diff --git a/llvm/include/llvm/Analysis/TargetTransformInfo.h b/llvm/include/llvm/Analysis/TargetTransformInfo.h
--- a/llvm/include/llvm/Analysis/TargetTransformInfo.h
+++ b/llvm/include/llvm/Analysis/TargetTransformInfo.h
@@ -613,6 +613,9 @@
   /// cost should return false, otherwise return true.
   bool isNumRegsMajorCostOfLSR() const;
 
+  /// Return true if LSR attempts to replace primary IV with other IV.
+  bool isAllowTerminatingConditionFoldingAfterLSR() const;
+
   /// \returns true if LSR should not optimize a chain that includes \p I.
   bool isProfitableLSRChainElement(Instruction *I) const;
 
@@ -1622,6 +1625,7 @@
   virtual bool isLSRCostLess(const TargetTransformInfo::LSRCost &C1,
                              const TargetTransformInfo::LSRCost &C2) = 0;
   virtual bool isNumRegsMajorCostOfLSR() = 0;
+  virtual bool isAllowTerminatingConditionFoldingAfterLSR() const = 0;
   virtual bool isProfitableLSRChainElement(Instruction *I) = 0;
   virtual bool canMacroFuseCmp() = 0;
   virtual bool canSaveCmp(Loop *L, BranchInst **BI, ScalarEvolution *SE,
@@ -2035,6 +2039,9 @@
   bool isNumRegsMajorCostOfLSR() override {
     return Impl.isNumRegsMajorCostOfLSR();
   }
+  bool isAllowTerminatingConditionFoldingAfterLSR() const override {
+    return Impl.isAllowTerminatingConditionFoldingAfterLSR();
+  }
   bool isProfitableLSRChainElement(Instruction *I) override {
     return Impl.isProfitableLSRChainElement(I);
   }
diff --git a/llvm/include/llvm/Analysis/TargetTransformInfoImpl.h b/llvm/include/llvm/Analysis/TargetTransformInfoImpl.h
--- a/llvm/include/llvm/Analysis/TargetTransformInfoImpl.h
+++ b/llvm/include/llvm/Analysis/TargetTransformInfoImpl.h
@@ -220,6 +220,8 @@
 
   bool isNumRegsMajorCostOfLSR() const { return true; }
 
+  bool isAllowTerminatingConditionFoldingAfterLSR() const { return false; }
+
   bool isProfitableLSRChainElement(Instruction *I) const { return false; }
 
   bool canMacroFuseCmp() const { return false; }
diff --git a/llvm/include/llvm/CodeGen/BasicTTIImpl.h b/llvm/include/llvm/CodeGen/BasicTTIImpl.h
--- a/llvm/include/llvm/CodeGen/BasicTTIImpl.h
+++ b/llvm/include/llvm/CodeGen/BasicTTIImpl.h
@@ -356,6 +356,11 @@
     return TargetTransformInfoImplBase::isNumRegsMajorCostOfLSR();
   }
 
+  bool isAllowTerminatingConditionFoldingAfterLSR() const {
+    return TargetTransformInfoImplBase::
+        isAllowTerminatingConditionFoldingAfterLSR();
+  }
+
   bool isProfitableLSRChainElement(Instruction *I) {
     return TargetTransformInfoImplBase::isProfitableLSRChainElement(I);
   }
diff --git a/llvm/lib/Analysis/TargetTransformInfo.cpp b/llvm/lib/Analysis/TargetTransformInfo.cpp
--- a/llvm/lib/Analysis/TargetTransformInfo.cpp
+++ b/llvm/lib/Analysis/TargetTransformInfo.cpp
@@ -364,6 +364,10 @@
   return TTIImpl->isNumRegsMajorCostOfLSR();
 }
 
+bool TargetTransformInfo::isAllowTerminatingConditionFoldingAfterLSR() const {
+  return TTIImpl->isAllowTerminatingConditionFoldingAfterLSR();
+}
+
 bool TargetTransformInfo::isProfitableLSRChainElement(Instruction *I) const {
   return TTIImpl->isProfitableLSRChainElement(I);
 }
diff --git a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.h b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.h
--- a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.h
+++ b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.h
@@ -308,6 +308,8 @@
     }
     llvm_unreachable("unknown register class");
   }
+
+  bool isAllowTerminatingConditionFoldingAfterLSR() const;
 };
 
