[PATCH, rs6000] testcase coverage for vec_madd built-in functions

[PATCH, rs6000] testcase coverage for vec_madd built-in functions

[Will Schmidt]

Hi, 

[Patch, rs6000] testcase coverage for vec_madd built-in function
    
Add testcase coverage for vec_madd built-in function.

Tested across power platforms (p6 and newer).  
OK for trunk?

Thanks, 
-Will
    
[testsuite]

        2017-08-08  Will Schmidt  <will_schmidt@vnet.ibm.com>

	* fold-vec-madd-double.c: New.
	* fold-vec-madd-float.c: New.
	* fold-vec-madd-short.c: New.

diff --git a/gcc/testsuite/gcc.target/powerpc/fold-vec-madd-double.c b/gcc/testsuite/gcc.target/powerpc/fold-vec-madd-double.c
new file mode 100644
index 0000000..0fe7824
--- /dev/null
+++ b/gcc/testsuite/gcc.target/powerpc/fold-vec-madd-double.c
@@ -0,0 +1,17 @@
+/* Verify that overloaded built-ins for vec_madd with
+   double inputs produce the right results.  */
+
+/* { dg-do compile } */
+/* { dg-require-effective-target powerpc_vsx_ok } */
+/* { dg-options "-mvsx -O2" } */
+
+#include <altivec.h>
+
+vector double
+testd_l (vector double vd2, vector double vd3, vector double vd4)
+{
+  return vec_madd (vd2, vd3, vd4);
+}
+
+/* { dg-final { scan-assembler-times "xvmaddmdp|xvmaddadp" 1 } } */
+
diff --git a/gcc/testsuite/gcc.target/powerpc/fold-vec-madd-float.c b/gcc/testsuite/gcc.target/powerpc/fold-vec-madd-float.c
new file mode 100644
index 0000000..fcfe0c3
--- /dev/null
+++ b/gcc/testsuite/gcc.target/powerpc/fold-vec-madd-float.c
@@ -0,0 +1,17 @@
+/* Verify that overloaded built-ins for vec_madd with float
+   inputs produce the right results.  */
+
+/* { dg-do compile } */
+/* { dg-require-effective-target powerpc_vsx_ok } */
+/* { dg-options "-mvsx -O2" } */
+
+#include <altivec.h>
+
+vector float
+testf_l (vector float vf2, vector float vf3, vector float vf4)
+{
+  return vec_madd (vf2, vf3, vf4);
+}
+
+/* { dg-final { scan-assembler-times "xvmaddmsp|xvmaddasp" 1 } } */
+
diff --git a/gcc/testsuite/gcc.target/powerpc/fold-vec-madd-short.c b/gcc/testsuite/gcc.target/powerpc/fold-vec-madd-short.c
new file mode 100644
index 0000000..0e78f35
--- /dev/null
+++ b/gcc/testsuite/gcc.target/powerpc/fold-vec-madd-short.c
@@ -0,0 +1,38 @@
+/* Verify that overloaded built-ins for vec_madd with short
+   inputs produce the right results.  */
+
+/* { dg-do compile } */
+/* { dg-require-effective-target powerpc_altivec_ok } */
+/* { dg-options "-maltivec -O2" } */
+
+#include <altivec.h>
+
+vector signed short
+test_mad_sss (vector signed short vss2, vector signed short vss3,
+	      vector signed short vss4)
+{
+  return vec_madd (vss2, vss3, vss4);
+}
+
+vector signed short
+test_mad_suu (vector signed short vss2, vector unsigned short vus3,
+		vector unsigned short vus4)
+{
+  return vec_madd (vss2, vus3, vus4);
+}
+
+vector signed short
+test_mad_uss (vector unsigned short vus2, vector signed short vss3,
+	      vector signed short vss4)
+{
+  return vec_madd (vus2, vss3, vss4);
+}
+
+vector unsigned short
+test_mad_uuu (vector unsigned short vus2, vector unsigned short vus3,
+		vector unsigned short vus4)
+{
+  return vec_madd (vus2, vus3, vus4);
+}
+
+/* { dg-final { scan-assembler-times "vmladduhm" 4 } } */

Re: [PATCH, rs6000] testcase coverage for vec_madd built-in functions

[Segher Boessenkool]

Hi!

On Tue, Aug 08, 2017 at 04:20:14PM -0500, Will Schmidt wrote:
> 	* fold-vec-madd-double.c: New.
> 	* fold-vec-madd-float.c: New.
> 	* fold-vec-madd-short.c: New.

Same as before, it needs the path in here (the path is relative to the
directory the changelog is in).

The actual tests look fine, please commit.  Thanks!


Segher