#include "df.h"
#include "vecprim.h"
#include "ggc.h"
+#include "cfgloop.h"
+#include "params.h"
#ifdef XCOFF_DEBUGGING_INFO
#include "xcoffout.h" /* Needed for external data
{
int log, max_skip, max_log;
basic_block bb;
+ int freq_max = 0;
+ int freq_threshold = 0;
if (label_align)
{
if (! optimize || optimize_size)
return 0;
+ if (dump_file)
+ {
+ dump_flow_info (dump_file, TDF_DETAILS);
+ flow_loops_dump (dump_file, NULL, 1);
+ loop_optimizer_init (AVOID_CFG_MODIFICATIONS);
+ }
+ FOR_EACH_BB (bb)
+ if (bb->frequency > freq_max)
+ freq_max = bb->frequency;
+ freq_threshold = freq_max / PARAM_VALUE (PARAM_ALIGN_THRESHOLD);
+
+ if (dump_file)
+ fprintf(dump_file, "freq_max: %i\n",freq_max);
FOR_EACH_BB (bb)
{
rtx label = BB_HEAD (bb);
if (!LABEL_P (label)
|| probably_never_executed_bb_p (bb))
- continue;
+ {
+ if (dump_file)
+ fprintf(dump_file, "BB %4i freq %4i loop %2i loop_depth %2i skipped.\n",
+ bb->index, bb->frequency, bb->loop_father->num, bb->loop_depth);
+ continue;
+ }
max_log = LABEL_ALIGN (label);
max_skip = LABEL_ALIGN_MAX_SKIP;
else
branch_frequency += EDGE_FREQUENCY (e);
}
+ if (dump_file)
+ {
+ fprintf(dump_file, "BB %4i freq %4i loop %2i loop_depth %2i fall %4i branch %4i",
+ bb->index, bb->frequency, bb->loop_father->num,
+ bb->loop_depth,
+ fallthru_frequency, branch_frequency);
+ if (!bb->loop_father->inner && bb->loop_father->num)
+ fprintf (dump_file, " inner_loop");
+ if (bb->loop_father->header == bb)
+ fprintf (dump_file, " loop_header");
+ fprintf (dump_file, "\n");
+ }
/* There are two purposes to align block with no fallthru incoming edge:
1) to avoid fetch stalls when branch destination is near cache boundary
when function is called. */
if (!has_fallthru
- && (branch_frequency > BB_FREQ_MAX / 10
+ && (branch_frequency > freq_threshold
|| (bb->frequency > bb->prev_bb->frequency * 10
&& (bb->prev_bb->frequency
<= ENTRY_BLOCK_PTR->frequency / 2))))
{
log = JUMP_ALIGN (label);
+ if (dump_file)
+ fprintf(dump_file, " jump alignment added.\n");
if (max_log < log)
{
max_log = log;
align it. It is most likely a first block of loop. */
if (has_fallthru
&& maybe_hot_bb_p (bb)
- && branch_frequency + fallthru_frequency > BB_FREQ_MAX / 10
- && branch_frequency > fallthru_frequency * 2)
+ && branch_frequency + fallthru_frequency > freq_threshold
+ && (branch_frequency
+ > fallthru_frequency * PARAM_VALUE (PARAM_ALIGN_LOOP_ITERATIONS)))
{
log = LOOP_ALIGN (label);
+ if (dump_file)
+ fprintf(dump_file, " internal loop alignment added.\n");
if (max_log < log)
{
max_log = log;
LABEL_TO_ALIGNMENT (label) = max_log;
LABEL_TO_MAX_SKIP (label) = max_skip;
}
+
+ if (dump_file)
+ loop_optimizer_finalize ();
return 0;
}
struct tree_opt_pass pass_compute_alignments =
{
- NULL, /* name */
+ "alignments", /* name */
NULL, /* gate */
compute_alignments, /* execute */
NULL, /* sub */
0, /* properties_provided */
0, /* properties_destroyed */
0, /* todo_flags_start */
- 0, /* todo_flags_finish */
+ TODO_dump_func | TODO_verify_rtl_sharing
+ | TODO_ggc_collect, /* todo_flags_finish */
0 /* letter */
};
"Select fraction of the maximal frequency of executions of basic block in function given basic block needs to have to be considered hot",
1000, 0, 0)
+DEFPARAM (PARAM_ALIGN_THRESHOLD,
+ "align-threshold",
+ "Select fraction of the maximal frequency of executions of basic block in function given basic block get alignment",
+ 100, 0, 0)
+
+DEFPARAM (PARAM_ALIGN_LOOP_ITERATIONS,
+ "align-loop-iterations",
+ "Loops iterating at least selected number of iterations will get loop alignement.",
+ 4, 0, 0)
+
/* For guessed profiles, the loops having unknown number of iterations
are predicted to iterate relatively few (10) times at average.
For functions containing one loop with large known number of iterations