[llvm] [RISCV][VLOPT] Add fsqrt instruction to isSupportInstr (PR #127462)

[via llvm-commits]

llvmbot wrote:


<!--LLVM PR SUMMARY COMMENT-->

@llvm/pr-subscribers-backend-risc-v

Author: LiqinWeng (LiqinWeng)

<details>
<summary>Changes</summary>



---
Full diff: https://github.com/llvm/llvm-project/pull/127462.diff


3 Files Affected:

- (modified) llvm/lib/Target/RISCV/RISCVVLOptimizer.cpp (+4-1) 
- (modified) llvm/test/CodeGen/RISCV/rvv/fixed-vectors-fp.ll (+2-4) 
- (modified) llvm/test/CodeGen/RISCV/rvv/vl-opt.mir (+20) 


``````````diff
diff --git a/llvm/lib/Target/RISCV/RISCVVLOptimizer.cpp b/llvm/lib/Target/RISCV/RISCVVLOptimizer.cpp
index 1ba7f0b522a2b..0ed15afc7533d 100644
--- a/llvm/lib/Target/RISCV/RISCVVLOptimizer.cpp
+++ b/llvm/lib/Target/RISCV/RISCVVLOptimizer.cpp
@@ -290,7 +290,6 @@ getOperandLog2EEW(const MachineOperand &MO, const MachineRegisterInfo *MRI) {
       return MILog2SEW;
     return 6;
   }
-
   // Vector Integer Arithmetic Instructions
   // Vector Single-Width Integer Add and Subtract
   case RISCV::VADD_VI:
@@ -1141,6 +1140,10 @@ static bool isSupportedInstr(const MachineInstr &MI) {
   case RISCV::VFNCVT_F_F_W:
   case RISCV::VFNCVT_ROD_F_F_W:
   case RISCV::VFNCVTBF16_F_F_W:
+  // Vector Floating-Point Square-Root Instruction
+  case RISCV::VFSQRT_V:
+  // Vector Floating-Point Reciprocal Square-Root Estimate Instruction
+  case RISCV::VFRSQRT7_V:
     return true;
   }
 
diff --git a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-fp.ll b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-fp.ll
index 585a331e55094..bef29dfecef4c 100644
--- a/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-fp.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/fixed-vectors-fp.ll
@@ -1318,11 +1318,10 @@ define void @sqrt_v6bf16(ptr %x) {
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    vsetivli zero, 6, e16, m1, ta, ma
 ; CHECK-NEXT:    vle16.v v8, (a0)
-; CHECK-NEXT:    vsetivli zero, 8, e16, m1, ta, ma
 ; CHECK-NEXT:    vfwcvtbf16.f.f.v v10, v8
 ; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; CHECK-NEXT:    vfsqrt.v v8, v10
-; CHECK-NEXT:    vsetivli zero, 6, e16, m1, ta, ma
+; CHECK-NEXT:    vsetvli zero, zero, e16, m1, ta, ma
 ; CHECK-NEXT:    vfncvtbf16.f.f.w v10, v8
 ; CHECK-NEXT:    vse16.v v10, (a0)
 ; CHECK-NEXT:    ret
@@ -1371,11 +1370,10 @@ define void @sqrt_v6f16(ptr %x) {
 ; ZVFHMIN:       # %bb.0:
 ; ZVFHMIN-NEXT:    vsetivli zero, 6, e16, m1, ta, ma
 ; ZVFHMIN-NEXT:    vle16.v v8, (a0)
-; ZVFHMIN-NEXT:    vsetivli zero, 8, e16, m1, ta, ma
 ; ZVFHMIN-NEXT:    vfwcvt.f.f.v v10, v8
 ; ZVFHMIN-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
 ; ZVFHMIN-NEXT:    vfsqrt.v v8, v10
-; ZVFHMIN-NEXT:    vsetivli zero, 6, e16, m1, ta, ma
+; ZVFHMIN-NEXT:    vsetvli zero, zero, e16, m1, ta, ma
 ; ZVFHMIN-NEXT:    vfncvt.f.f.w v10, v8
 ; ZVFHMIN-NEXT:    vse16.v v10, (a0)
 ; ZVFHMIN-NEXT:    ret
diff --git a/llvm/test/CodeGen/RISCV/rvv/vl-opt.mir b/llvm/test/CodeGen/RISCV/rvv/vl-opt.mir
index 0475a988e9851..3729441db8a71 100644
--- a/llvm/test/CodeGen/RISCV/rvv/vl-opt.mir
+++ b/llvm/test/CodeGen/RISCV/rvv/vl-opt.mir
@@ -141,6 +141,26 @@ body: |
     %y:vr = PseudoVADD_VV_M1 $noreg, %x, $noreg, 1, 4 /* e16 */, 0
 ...
 ---
+name: vfsqr
+body: |
+  bb.0:
+    ; CHECK-LABEL: name: vfsqr
+    ; CHECK: %x:vrm2 = nofpexcept PseudoVFSQRT_V_M2_E32 $noreg, $noreg, 7, 6, 5 /* e32 */, 3 /* ta, ma */, implicit $frm
+    ; CHECK-NEXT: early-clobber %y:vr = nofpexcept PseudoVFNCVTBF16_F_F_W_M1_E16 $noreg, %x, 7, 6, 4 /* e16 */, 3 /* ta, ma */, implicit $frm
+    %x:vrm2 = nofpexcept PseudoVFSQRT_V_M2_E32 $noreg, $noreg, 7, 8, 5, 3, implicit $frm
+    early-clobber %y:vr = nofpexcept PseudoVFNCVTBF16_F_F_W_M1_E16 $noreg, %x, 7, 6, 4, 3, implicit $frm
+...
+---
+name: vfsqr_nofpexcept
+body: |
+  bb.0:
+    ; CHECK-LABEL: name: vfsqr_nofpexcept
+    ; CHECK: %x:vrm2 = PseudoVFSQRT_V_M2_E32 $noreg, $noreg, 7, 8, 5 /* e32 */, 3 /* ta, ma */, implicit $frm
+    ; CHECK-NEXT: early-clobber %y:vr = nofpexcept PseudoVFNCVTBF16_F_F_W_M1_E16 $noreg, %x, 7, 6, 4 /* e16 */, 3 /* ta, ma */, implicit $frm
+    %x:vrm2 = PseudoVFSQRT_V_M2_E32 $noreg, $noreg, 7, 8, 5, 3, implicit $frm
+    early-clobber %y:vr = nofpexcept PseudoVFNCVTBF16_F_F_W_M1_E16 $noreg, %x, 7, 6, 4, 3, implicit $frm
+...
+---
 name: vwadd_tied_vs1
 body: |
   bb.0:

``````````

</details>


https://github.com/llvm/llvm-project/pull/127462