 } // end namespace llvm
diff --git a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
--- a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
+++ b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
@@ -968,3 +968,7 @@
   // TODO: Figure out constant materialization cost modeling and remove.
   return SLPMaxVF;
 }
+
+bool RISCVTTIImpl::isAllowTerminatingConditionFoldingAfterLSR() const {
+  return true;
+}
diff --git a/llvm/lib/Transforms/Scalar/LoopStrengthReduce.cpp b/llvm/lib/Transforms/Scalar/LoopStrengthReduce.cpp
--- a/llvm/lib/Transforms/Scalar/LoopStrengthReduce.cpp
+++ b/llvm/lib/Transforms/Scalar/LoopStrengthReduce.cpp
@@ -6771,7 +6771,8 @@
     }
   }
 
-  if (AllowTerminatingConditionFoldingAfterLSR) {
+  if (AllowTerminatingConditionFoldingAfterLSR ||
+      TTI.isAllowTerminatingConditionFoldingAfterLSR()) {
     auto CanFoldTerminatingCondition = canFoldTermCondOfLoop(L, SE, DT, LI);
     if (CanFoldTerminatingCondition) {
       BasicBlock *LoopPreheader = L->getLoopPreheader();
diff --git a/llvm/test/CodeGen/RISCV/rvv/fixed-vector-strided-load-store-asm.ll b/llvm/test/CodeGen/RISCV/rvv/fixed-vector-strided-load-store-asm.ll
--- a/llvm/test/CodeGen/RISCV/rvv/fixed-vector-strided-load-store-asm.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/fixed-vector-strided-load-store-asm.ll
@@ -778,19 +778,18 @@
 ; CHECK-NEXT:    add a6, a6, a2
 ; CHECK-NEXT:    add a2, a0, a2
 ; CHECK-NEXT:    add a6, a1, a6
-; CHECK-NEXT:    li a7, 32
-; CHECK-NEXT:    li t0, 5
-; CHECK-NEXT:    mv t1, a5
+; CHECK-NEXT:    add a7, a3, a0
+; CHECK-NEXT:    li t0, 32
+; CHECK-NEXT:    li t1, 5
 ; CHECK-NEXT:  .LBB12_3: # =>This Inner Loop Header: Depth=1
-; CHECK-NEXT:    vsetvli zero, a7, e8, m1, ta, mu
-; CHECK-NEXT:    vlse8.v v8, (a6), t0
+; CHECK-NEXT:    vsetvli zero, t0, e8, m1, ta, mu
+; CHECK-NEXT:    vlse8.v v8, (a6), t1
 ; CHECK-NEXT:    vle8.v v9, (a2)
 ; CHECK-NEXT:    vadd.vv v8, v9, v8
 ; CHECK-NEXT:    vse8.v v8, (a2)
-; CHECK-NEXT:    addi t1, t1, -32
 ; CHECK-NEXT:    addi a2, a2, 32
 ; CHECK-NEXT:    addi a6, a6, 160
-; CHECK-NEXT:    bnez t1, .LBB12_3
+; CHECK-NEXT:    bne a2, a7, .LBB12_3
 ; CHECK-NEXT:  # %bb.4:
 ; CHECK-NEXT:    beq a4, a5, .LBB12_7
 ; CHECK-NEXT:  .LBB12_5:
diff --git a/llvm/test/CodeGen/RISCV/rvv/sink-splat-operands.ll b/llvm/test/CodeGen/RISCV/rvv/sink-splat-operands.ll
--- a/llvm/test/CodeGen/RISCV/rvv/sink-splat-operands.ll
+++ b/llvm/test/CodeGen/RISCV/rvv/sink-splat-operands.ll
@@ -283,17 +283,17 @@
 ; CHECK-NEXT:  # %bb.4: # %middle.block
 ; CHECK-NEXT:    beqz a4, .LBB7_7
 ; CHECK-NEXT:  .LBB7_5: # %for.body.preheader
