[AArch64 costs 12/18] Improve costs for sign/zero extracts

git cherry-pick d515bbc

gcc/

	* config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
	SIGN/ZERO_EXTRACT.

Change-Id: Ica7dd6d1721d38e1f8049d7be25db024ac7ed7b5
git-svn-id: svn+ssh://gcc.gnu.org/svn/gcc/trunk@210504 138bc75d-0d04-0410-961f-82ee72b054a4

diff --git a/gcc/ChangeLog b/gcc/ChangeLog
index b318f21..de89b7e 100644 (file)
--- a/gcc/ChangeLog
+++ b/gcc/ChangeLog
@@ -1,3 +1,9 @@
+2014-03-16  James Greenhalgh  <james.greenhalgh@arm.com>
+           Philipp Tomsich  <philipp.tomsich@theobroma-systems.com>
+
+       * config/aarch64/aarch64.c (aarch64_rtx_arith_op_extract_p): New.
+       (aarch64_rtx_costs): Improve costs for SIGN/ZERO_EXTRACT.
+
 2014-05-16  James Greenhalgh  <james.greenhalgh@arm.com>
            Philipp Tomsich  <philipp.tomsich@theobroma-systems.com>
 

diff --git a/gcc/config/aarch64/aarch64.c b/gcc/config/aarch64/aarch64.c
index ac7de2d..f687194 100644 (file)
--- a/gcc/config/aarch64/aarch64.c
+++ b/gcc/config/aarch64/aarch64.c
@@ -4783,6 +4783,35 @@ aarch64_address_cost (rtx x,
   return cost;
 }
 
+/* Return true if the RTX X in mode MODE is a zero or sign extract
+   usable in an ADD or SUB (extended register) instruction.  */
+static bool
+aarch64_rtx_arith_op_extract_p (rtx x, enum machine_mode mode)
+{
+  /* Catch add with a sign extract.
+     This is add_<optab><mode>_multp2.  */
+  if (GET_CODE (x) == SIGN_EXTRACT
+      || GET_CODE (x) == ZERO_EXTRACT)
+    {
+      rtx op0 = XEXP (x, 0);
+      rtx op1 = XEXP (x, 1);
+      rtx op2 = XEXP (x, 2);
+
+      if (GET_CODE (op0) == MULT
+         && CONST_INT_P (op1)
+         && op2 == const0_rtx
+         && CONST_INT_P (XEXP (op0, 1))
+         && aarch64_is_extend_from_extract (mode,
+                                            XEXP (op0, 1),
+                                            op1))
+       {
+         return true;
+       }
+    }
+
+  return false;
+}
+
 /* Calculate the cost of calculating X, storing it in *COST.  Result
    is true if the total cost of the operation has now been calculated.  */
 static bool
@@ -5047,6 +5076,18 @@ cost_minus:
 
          }
 
+       /* Look for SUB (extended register).  */
+        if (aarch64_rtx_arith_op_extract_p (op1, mode))
+         {
+           if (speed)
+             *cost += extra_cost->alu.arith_shift;
+
+           *cost += rtx_cost (XEXP (XEXP (op1, 0), 0),
+                              (enum rtx_code) GET_CODE (op1),
+                              0, speed);
+           return true;
+         }
+
        rtx new_op1 = aarch64_strip_extend (op1);
 
        /* Cost this as an FMA-alike operation.  */
@@ -5103,6 +5144,18 @@ cost_minus:
            return true;
          }
 
+       /* Look for ADD (extended register).  */
+        if (aarch64_rtx_arith_op_extract_p (op0, mode))
+         {
+           if (speed)
+             *cost += extra_cost->alu.arith_shift;
+
+           *cost += rtx_cost (XEXP (XEXP (op0, 0), 0),
+                              (enum rtx_code) GET_CODE (op0),
+                              0, speed);
+           return true;
+         }
+
        /* Strip any extend, leave shifts behind as we will
           cost them through mult_cost.  */
        new_op0 = aarch64_strip_extend (op0);
@@ -5338,7 +5391,13 @@ cost_minus:
 
     case ZERO_EXTRACT:
     case SIGN_EXTRACT:
-      *cost += rtx_cost (XEXP (x, 0), ZERO_EXTRACT, 0, speed);
+      /* UBFX/SBFX.  */
+      if (speed)
+       *cost += extra_cost->alu.bfx;
+
+      /* We can trust that the immediates used will be correct (there
+        are no by-register forms), so we need only cost op0.  */
+      *cost += rtx_cost (XEXP (x, 0), (enum rtx_code) code, 0, speed);
       return true;
 
     case MULT:
@@ -9145,7 +9204,7 @@ aarch64_cannot_change_mode_class (enum machine_mode from,
 #define TARGET_RETURN_IN_MSB aarch64_return_in_msb
 
 #undef TARGET_RTX_COSTS
-#define TARGET_RTX_COSTS aarch64_rtx_costs
+#define TARGET_RTX_COSTS aarch64_rtx_costs_wrapper
 
 #undef TARGET_SCHED_ISSUE_RATE
 #define TARGET_SCHED_ISSUE_RATE aarch64_sched_issue_rate