/* { dg-do compile { target { s390*-*-* } } } */ /* { dg-options "-march=z13 -mzarch -mzvector" } */ /* With IBM z14 a hardware instruction for floating point min and max has been added while for IBM z13 we emulated min/max for vector double with compare and select. This testcase makes sure that we fall back to the emulated variant when compiling for z13. */ #include vector double foo (vector double a, vector double b) { return vec_min (a, b); } vector double bar (vector double a, vector double b) { return vec_max (a, b); }