From: Alan Modra <amodra@gmail.com>
Date: Tue, 21 May 2019 13:36:04 +0000 (+0930)
Subject: PR90545, gcc.target/powerpc/fold-vec-splats-floatdouble.c fails
X-Git-Url: https://git.libre-soc.org/?a=commitdiff_plain;h=78e5da0aa817ac7e84a5f0e803a5b8b94fa122e0;p=gcc.git

PR90545, gcc.target/powerpc/fold-vec-splats-floatdouble.c fails

I figure a tweak to register_move_cost is better than sprinkling ?s
in instruction operand alternatives.

	PR target/90545
	* config/rs6000/rs6000.c (rs6000_register_move_cost): Increase
	power9 direct move cost.
	* testsuite/gcc.target/powerpc/fold-vec-splats-floatdouble.c:
	Correct comments and rename functions to suit parameters.

From-SVN: r271464
---

diff --git a/gcc/ChangeLog b/gcc/ChangeLog
index cd55c4dffad..eff11afb0d2 100644
--- a/gcc/ChangeLog
+++ b/gcc/ChangeLog
@@ -1,3 +1,11 @@
+2019-05-21  Alan Modra  <amodra@gmail.com>
+
+	PR target/90545
+	* config/rs6000/rs6000.c (rs6000_register_move_cost): Increase
+	power9 direct move cost.
+	* testsuite/gcc.target/powerpc/fold-vec-splats-floatdouble.c:
+	Correct comments and rename functions to suit parameters.
+
 2019-05-21  Richard Biener  <rguenther@suse.de>
 
 	PR middle-end/90510
diff --git a/gcc/config/rs6000/rs6000.c b/gcc/config/rs6000/rs6000.c
index 98429817539..693f2e1c91f 100644
--- a/gcc/config/rs6000/rs6000.c
+++ b/gcc/config/rs6000/rs6000.c
@@ -34657,8 +34657,14 @@ rs6000_register_move_cost (machine_mode mode,
 	{
 	  if (TARGET_DIRECT_MOVE)
 	    {
+	      /* Keep the cost for direct moves above that for within
+		 a register class even if the actual processor cost is
+		 comparable.  We do this because a direct move insn
+		 can't be a nop, whereas with ideal register
+		 allocation a move within the same class might turn
+		 out to be a nop.  */
 	      if (rs6000_tune == PROCESSOR_POWER9)
-		ret = 2 * hard_regno_nregs (FIRST_GPR_REGNO, mode);
+		ret = 3 * hard_regno_nregs (FIRST_GPR_REGNO, mode);
 	      else
 		ret = 4 * hard_regno_nregs (FIRST_GPR_REGNO, mode);
 	      /* SFmode requires a conversion when moving between gprs
diff --git a/gcc/testsuite/gcc.target/powerpc/fold-vec-splats-floatdouble.c b/gcc/testsuite/gcc.target/powerpc/fold-vec-splats-floatdouble.c
index c4544f1a452..3c7cc7c9a67 100644
--- a/gcc/testsuite/gcc.target/powerpc/fold-vec-splats-floatdouble.c
+++ b/gcc/testsuite/gcc.target/powerpc/fold-vec-splats-floatdouble.c
@@ -8,20 +8,20 @@
 #include <altivec.h>
 
 vector float
-test1d (float x)
+test1f (float x)
 {
   return vec_splats (x);
 }
 
 vector double
-test1f (double x)
+test1d (double x)
 {
   return vec_splats (x);
 }
 
-// float test generates the permute instruction.
+// double test generates the permute instruction.
 /* { dg-final { scan-assembler-times "xxpermdi" 1 } } */
 
-// double test generates a convert (double to single non-signalling) followed by a splat.
+// float test generates a convert (double to single non-signalling) followed by a splat.
 /* { dg-final { scan-assembler-times {\mxscvdpspn?\M} 1 } } */
 /* { dg-final { scan-assembler-times {\mvspltw\M|\mxxspltw\M} 1 } } */