RISC-V: Support RVV VFWREDOSUM.VS rounding mode intrinsic API

This patch would like to support the rounding mode API for the
VFWREDOSUM.VS as the below samples

* __riscv_vfwredosum_vs_f32m1_f64m1_rm
* __riscv_vfwredosum_vs_f32m1_f64m1_rm_m

Signed-off-by: Pan Li <pan2.li@intel.com>
gcc/ChangeLog:

	* config/riscv/riscv-vector-builtins-bases.cc
	(widen_freducop): Add frm_opt_type template arg.
	(vfwredosum_frm_obj): New declaration.
	(BASE): Ditto.
	* config/riscv/riscv-vector-builtins-bases.h: Ditto.
	* config/riscv/riscv-vector-builtins-functions.def
	(vfwredosum_frm): New intrinsic function def.

gcc/testsuite/ChangeLog:

	* gcc.target/riscv/rvv/base/float-point-wredosum.c: New test.

diff --git a/gcc/config/riscv/riscv-vector-builtins-bases.cc b/gcc/config/riscv/riscv-vector-builtins-bases.cc
index ef2991359dad6c6be2791b4974c566b3095a2af5..abf03bab0dac012213b4a704720396af3101c726 100644 (file)
--- a/gcc/config/riscv/riscv-vector-builtins-bases.cc
+++ b/gcc/config/riscv/riscv-vector-builtins-bases.cc
@@ -1866,10 +1866,15 @@ public:
 };
 
 /* Implements widening floating-point reduction instructions.  */
-template<int UNSPEC>
+template<int UNSPEC, enum frm_op_type FRM_OP = NO_FRM>
 class widen_freducop : public function_base
 {
 public:
+  bool has_rounding_mode_operand_p () const override
+  {
+    return FRM_OP == HAS_FRM;
+  }
+
   bool apply_mask_policy_p () const override { return false; }
 
   rtx expand (function_expander &e) const override
@@ -2544,6 +2549,7 @@ static CONSTEXPR const reducop<SMAX> vfredmax_obj;
 static CONSTEXPR const reducop<SMIN> vfredmin_obj;
 static CONSTEXPR const widen_freducop<UNSPEC_UNORDERED> vfwredusum_obj;
 static CONSTEXPR const widen_freducop<UNSPEC_ORDERED> vfwredosum_obj;
+static CONSTEXPR const widen_freducop<UNSPEC_ORDERED, HAS_FRM> vfwredosum_frm_obj;
 static CONSTEXPR const vmv vmv_x_obj;
 static CONSTEXPR const vmv_s vmv_s_obj;
 static CONSTEXPR const vmv vfmv_f_obj;
@@ -2802,6 +2808,7 @@ BASE (vfredosum_frm)
 BASE (vfredmax)
 BASE (vfredmin)
 BASE (vfwredosum)
+BASE (vfwredosum_frm)
 BASE (vfwredusum)
 BASE (vmv_x)
 BASE (vmv_s)

diff --git a/gcc/config/riscv/riscv-vector-builtins-bases.h b/gcc/config/riscv/riscv-vector-builtins-bases.h
index da8412b66df972b38fe345d1a1a9e4206f167a46..c1bb164a712c61ab744ddb1990f5b9e66a09a62e 100644 (file)
--- a/gcc/config/riscv/riscv-vector-builtins-bases.h
+++ b/gcc/config/riscv/riscv-vector-builtins-bases.h
@@ -245,6 +245,7 @@ extern const function_base *const vfredosum_frm;
 extern const function_base *const vfredmax;
 extern const function_base *const vfredmin;
 extern const function_base *const vfwredosum;
+extern const function_base *const vfwredosum_frm;
 extern const function_base *const vfwredusum;
 extern const function_base *const vmv_x;
 extern const function_base *const vmv_s;

diff --git a/gcc/config/riscv/riscv-vector-builtins-functions.def b/gcc/config/riscv/riscv-vector-builtins-functions.def
index 80e65bfb14bd54d692c3605eef0adb329e6767d9..da1157f5a5672650b78969559678b4be17f8727f 100644 (file)
--- a/gcc/config/riscv/riscv-vector-builtins-functions.def
+++ b/gcc/config/riscv/riscv-vector-builtins-functions.def
@@ -507,6 +507,8 @@ DEF_RVV_FUNCTION (vfredosum_frm, reduc_alu_frm, no_mu_preds, f_vs_ops)
 DEF_RVV_FUNCTION (vfwredosum, reduc_alu, no_mu_preds, wf_vs_ops)
 DEF_RVV_FUNCTION (vfwredusum, reduc_alu, no_mu_preds, wf_vs_ops)
 
+DEF_RVV_FUNCTION (vfwredosum_frm, reduc_alu_frm, no_mu_preds, wf_vs_ops)
+
 /* 15. Vector Mask Instructions.  */
 
 // 15.1. Vector Mask-Register Logical Instructions

diff --git a/gcc/testsuite/gcc.target/riscv/rvv/base/float-point-wredosum.c b/gcc/testsuite/gcc.target/riscv/rvv/base/float-point-wredosum.c
new file mode 100644 (file)
index 0000000..acf7956
--- /dev/null
+++ b/gcc/testsuite/gcc.target/riscv/rvv/base/float-point-wredosum.c
@@ -0,0 +1,33 @@
+/* { dg-do compile } */
+/* { dg-options "-march=rv64gcv -mabi=lp64 -O3 -Wno-psabi" } */
+
+#include "riscv_vector.h"
+
+vfloat64m1_t
+test_riscv_vfwredosum_vs_f32m1_f64m1_rm (vfloat32m1_t op1, vfloat64m1_t op2,
+                                        size_t vl) {
+  return __riscv_vfwredosum_vs_f32m1_f64m1_rm (op1, op2, 0, vl);
+}
+
+vfloat64m1_t
+test_vfwredosum_vs_f32m1_f64m1_rm_m (vbool32_t mask, vfloat32m1_t op1,
+                                    vfloat64m1_t op2, size_t vl) {
+  return __riscv_vfwredosum_vs_f32m1_f64m1_rm_m (mask, op1, op2, 1, vl);
+}
+
+vfloat64m1_t
+test_riscv_vfwredosum_vs_f32m1_f64m1 (vfloat32m1_t op1, vfloat64m1_t op2,
+                                     size_t vl) {
+  return __riscv_vfwredosum_vs_f32m1_f64m1 (op1, op2, vl);
+}
+
+vfloat64m1_t
+test_vfwredosum_vs_f32m1_f64m1_m (vbool32_t mask, vfloat32m1_t op1,
+                                 vfloat64m1_t op2, size_t vl) {
+  return __riscv_vfwredosum_vs_f32m1_f64m1_m (mask, op1, op2, vl);
+}
+
+/* { dg-final { scan-assembler-times {vfwredosum\.vs\s+v[0-9]+,\s*v[0-9]+} 4 } } */
+/* { dg-final { scan-assembler-times {frrm\s+[axs][0-9]+} 2 } } */
+/* { dg-final { scan-assembler-times {fsrm\s+[axs][0-9]+} 2 } } */
+/* { dg-final { scan-assembler-times {fsrmi\s+[01234]} 2 } } */