
+ isl_die(ctx, isl_error_internal,
+ "backward validity edge", return -1);
+ continue;
+ }
+
map = isl_map_copy(edge->map);
- dst->edge[dst->n_edge].src =
- graph_find_node(ctx, dst, edge->src->dim);
- dst->edge[dst->n_edge].dst =
- graph_find_node(ctx, dst, edge->dst->dim);
+ dst->edge[dst->n_edge].src = dst_src;
+ dst->edge[dst->n_edge].dst = dst_dst;
dst->edge[dst->n_edge].map = map;
dst->edge[dst->n_edge].validity = edge->validity;
dst->edge[dst->n_edge].proximity = edge->proximity;
* It would be possible to reuse them as the first rows in the next
* band, but recomputing them may result in better rows as we are looking
* at a smaller part of the dependence graph.
+ * compute_split_schedule is only called when no zero-distance schedule row
+ * could be found on the entire graph, so we wark the splitting row as
+ * non zero-distance.
*
* The band_id of the second group is set to n, where n is the number
* of nodes in the first group. This ensures that the band_ids over
node->sched = isl_mat_set_element_si(node->sched,
row, j, 0);
node->band[graph->n_total_row] = graph->n_band;
+ node->zero[graph->n_total_row] = 0;
}
e1 = e2 = 0;
for (i = 1; i < graph->n; ++i) {
struct isl_sched_node *node = &graph->node[i];
+ if (isl_mat_cols(node->sched) != cols)
+ return 0;
if (!isl_seq_eq(node0->sched->row[row] + 1,
node->sched->row[row] + 1, cols - 1))
return 0;
return compute_next_band(ctx, graph);
}
+/* Are there any validity edges in the graph?
+ */
+static int has_validity_edges(struct isl_sched_graph *graph)
+{
+ int i;
+
+ for (i = 0; i < graph->n_edge; ++i)
+ if (graph->edge[i].validity)
+ return 1;
+
+ return 0;
+}
+
+/* Should we apply a Feautrier step?
+ * That is, did the user request the Feautrier algorithm and are
+ * there any validity dependences (left)?
+ */
+static int need_feautrier_step(isl_ctx *ctx, struct isl_sched_graph *graph)
+{
+ if (ctx->opt->schedule_algorithm != ISL_SCHEDULE_ALGORITHM_FEAUTRIER)
+ return 0;
+
+ return has_validity_edges(graph);
+}
+
+/* Compute a schedule for a connected dependence graph using Feautrier's
+ * multi-dimensional scheduling algorithm.
+ * The original algorithm is described in [1].
+ * The main idea is to minimize the number of scheduling dimensions, by
+ * trying to satisfy as many dependences as possible per scheduling dimension.
+ *
+ * [1] P. Feautrier, Some Efficient Solutions to the Affine Scheduling
+ * Problem, Part II: Multi-Dimensional Time.
+ * In Intl. Journal of Parallel Programming, 1992.
+ */
+static int compute_schedule_wcc_feautrier(isl_ctx *ctx,
+ struct isl_sched_graph *graph)
+{
+ return carry_dependences(ctx, graph);
+}
+
/* Compute a schedule for a connected dependence graph.
* We try to find a sequence of as many schedule rows as possible that result
* in non-negative dependence distances (independent of the previous rows
* - try to carry as many dependences as possible and continue with the next
* band
*
+ * If Feautrier's algorithm is selected, we first recursively try to satisfy
+ * as many validity dependences as possible. When all validity dependences
+ * are satisfied we extend the schedule to a full-dimensional schedule.
+ *
* If we manage to complete the schedule, we finish off by topologically
* sorting the statements based on the remaining dependences.
*
if (compute_maxvar(graph) < 0)
return -1;
+ if (need_feautrier_step(ctx, graph))
+ return compute_schedule_wcc_feautrier(ctx, graph);
+
if (ctx->opt->schedule_outer_zero_distance)
force_zero = 1;
/* Compute a schedule for the given dependence graph.
* We first check if the graph is connected (through validity dependences)
- * and if so compute a schedule for each component separately.
+ * and, if not, compute a schedule for each component separately.
*/
static int compute_schedule(isl_ctx *ctx, struct isl_sched_graph *graph)
{
}
/* Compute a schedule for the given union of domains that respects
- * all the validity dependences and tries to minimize the dependence
- * distances over the proximity dependences.
+ * all the validity dependences.
+ * If the default isl scheduling algorithm is used, it tries to minimize
+ * the dependence distances over the proximity dependences.
+ * If Feautrier's scheduling algorithm is used, the proximity dependence
+ * distances are only minimized during the extension to a full-dimensional
+ * schedule.
*/
__isl_give isl_schedule *isl_union_set_compute_schedule(
__isl_take isl_union_set *domain,
if (graph_alloc(ctx, &graph, graph.n,
isl_union_map_n_map(validity) + isl_union_map_n_map(proximity)) < 0)
goto error;
+ if (compute_max_row(&graph, domain) < 0)
+ goto error;
graph.root = 1;
graph.n = 0;
if (isl_union_set_foreach_set(domain, &extract_node, &graph) < 0)