if (gimple_get_lhs (stmt) == NULL_TREE)
{
+ if (is_gimple_call (stmt))
+ {
+ /* Ignore calls with no lhs. These must be calls to
+ #pragma omp simd functions, and what vectorization factor
+ it really needs can't be determined until
+ vectorizable_simd_clone_call. */
+ if (!analyze_pattern_stmt && gsi_end_p (pattern_def_si))
+ {
+ pattern_def_seq = NULL;
+ gsi_next (&si);
+ }
+ continue;
+ }
if (dump_enabled_p ())
{
dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location,
/* Function vect_get_loop_niters.
Determine how many iterations the loop is executed and place it
- in NUMBER_OF_ITERATIONS.
+ in NUMBER_OF_ITERATIONS. Place the number of latch iterations
+ in NUMBER_OF_ITERATIONSM1.
Return the loop exit condition. */
static gimple
-vect_get_loop_niters (struct loop *loop, tree *number_of_iterations)
+vect_get_loop_niters (struct loop *loop, tree *number_of_iterations,
+ tree *number_of_iterationsm1)
{
tree niters;
"=== get_loop_niters ===\n");
niters = number_of_latch_executions (loop);
+ *number_of_iterationsm1 = niters;
+
/* We want the number of loop header executions which is the number
of latch executions plus one.
??? For UINT_MAX latch executions this number overflows to zero
for loops like do { n++; } while (n != 0); */
if (niters && !chrec_contains_undetermined (niters))
- niters = fold_build2 (PLUS_EXPR, TREE_TYPE (niters), niters,
+ niters = fold_build2 (PLUS_EXPR, TREE_TYPE (niters), unshare_expr (niters),
build_int_cst (TREE_TYPE (niters), 1));
*number_of_iterations = niters;
gcc_assert (nbbs == loop->num_nodes);
LOOP_VINFO_BBS (res) = bbs;
+ LOOP_VINFO_NITERSM1 (res) = NULL;
LOOP_VINFO_NITERS (res) = NULL;
LOOP_VINFO_NITERS_UNCHANGED (res) = NULL;
LOOP_VINFO_COST_MODEL_MIN_ITERS (res) = 0;
{
loop_vec_info loop_vinfo;
gimple loop_cond;
- tree number_of_iterations = NULL;
+ tree number_of_iterations = NULL, number_of_iterationsm1 = NULL;
loop_vec_info inner_loop_vinfo = NULL;
if (dump_enabled_p ())
}
}
- loop_cond = vect_get_loop_niters (loop, &number_of_iterations);
+ loop_cond = vect_get_loop_niters (loop, &number_of_iterations,
+ &number_of_iterationsm1);
if (!loop_cond)
{
if (dump_enabled_p ())
}
loop_vinfo = new_loop_vec_info (loop);
+ LOOP_VINFO_NITERSM1 (loop_vinfo) = number_of_iterationsm1;
LOOP_VINFO_NITERS (loop_vinfo) = number_of_iterations;
LOOP_VINFO_NITERS_UNCHANGED (loop_vinfo) = number_of_iterations;
void *target_cost_data = LOOP_VINFO_TARGET_COST_DATA (loop_vinfo);
/* Cost model disabled. */
- if (unlimited_cost_model ())
+ if (unlimited_cost_model (LOOP_VINFO_LOOP (loop_vinfo)))
{
dump_printf_loc (MSG_NOTE, vect_location, "cost model disabled.\n");
*ret_min_profitable_niters = 0;
/* vector version will never be profitable. */
else
{
+ if (LOOP_VINFO_LOOP (loop_vinfo)->force_vect)
+ warning_at (vect_location, OPT_Wopenmp_simd, "vectorization "
+ "did not happen for a simd loop");
+
if (dump_enabled_p ())
dump_printf_loc (MSG_MISSED_OPTIMIZATION, vect_location,
"cost model: the vector iteration cost = %d "
struct loop *iv_loop;
basic_block new_bb;
tree new_vec, vec_init, vec_step, t;
- tree access_fn;
tree new_var;
tree new_name;
gimple init_stmt, induction_phi, new_stmt;
tree init_expr, step_expr;
int vf = LOOP_VINFO_VECT_FACTOR (loop_vinfo);
int i;
- bool ok;
int ncopies;
tree expr;
stmt_vec_info phi_info = vinfo_for_stmt (iv_phi);
latch_e = loop_latch_edge (iv_loop);
loop_arg = PHI_ARG_DEF_FROM_EDGE (iv_phi, latch_e);
- access_fn = analyze_scalar_evolution (iv_loop, PHI_RESULT (iv_phi));
- gcc_assert (access_fn);
- STRIP_NOPS (access_fn);
- ok = vect_is_simple_iv_evolution (iv_loop->num, access_fn,
- &init_expr, &step_expr);
- gcc_assert (ok);
+ step_expr = STMT_VINFO_LOOP_PHI_EVOLUTION_PART (phi_info);
+ gcc_assert (step_expr != NULL_TREE);
+
pe = loop_preheader_edge (iv_loop);
+ init_expr = PHI_ARG_DEF_FROM_EDGE (iv_phi,
+ loop_preheader_edge (iv_loop));
vectype = get_vectype_for_scalar_type (TREE_TYPE (init_expr));
resvectype = get_vectype_for_scalar_type (TREE_TYPE (PHI_RESULT (iv_phi)));
gcc_assert (phi_info);
gcc_assert (ncopies >= 1);
+ /* Convert the step to the desired type. */
