re PR tree-optimization/71347 (Performance drop after r235513 on x86-64 in 32-bit mode.)

	PR tree-optimization/71347
	* tree-ssa-loop-ivopts.c (determine_group_iv_cost_address): Compute
	cost for all uses in group.

	PR tree-optimization/71347
	* gcc.dg/tree-ssa/pr71347.c: New test.

From-SVN: r237552

gcc/ChangeLog

 2016-06-17  Bin Cheng  <bin.cheng@arm.com>
 
+	PR tree-optimization/71347
+	* tree-ssa-loop-ivopts.c (determine_group_iv_cost_address): Compute
+	cost for all uses in group.
+
+2016-06-17 Bin Cheng  <bin.cheng@arm.com>
+
 	* tree-vect-loop.c (vect_create_epilog_for_reduction): Only
 	insert gimple seq if it's not empty.
 

gcc/testsuite/ChangeLog

 2016-06-17  Bin Cheng  <bin.cheng@arm.com>
 
+	PR tree-optimization/71347
+	* gcc.dg/tree-ssa/pr71347.c: New test.
+
+2016-06-17  Bin Cheng  <bin.cheng@arm.com>
+
 	PR tree-optimization/71354
 	* gcc.dg/vect/vect-23.c: Add VECT_COND requirement.
 

gcc/testsuite/gcc.dg/tree-ssa/pr71347.c

+/* { dg-do compile } */
+/* { dg-options "-O2 -fdump-tree-optimized" } */
+
+double in;
+extern void Write (double);
+void foo (void)
+{
+  static double X[9];
+  int i;
+        X[1] = in * in;
+        for (i = 2; i <= 8; i++)
+            X[i] = X[i - 1] * X[1];
+        Write (X[5]);
+}
+
+/* Load of X[i - i] can be omitted by reusing X[i] in previous iteration.  */
+/* { dg-final { scan-tree-dump-not ".* = MEM.*;" "optimized"} } */

gcc/tree-ssa-loop-ivopts.c

@@ -5115,7 +5115,7 @@ determine_group_iv_cost_address (struct ivopts_data *data,
 {
   unsigned i;
   bitmap depends_on;
-  bool can_autoinc, first = true;
+  bool can_autoinc;
   iv_inv_expr_ent *inv_expr = NULL;
   struct iv_use *use = group->vuses[0];
   comp_cost sum_cost = no_cost, cost;
@@ -5142,30 +5142,11 @@ determine_group_iv_cost_address (struct ivopts_data *data,
     {
       struct iv_use *next = group->vuses[i];
 
-      /* Compute cost for the first use with different offset to the main
-	 use and add it afterwards.  Costs for these uses could be quite
-	 different.  Given below uses in a group:
-	   use 0  : {base + A + offset_0, step}
-	   use 0.1: {base + A + offset_0, step}
-	   use 0.2: {base + A + offset_1, step}
-	   use 0.3: {base + A + offset_2, step}
-	 when we need to compute costs with candidate:
-	   cand 1 : {base + B + offset_0, step}
-
-	 The first use with different offset is use 0.2, its cost is larger
-	 than cost of use 0/0.1 because we need to compute:
-	   A - B + offset_1 - offset_0
-	   rather than:
-	   A - B.  */
-      if (first && next->addr_offset != use->addr_offset)
-	{
-	  first = false;
+      /* TODO: We could skip computing cost for sub iv_use when it has the
+	 same cost as the first iv_use, but the cost really depends on the
+	 offset and where the iv_use is.  */
 	cost = get_computation_cost (data, next, cand, true,
 				     NULL, &can_autoinc, NULL);
-	  /* Remove setup cost.  */
-	  if (!cost.infinite_cost_p ())
-	    cost -= cost.scratch;
-	}
       sum_cost += cost;
     }
   set_group_iv_cost (data, group, cand, sum_cost, depends_on,