Re: [PATCH v1] RISC-V: Support RVV VFMSAC rounding mode intrinsic API

["juzhe.zhong@rivai.ai" <juzhe.zhong@rivai.ai>]

[-- Attachment #1: Type: text/plain, Size: 6288 bytes --]

ok



juzhe.zhong@rivai.ai
 
From: pan2.li
Date: 2023-08-04 10:58
To: gcc-patches
CC: juzhe.zhong; kito.cheng; pan2.li; yanzhang.wang
Subject: [PATCH v1] RISC-V: Support RVV VFMSAC rounding mode intrinsic API
From: Pan Li <pan2.li@intel.com>
 
This patch would like to support the rounding mode API for the
VFMSAC for the below samples.
 
* __riscv_vfmsac_vv_f32m1_rm
* __riscv_vfmsac_vv_f32m1_rm_m
* __riscv_vfmsac_vf_f32m1_rm
* __riscv_vfmsac_vf_f32m1_rm_m
 
Signed-off-by: Pan Li <pan2.li@intel.com>
 
gcc/ChangeLog:
 
* config/riscv/riscv-vector-builtins-bases.cc
(class vfmsac_frm): New class for vfmsac frm.
(vfmsac_frm_obj): New declaration.
(BASE): Ditto.
* config/riscv/riscv-vector-builtins-bases.h: Ditto.
* config/riscv/riscv-vector-builtins-functions.def
(vfmsac_frm): New function definition.
 
gcc/testsuite/ChangeLog:
 
* gcc.target/riscv/rvv/base/float-point-single-multiply-sub.c: New test.
---
.../riscv/riscv-vector-builtins-bases.cc      | 24 ++++++++++
.../riscv/riscv-vector-builtins-bases.h       |  1 +
.../riscv/riscv-vector-builtins-functions.def |  2 +
.../base/float-point-single-multiply-sub.c    | 47 +++++++++++++++++++
4 files changed, 74 insertions(+)
create mode 100644 gcc/testsuite/gcc.target/riscv/rvv/base/float-point-single-multiply-sub.c
 