-; CHECK-NEXT:    addi a3, a2, -1024
 ; CHECK-NEXT:    slli a2, a2, 2
-; CHECK-NEXT:    add a0, a0, a2
+; CHECK-NEXT:    add a2, a0, a2
+; CHECK-NEXT:    lui a3, 1
+; CHECK-NEXT:    add a0, a0, a3
 ; CHECK-NEXT:  .LBB7_6: # %for.body
 ; CHECK-NEXT:    # =>This Inner Loop Header: Depth=1
-; CHECK-NEXT:    lw a2, 0(a0)
-; CHECK-NEXT:    mulw a2, a2, a1
-; CHECK-NEXT:    sw a2, 0(a0)
-; CHECK-NEXT:    addi a3, a3, 1
-; CHECK-NEXT:    addi a0, a0, 4
-; CHECK-NEXT:    bnez a3, .LBB7_6
+; CHECK-NEXT:    lw a3, 0(a2)
+; CHECK-NEXT:    mulw a3, a3, a1
+; CHECK-NEXT:    sw a3, 0(a2)
+; CHECK-NEXT:    addi a2, a2, 4
+; CHECK-NEXT:    bne a2, a0, .LBB7_6
 ; CHECK-NEXT:  .LBB7_7: # %for.cond.cleanup
 ; CHECK-NEXT:    ret
 entry:
@@ -376,17 +376,17 @@
 ; CHECK-NEXT:  # %bb.4: # %middle.block
 ; CHECK-NEXT:    beqz a4, .LBB8_7
 ; CHECK-NEXT:  .LBB8_5: # %for.body.preheader
-; CHECK-NEXT:    addi a3, a2, -1024
 ; CHECK-NEXT:    slli a2, a2, 2
-; CHECK-NEXT:    add a0, a0, a2
+; CHECK-NEXT:    add a2, a0, a2
+; CHECK-NEXT:    lui a3, 1
+; CHECK-NEXT:    add a0, a0, a3
 ; CHECK-NEXT:  .LBB8_6: # %for.body
 ; CHECK-NEXT:    # =>This Inner Loop Header: Depth=1
-; CHECK-NEXT:    lw a2, 0(a0)
-; CHECK-NEXT:    addw a2, a2, a1
-; CHECK-NEXT:    sw a2, 0(a0)
-; CHECK-NEXT:    addi a3, a3, 1
-; CHECK-NEXT:    addi a0, a0, 4
-; CHECK-NEXT:    bnez a3, .LBB8_6
+; CHECK-NEXT:    lw a3, 0(a2)
+; CHECK-NEXT:    addw a3, a3, a1
+; CHECK-NEXT:    sw a3, 0(a2)
+; CHECK-NEXT:    addi a2, a2, 4
+; CHECK-NEXT:    bne a2, a0, .LBB8_6
 ; CHECK-NEXT:  .LBB8_7: # %for.cond.cleanup
 ; CHECK-NEXT:    ret
 entry:
@@ -469,17 +469,17 @@
 ; CHECK-NEXT:  # %bb.4: # %middle.block
 ; CHECK-NEXT:    beqz a4, .LBB9_7
 ; CHECK-NEXT:  .LBB9_5: # %for.body.preheader
-; CHECK-NEXT:    addi a3, a2, -1024
 ; CHECK-NEXT:    slli a2, a2, 2
-; CHECK-NEXT:    add a0, a0, a2
+; CHECK-NEXT:    add a2, a0, a2
+; CHECK-NEXT:    lui a3, 1
+; CHECK-NEXT:    add a0, a0, a3
 ; CHECK-NEXT:  .LBB9_6: # %for.body
 ; CHECK-NEXT:    # =>This Inner Loop Header: Depth=1
-; CHECK-NEXT:    lw a2, 0(a0)
-; CHECK-NEXT:    addw a2, a2, a1
-; CHECK-NEXT:    sw a2, 0(a0)
-; CHECK-NEXT:    addi a3, a3, 1
-; CHECK-NEXT:    addi a0, a0, 4
-; CHECK-NEXT:    bnez a3, .LBB9_6
+; CHECK-NEXT:    lw a3, 0(a2)
+; CHECK-NEXT:    addw a3, a3, a1
+; CHECK-NEXT:    sw a3, 0(a2)
+; CHECK-NEXT:    addi a2, a2, 4
+; CHECK-NEXT:    bne a2, a0, .LBB9_6
 ; CHECK-NEXT:  .LBB9_7: # %for.cond.cleanup
 ; CHECK-NEXT:    ret
 entry:
@@ -562,17 +562,17 @@
 ; CHECK-NEXT:  # %bb.4: # %middle.block
 ; CHECK-NEXT:    beqz a4, .LBB10_7
