* param.h (MIN_VECT_LOOP_BOUND): New.

        * params.def (MIN_VECT_LOOP_BOUND): New.
        * tree-vectorizer.c (slpeel_tree_peel_loop_to_edge): Takes another
        argument - minimum threshold for number of iterations.
        * tree-vectorizer.h (slpeel_tree_peel_loop_to_edge): Add another
        argument to declaration.
        * tree-vect-analyze.c (vect_analyze_operations): Check value of
        MIN_VECT_LOOP_BOUND.
        * tree-vect-transform.c (vect_do_peeling_for_loop_bound): Call
        slpeel_tree_peel_loop_to_edge with additional argument.
        (vect_do_peeling_for_alignment): Likewise.
        * doc/invoke.texi (min-vect-loop-bound): Document new param option.



git-svn-id: svn+ssh://gcc.gnu.org/svn/gcc/trunk@120770 138bc75d-0d04-0410-961f-82ee72b054a4

8 files changed, 45 insertions, 6 deletions

diff --git a/gcc/ChangeLog b/gcc/ChangeLog
index 7e785b63d20..646e2928035 100644
--- a/gcc/ChangeLog
+++ b/gcc/ChangeLog
@@ -1,3 +1,18 @@
+2007-01-14  Dorit Nuzman  <dorit@il.ibm.com>
+
+	* param.h (MIN_VECT_LOOP_BOUND): New.
+	* params.def (MIN_VECT_LOOP_BOUND): New.
+	* tree-vectorizer.c (slpeel_tree_peel_loop_to_edge): Takes another
+	argument - minimum threshold for number of iterations. 
+	* tree-vectorizer.h (slpeel_tree_peel_loop_to_edge): Add another
+	argument to declaration.
+	* tree-vect-analyze.c (vect_analyze_operations): Check value of 
+	MIN_VECT_LOOP_BOUND.
+	* tree-vect-transform.c (vect_do_peeling_for_loop_bound): Call
+	slpeel_tree_peel_loop_to_edge with additional argument.
+	(vect_do_peeling_for_alignment): Likewise.
+	* doc/invoke.texi (min-vect-loop-bound): Document new param option.
+
 2007-01-14  Uros Bizjak  <ubizjak@gmail.com>
 
 	PR target/30413
diff --git a/gcc/doc/invoke.texi b/gcc/doc/invoke.texi
index f44e28d1486..b896aa5a6b2 100644
--- a/gcc/doc/invoke.texi
+++ b/gcc/doc/invoke.texi
@@ -6130,6 +6130,12 @@ inlining for code having large abstraction penalty (many functions that just
 pass the arguments to other functions) and decrease inlining for code with low
 abstraction penalty.  The default value is 16.
 
+@item min-vect-loop-bound
+The minimum number of iterations under which a loop will not get vectorized 
+when @option{-ftree-vectorize} is used.  The number of iterations after 
+vectorization needs to be greater than the value specified by this option
+to allow vectorization.  The default value is 0.
+
 @item max-unrolled-insns
 The maximum number of instructions that a loop should have if that loop
 is unrolled, and if the loop is unrolled, it determines how many times
diff --git a/gcc/params.def b/gcc/params.def
index 41a4e417ccf..6528361e6e9 100644
--- a/gcc/params.def
+++ b/gcc/params.def
@@ -146,6 +146,12 @@ DEFPARAM (PARAM_MAX_VARIABLE_EXPANSIONS,
 	  "If -fvariable-expansion-in-unroller is used, the maximum number of times that an individual variable will be expanded during loop unrolling",
           1, 0, 0)
      
+/* Limit loop autovectorization to loops with large enough iteration count.  */
+DEFPARAM (PARAM_MIN_VECT_LOOP_BOUND,
+	  "min-vect-loop-bound",
+	  "If -ftree-vectorize is used, the minimal loop bound of a loop to be considered for vectorization",
+	  0, 0, 0)
+
 /* The maximum number of instructions to consider when looking for an
    instruction to fill a delay slot.  If more than this arbitrary
    number of instructions is searched, the time savings from filling
diff --git a/gcc/params.h b/gcc/params.h
index 79a656c31e9..296db6141f5 100644
--- a/gcc/params.h
+++ b/gcc/params.h
@@ -118,6 +118,8 @@ typedef enum compiler_param
   PARAM_VALUE (PARAM_MAX_INLINE_INSNS_AUTO)
 #define MAX_VARIABLE_EXPANSIONS \
   PARAM_VALUE (PARAM_MAX_VARIABLE_EXPANSIONS)
+#define MIN_VECT_LOOP_BOUND \
+  PARAM_VALUE (PARAM_MIN_VECT_LOOP_BOUND)
 #define MAX_DELAY_SLOT_INSN_SEARCH \
   PARAM_VALUE (PARAM_MAX_DELAY_SLOT_INSN_SEARCH)
 #define MAX_DELAY_SLOT_LIVE_SEARCH \
diff --git a/gcc/tree-vect-analyze.c b/gcc/tree-vect-analyze.c
index c8b2bf858d6..83293d6051e 100644
--- a/gcc/tree-vect-analyze.c
+++ b/gcc/tree-vect-analyze.c
@@ -368,7 +368,10 @@ vect_analyze_operations (loop_vec_info loop_vinfo)
         vectorization_factor, LOOP_VINFO_INT_NITERS (loop_vinfo));
 