diff --git a/gcc/config/riscv/riscv-vector-builtins-bases.cc b/gcc/config/riscv/riscv-vector-builtins-bases.cc
index 1d4a5a18bf9..e73051bbd89 100644
--- a/gcc/config/riscv/riscv-vector-builtins-bases.cc
+++ b/gcc/config/riscv/riscv-vector-builtins-bases.cc
@@ -401,6 +401,28 @@ public:
   }
};
+/* Implements below instructions for frm
+   - vfmsac
+*/
+class vfmsac_frm : public function_base
+{
+public:
+  bool has_rounding_mode_operand_p () const override { return true; }
+
+  bool has_merge_operand_p () const override { return false; }
+
+  rtx expand (function_expander &e) const override
+  {
+    if (e.op_info->op == OP_TYPE_vf)
+      return e.use_ternop_insn (
+ true, code_for_pred_mul_scalar (MINUS, e.vector_mode ()));
+    if (e.op_info->op == OP_TYPE_vv)
+      return e.use_ternop_insn (
+ true, code_for_pred_mul (MINUS, e.vector_mode ()));
+    gcc_unreachable ();
+  }
+};
+
/* Implements vrsub.  */
class vrsub : public function_base
{
@@ -2168,6 +2190,7 @@ static CONSTEXPR const vfnmsub vfnmsub_obj;
static CONSTEXPR const vfnmacc vfnmacc_obj;
static CONSTEXPR const vfnmacc_frm vfnmacc_frm_obj;
static CONSTEXPR const vfmsac vfmsac_obj;
+static CONSTEXPR const vfmsac_frm vfmsac_frm_obj;
static CONSTEXPR const vfnmadd vfnmadd_obj;
static CONSTEXPR const vfmsub vfmsub_obj;
static CONSTEXPR const vfwmacc vfwmacc_obj;
@@ -2405,6 +2428,7 @@ BASE (vfnmsub)
BASE (vfnmacc)
BASE (vfnmacc_frm)
BASE (vfmsac)
+BASE (vfmsac_frm)
BASE (vfnmadd)
BASE (vfmsub)
BASE (vfwmacc)
diff --git a/gcc/config/riscv/riscv-vector-builtins-bases.h b/gcc/config/riscv/riscv-vector-builtins-bases.h
index 247074d0868..ca8a6dc1cc3 100644
--- a/gcc/config/riscv/riscv-vector-builtins-bases.h
+++ b/gcc/config/riscv/riscv-vector-builtins-bases.h
@@ -167,6 +167,7 @@ extern const function_base *const vfnmsub;
extern const function_base *const vfnmacc;
extern const function_base *const vfnmacc_frm;
extern const function_base *const vfmsac;
+extern const function_base *const vfmsac_frm;
extern const function_base *const vfnmadd;
extern const function_base *const vfmsub;
extern const function_base *const vfwmacc;
diff --git a/gcc/config/riscv/riscv-vector-builtins-functions.def b/gcc/config/riscv/riscv-vector-builtins-functions.def
index 223e8346cd8..8bae7e616ba 100644
--- a/gcc/config/riscv/riscv-vector-builtins-functions.def
+++ b/gcc/config/riscv/riscv-vector-builtins-functions.def
@@ -352,6 +352,8 @@ DEF_RVV_FUNCTION (vfmacc_frm, alu_frm, full_preds, f_vvvv_ops)
DEF_RVV_FUNCTION (vfmacc_frm, alu_frm, full_preds, f_vvfv_ops)
DEF_RVV_FUNCTION (vfnmacc_frm, alu_frm, full_preds, f_vvvv_ops)
DEF_RVV_FUNCTION (vfnmacc_frm, alu_frm, full_preds, f_vvfv_ops)
+DEF_RVV_FUNCTION (vfmsac_frm, alu_frm, full_preds, f_vvvv_ops)
+DEF_RVV_FUNCTION (vfmsac_frm, alu_frm, full_preds, f_vvfv_ops)
// 13.7. Vector Widening Floating-Point Fused Multiply-Add Instructions
DEF_RVV_FUNCTION (vfwmacc, alu, full_preds, f_wwvv_ops)
diff --git a/gcc/testsuite/gcc.target/riscv/rvv/base/float-point-single-multiply-sub.c b/gcc/testsuite/gcc.target/riscv/rvv/base/float-point-single-multiply-sub.c
new file mode 100644
index 00000000000..8fee552dd30
--- /dev/null
+++ b/gcc/testsuite/gcc.target/riscv/rvv/base/float-point-single-multiply-sub.c
@@ -0,0 +1,47 @@
+/* { dg-do compile } */
+/* { dg-options "-march=rv64gcv -mabi=lp64 -O3 -Wno-psabi" } */
+
+#include "riscv_vector.h"
+
+typedef float float32_t;
+
+vfloat32m1_t
+test_riscv_vfmsac_vv_f32m1_rm (vfloat32m1_t vd, vfloat32m1_t op1,
+        vfloat32m1_t op2, size_t vl) {
+  return __riscv_vfmsac_vv_f32m1_rm (vd, op1, op2, 0, vl);
+}
+
+vfloat32m1_t
+test_vfmsac_vv_f32m1_rm_m (vbool32_t mask, vfloat32m1_t vd, vfloat32m1_t op1,
+    vfloat32m1_t op2, size_t vl) {
+  return __riscv_vfmsac_vv_f32m1_rm_m (mask, vd, op1, op2, 1, vl);
+}
+
+vfloat32m1_t
+test_vfmsac_vf_f32m1_rm (vfloat32m1_t vd, float32_t op1, vfloat32m1_t op2,
+ size_t vl) {
+  return __riscv_vfmsac_vf_f32m1_rm (vd, op1, op2, 2, vl);
+}
+
+vfloat32m1_t
+test_vfmsac_vf_f32m1_rm_m (vfloat32m1_t vd, vbool32_t mask, float32_t op1,
+    vfloat32m1_t op2, size_t vl) {
+  return __riscv_vfmsac_vf_f32m1_rm_m (mask, vd, op1, op2, 3, vl);
+}
+
+vfloat32m1_t
+test_riscv_vfmsac_vv_f32m1 (vfloat32m1_t vd, vfloat32m1_t op1,
+     vfloat32m1_t op2, size_t vl) {
+  return __riscv_vfmsac_vv_f32m1 (vd, op1, op2, vl);
+}
+
+vfloat32m1_t
+test_vfmsac_vv_f32m1_m (vbool32_t mask, vfloat32m1_t vd, vfloat32m1_t op1,
+ vfloat32m1_t op2, size_t vl) {
+  return __riscv_vfmsac_vv_f32m1_m (mask, vd, op1, op2, vl);
+}
+
+/* { dg-final { scan-assembler-times {vfmsac\.v[vf]\s+v[0-9]+,\s*[fav]+[0-9]+,\s*v[0-9]+} 6 } } */
+/* { dg-final { scan-assembler-times {frrm\s+[axs][0-9]+} 4 } } */
+/* { dg-final { scan-assembler-times {fsrm\s+[axs][0-9]+} 4 } } */
+/* { dg-final { scan-assembler-times {fsrmi\s+[01234]} 4 } } */
-- 
2.34.1