 ; CHECK-NEXT:  .LBB10_5: # %for.body.preheader
-; CHECK-NEXT:    addi a3, a2, -1024
 ; CHECK-NEXT:    slli a2, a2, 2
-; CHECK-NEXT:    add a0, a0, a2
+; CHECK-NEXT:    add a2, a0, a2
+; CHECK-NEXT:    lui a3, 1
+; CHECK-NEXT:    add a0, a0, a3
 ; CHECK-NEXT:  .LBB10_6: # %for.body
 ; CHECK-NEXT:    # =>This Inner Loop Header: Depth=1
-; CHECK-NEXT:    lw a2, 0(a0)
-; CHECK-NEXT:    subw a2, a1, a2
-; CHECK-NEXT:    sw a2, 0(a0)
-; CHECK-NEXT:    addi a3, a3, 1
-; CHECK-NEXT:    addi a0, a0, 4
-; CHECK-NEXT:    bnez a3, .LBB10_6
+; CHECK-NEXT:    lw a3, 0(a2)
+; CHECK-NEXT:    subw a3, a1, a3
+; CHECK-NEXT:    sw a3, 0(a2)
+; CHECK-NEXT:    addi a2, a2, 4
+; CHECK-NEXT:    bne a2, a0, .LBB10_6
 ; CHECK-NEXT:  .LBB10_7: # %for.cond.cleanup
 ; CHECK-NEXT:    ret
 entry:
@@ -655,17 +655,17 @@
 ; CHECK-NEXT:  # %bb.4: # %middle.block
 ; CHECK-NEXT:    beqz a4, .LBB11_7
 ; CHECK-NEXT:  .LBB11_5: # %for.body.preheader
-; CHECK-NEXT:    addi a3, a2, -1024
 ; CHECK-NEXT:    slli a2, a2, 2
-; CHECK-NEXT:    add a0, a0, a2
+; CHECK-NEXT:    add a2, a0, a2
+; CHECK-NEXT:    lui a3, 1
+; CHECK-NEXT:    add a0, a0, a3
 ; CHECK-NEXT:  .LBB11_6: # %for.body
 ; CHECK-NEXT:    # =>This Inner Loop Header: Depth=1
-; CHECK-NEXT:    lw a2, 0(a0)
-; CHECK-NEXT:    and a2, a2, a1
-; CHECK-NEXT:    sw a2, 0(a0)
-; CHECK-NEXT:    addi a3, a3, 1
-; CHECK-NEXT:    addi a0, a0, 4
-; CHECK-NEXT:    bnez a3, .LBB11_6
+; CHECK-NEXT:    lw a3, 0(a2)
+; CHECK-NEXT:    and a3, a3, a1
+; CHECK-NEXT:    sw a3, 0(a2)
+; CHECK-NEXT:    addi a2, a2, 4
+; CHECK-NEXT:    bne a2, a0, .LBB11_6
 ; CHECK-NEXT:  .LBB11_7: # %for.cond.cleanup
 ; CHECK-NEXT:    ret
 entry:
@@ -748,17 +748,17 @@
 ; CHECK-NEXT:  # %bb.4: # %middle.block
 ; CHECK-NEXT:    beqz a4, .LBB12_7
 ; CHECK-NEXT:  .LBB12_5: # %for.body.preheader
-; CHECK-NEXT:    addi a3, a2, -1024
 ; CHECK-NEXT:    slli a2, a2, 2
-; CHECK-NEXT:    add a0, a0, a2
+; CHECK-NEXT:    add a2, a0, a2
+; CHECK-NEXT:    lui a3, 1
+; CHECK-NEXT:    add a0, a0, a3
 ; CHECK-NEXT:  .LBB12_6: # %for.body
 ; CHECK-NEXT:    # =>This Inner Loop Header: Depth=1
-; CHECK-NEXT:    lw a2, 0(a0)
-; CHECK-NEXT:    or a2, a2, a1
-; CHECK-NEXT:    sw a2, 0(a0)
-; CHECK-NEXT:    addi a3, a3, 1
-; CHECK-NEXT:    addi a0, a0, 4
-; CHECK-NEXT:    bnez a3, .LBB12_6
+; CHECK-NEXT:    lw a3, 0(a2)
+; CHECK-NEXT:    or a3, a3, a1
+; CHECK-NEXT:    sw a3, 0(a2)
+; CHECK-NEXT:    addi a2, a2, 4
+; CHECK-NEXT:    bne a2, a0, .LBB12_6
 ; CHECK-NEXT:  .LBB12_7: # %for.cond.cleanup
 ; CHECK-NEXT:    ret
 entry:
@@ -841,17 +841,17 @@
 ; CHECK-NEXT:  # %bb.4: # %middle.block
 ; CHECK-NEXT:    beqz a4, .LBB13_7