   if (LOOP_VINFO_NITERS_KNOWN_P (loop_vinfo)
-      && LOOP_VINFO_INT_NITERS (loop_vinfo) < vectorization_factor)
+      && ((LOOP_VINFO_INT_NITERS (loop_vinfo) < vectorization_factor)
+	  || (LOOP_VINFO_INT_NITERS (loop_vinfo) <=
+		((unsigned) (PARAM_VALUE (PARAM_MIN_VECT_LOOP_BOUND)) 
+					   * vectorization_factor))))
     {
       if (vect_print_dump_info (REPORT_UNVECTORIZED_LOOPS))
 	fprintf (vect_dump, "not vectorized: iteration count too small.");
diff --git a/gcc/tree-vect-transform.c b/gcc/tree-vect-transform.c
index ec3f511e74e..70fa217453a 100644
--- a/gcc/tree-vect-transform.c
+++ b/gcc/tree-vect-transform.c
@@ -35,6 +35,7 @@ Software Foundation, 51 Franklin Street, Fifth Floor, Boston, MA
 #include "cfgloop.h"
 #include "expr.h"
 #include "optabs.h"
+#include "params.h"
 #include "recog.h"
 #include "tree-data-ref.h"
 #include "tree-chrec.h"
@@ -4276,6 +4277,7 @@ vect_do_peeling_for_loop_bound (loop_vec_info loop_vinfo, tree *ratio)
   edge update_e;
   basic_block preheader;
   int loop_num;
+  unsigned int th;
 
   if (vect_print_dump_info (REPORT_DETAILS))
     fprintf (vect_dump, "=== vect_do_peeling_for_loop_bound ===");
@@ -4291,8 +4293,11 @@ vect_do_peeling_for_loop_bound (loop_vec_info loop_vinfo, tree *ratio)
 				   &ratio_mult_vf_name, ratio);
 
   loop_num  = loop->num; 
+  /* Threshold for vectorized loop.  */
+  th = (PARAM_VALUE (PARAM_MIN_VECT_LOOP_BOUND)) * 
+			LOOP_VINFO_VECT_FACTOR (loop_vinfo);
   new_loop = slpeel_tree_peel_loop_to_edge (loop, single_exit (loop),
-					    ratio_mult_vf_name, ni_name, false);
+					    ratio_mult_vf_name, ni_name, false, th);
   gcc_assert (new_loop);
   gcc_assert (loop_num == loop->num);
 #ifdef ENABLE_CHECKING
@@ -4517,7 +4522,7 @@ vect_do_peeling_for_alignment (loop_vec_info loop_vinfo)
   /* Peel the prolog loop and iterate it niters_of_prolog_loop.  */
   new_loop = 
 	slpeel_tree_peel_loop_to_edge (loop, loop_preheader_edge (loop), 
-				       niters_of_prolog_loop, ni_name, true); 
+				       niters_of_prolog_loop, ni_name, true, 0); 
   gcc_assert (new_loop);
 #ifdef ENABLE_CHECKING
   slpeel_verify_cfg_after_peeling (new_loop, loop);
diff --git a/gcc/tree-vectorizer.c b/gcc/tree-vectorizer.c
index c3b64e2ed6e..7fb98577951 100644
--- a/gcc/tree-vectorizer.c
+++ b/gcc/tree-vectorizer.c
@@ -1064,7 +1064,8 @@ slpeel_verify_cfg_after_peeling (struct loop *first_loop,
 struct loop*
 slpeel_tree_peel_loop_to_edge (struct loop *loop, 
 			       edge e, tree first_niters, 
-			       tree niters, bool update_first_loop_count)
+			       tree niters, bool update_first_loop_count,
+			       unsigned int th)
 {
   struct loop *new_loop = NULL, *first_loop, *second_loop;
   edge skip_e;
@@ -1157,7 +1158,8 @@ slpeel_tree_peel_loop_to_edge (struct loop *loop,
 
   pre_condition =
     fold_build2 (LE_EXPR, boolean_type_node, first_niters, 
-                 build_int_cst (TREE_TYPE (first_niters), 0));
+	build_int_cst (TREE_TYPE (first_niters), th));
+
   skip_e = slpeel_add_loop_guard (bb_before_first_loop, pre_condition,
                                   bb_before_second_loop, bb_before_first_loop);
   slpeel_update_phi_nodes_for_guard1 (skip_e, first_loop,
diff --git a/gcc/tree-vectorizer.h b/gcc/tree-vectorizer.h
index 69509403963..92b556712f3 100644
--- a/gcc/tree-vectorizer.h
+++ b/gcc/tree-vectorizer.h
@@ -342,7 +342,7 @@ extern bitmap vect_memsyms_to_rename;
    divide by the vectorization factor, and to peel the first few iterations
    to force the alignment of data references in the loop.  */
 extern struct loop *slpeel_tree_peel_loop_to_edge 
-  (struct loop *, edge, tree, tree, bool);
+  (struct loop *, edge, tree, tree, bool, unsigned int);
 extern void slpeel_make_loop_iterate_ntimes (struct loop *, tree);
 extern bool slpeel_can_duplicate_loop_p (struct loop *, edge);
 #ifdef ENABLE_CHECKING