+ step_expr = force_gimple_operand (fold_convert (TREE_TYPE (vectype),
+ step_expr),
+ &stmts, true, NULL_TREE);
+ if (stmts)
+ {
+ new_bb = gsi_insert_seq_on_edge_immediate (pe, stmts);
+ gcc_assert (!new_bb);
+ }
+
/* Find the first insertion point in the BB. */
si = gsi_after_labels (bb);
/* iv_loop is nested in the loop to be vectorized. init_expr had already
been created during vectorization of previous stmts. We obtain it
from the STMT_VINFO_VEC_STMT of the defining stmt. */
- tree iv_def = PHI_ARG_DEF_FROM_EDGE (iv_phi,
- loop_preheader_edge (iv_loop));
- vec_init = vect_get_vec_def_for_operand (iv_def, iv_phi, NULL);
+ vec_init = vect_get_vec_def_for_operand (init_expr, iv_phi, NULL);
/* If the initial value is not of proper type, convert it. */
if (!useless_type_conversion_p (vectype, TREE_TYPE (vec_init)))
{
tree var;
tree ratio_name;
tree ratio_mult_vf_name;
- tree ni = LOOP_VINFO_NITERS (loop_vinfo);
int vf = LOOP_VINFO_VECT_FACTOR (loop_vinfo);
edge pe = loop_preheader_edge (LOOP_VINFO_LOOP (loop_vinfo));
tree log_vf;
- log_vf = build_int_cst (TREE_TYPE (ni), exact_log2 (vf));
+ log_vf = build_int_cst (TREE_TYPE (ni_name), exact_log2 (vf));
/* If epilogue loop is required because of data accesses with gaps, we
subtract one iteration from the total number of iterations here for
build_one_cst (TREE_TYPE (ni_name)));
if (!is_gimple_val (ni_minus_gap_name))
{
- var = create_tmp_var (TREE_TYPE (ni), "ni_gap");
+ var = create_tmp_var (TREE_TYPE (ni_name), "ni_gap");
gimple stmts = NULL;
ni_minus_gap_name = force_gimple_operand (ni_minus_gap_name, &stmts,
true, var);
ni_minus_gap_name = ni_name;
/* Create: ratio = ni >> log2(vf) */
-
- ratio_name = fold_build2 (RSHIFT_EXPR, TREE_TYPE (ni_minus_gap_name),
- ni_minus_gap_name, log_vf);
+ /* ??? As we have ni == number of latch executions + 1, ni could
+ have overflown to zero. So avoid computing ratio based on ni
+ but compute it using the fact that we know ratio will be at least
+ one, thus via (ni - vf) >> log2(vf) + 1. */
+ ratio_name
+ = fold_build2 (PLUS_EXPR, TREE_TYPE (ni_name),
+ fold_build2 (RSHIFT_EXPR, TREE_TYPE (ni_name),
+ fold_build2 (MINUS_EXPR, TREE_TYPE (ni_name),
+ ni_minus_gap_name,
+ build_int_cst
+ (TREE_TYPE (ni_name), vf)),
+ log_vf),
+ build_int_cst (TREE_TYPE (ni_name), 1));
if (!is_gimple_val (ratio_name))
{
- var = create_tmp_var (TREE_TYPE (ni), "bnd");
+ var = create_tmp_var (TREE_TYPE (ni_name), "bnd");
gimple stmts = NULL;
ratio_name = force_gimple_operand (ratio_name, &stmts, true, var);
gsi_insert_seq_on_edge_immediate (pe, stmts);
ratio_name, log_vf);
if (!is_gimple_val (ratio_mult_vf_name))
{
- var = create_tmp_var (TREE_TYPE (ni), "ratio_mult_vf");
+ var = create_tmp_var (TREE_TYPE (ni_name), "ratio_mult_vf");
gimple stmts = NULL;
ratio_mult_vf_name = force_gimple_operand (ratio_mult_vf_name, &stmts,
true, var);
int vectorization_factor = LOOP_VINFO_VECT_FACTOR (loop_vinfo);
bool grouped_store;
bool slp_scheduled = false;
- unsigned int nunits;
gimple stmt, pattern_stmt;
gimple_seq pattern_def_seq = NULL;
gimple_stmt_iterator pattern_def_si = gsi_none ();
transform_pattern_stmt = false;
}
- gcc_assert (STMT_VINFO_VECTYPE (stmt_info));
- nunits = (unsigned int) TYPE_VECTOR_SUBPARTS (
- STMT_VINFO_VECTYPE (stmt_info));
- if (!STMT_SLP_TYPE (stmt_info)
- && nunits != (unsigned int) vectorization_factor
- && dump_enabled_p ())
- /* For SLP VF is set according to unrolling factor, and not to
- vector size, hence for SLP this print is not valid. */
- dump_printf_loc (MSG_NOTE, vect_location,
- "multiple-types.\n");
+ if (STMT_VINFO_VECTYPE (stmt_info))
+ {
+ unsigned int nunits
+ = (unsigned int)
+ TYPE_VECTOR_SUBPARTS (STMT_VINFO_VECTYPE (stmt_info));
+ if (!STMT_SLP_TYPE (stmt_info)
+ && nunits != (unsigned int) vectorization_factor
+ && dump_enabled_p ())
+ /* For SLP VF is set according to unrolling factor, and not
+ to vector size, hence for SLP this print is not valid. */
+ dump_printf_loc (MSG_NOTE, vect_location, "multiple-types.\n");
+ }
/* SLP. Schedule all the SLP instances when the first SLP stmt is
reached. */