 ; CHECK-NEXT:  .LBB13_5: # %for.body.preheader
-; CHECK-NEXT:    addi a3, a2, -1024
 ; CHECK-NEXT:    slli a2, a2, 2
-; CHECK-NEXT:    add a0, a0, a2
+; CHECK-NEXT:    add a2, a0, a2
+; CHECK-NEXT:    lui a3, 1
+; CHECK-NEXT:    add a0, a0, a3
 ; CHECK-NEXT:  .LBB13_6: # %for.body
 ; CHECK-NEXT:    # =>This Inner Loop Header: Depth=1
-; CHECK-NEXT:    lw a2, 0(a0)
-; CHECK-NEXT:    xor a2, a2, a1
-; CHECK-NEXT:    sw a2, 0(a0)
-; CHECK-NEXT:    addi a3, a3, 1
-; CHECK-NEXT:    addi a0, a0, 4
-; CHECK-NEXT:    bnez a3, .LBB13_6
+; CHECK-NEXT:    lw a3, 0(a2)
+; CHECK-NEXT:    xor a3, a3, a1
+; CHECK-NEXT:    sw a3, 0(a2)
+; CHECK-NEXT:    addi a2, a2, 4
+; CHECK-NEXT:    bne a2, a0, .LBB13_6
 ; CHECK-NEXT:  .LBB13_7: # %for.cond.cleanup
 ; CHECK-NEXT:    ret
 entry:
@@ -1042,17 +1042,17 @@
 ; CHECK-NEXT:  # %bb.4: # %middle.block
 ; CHECK-NEXT:    beqz a4, .LBB17_7
 ; CHECK-NEXT:  .LBB17_5: # %for.body.preheader
-; CHECK-NEXT:    addi a3, a2, -1024
 ; CHECK-NEXT:    slli a2, a2, 2
-; CHECK-NEXT:    add a0, a0, a2
+; CHECK-NEXT:    add a2, a0, a2
+; CHECK-NEXT:    lui a3, 1
+; CHECK-NEXT:    add a0, a0, a3
 ; CHECK-NEXT:  .LBB17_6: # %for.body
 ; CHECK-NEXT:    # =>This Inner Loop Header: Depth=1
-; CHECK-NEXT:    lw a2, 0(a0)
-; CHECK-NEXT:    sllw a2, a2, a1
-; CHECK-NEXT:    sw a2, 0(a0)
-; CHECK-NEXT:    addi a3, a3, 1
-; CHECK-NEXT:    addi a0, a0, 4
-; CHECK-NEXT:    bnez a3, .LBB17_6
+; CHECK-NEXT:    lw a3, 0(a2)
+; CHECK-NEXT:    sllw a3, a3, a1
+; CHECK-NEXT:    sw a3, 0(a2)
+; CHECK-NEXT:    addi a2, a2, 4
+; CHECK-NEXT:    bne a2, a0, .LBB17_6
 ; CHECK-NEXT:  .LBB17_7: # %for.cond.cleanup
 ; CHECK-NEXT:    ret
 entry:
@@ -1135,17 +1135,17 @@
 ; CHECK-NEXT:  # %bb.4: # %middle.block
 ; CHECK-NEXT:    beqz a4, .LBB18_7
 ; CHECK-NEXT:  .LBB18_5: # %for.body.preheader
-; CHECK-NEXT:    addi a3, a2, -1024
 ; CHECK-NEXT:    slli a2, a2, 2
-; CHECK-NEXT:    add a0, a0, a2
+; CHECK-NEXT:    add a2, a0, a2
+; CHECK-NEXT:    lui a3, 1
+; CHECK-NEXT:    add a0, a0, a3
 ; CHECK-NEXT:  .LBB18_6: # %for.body
 ; CHECK-NEXT:    # =>This Inner Loop Header: Depth=1
-; CHECK-NEXT:    lw a2, 0(a0)
-; CHECK-NEXT:    srlw a2, a2, a1
-; CHECK-NEXT:    sw a2, 0(a0)
-; CHECK-NEXT:    addi a3, a3, 1
-; CHECK-NEXT:    addi a0, a0, 4
-; CHECK-NEXT:    bnez a3, .LBB18_6
+; CHECK-NEXT:    lw a3, 0(a2)
+; CHECK-NEXT:    srlw a3, a3, a1
+; CHECK-NEXT:    sw a3, 0(a2)
+; CHECK-NEXT:    addi a2, a2, 4
+; CHECK-NEXT:    bne a2, a0, .LBB18_6
 ; CHECK-NEXT:  .LBB18_7: # %for.cond.cleanup
 ; CHECK-NEXT:    ret
 entry:
@@ -1228,17 +1228,17 @@
 ; CHECK-NEXT:  # %bb.4: # %middle.block
 ; CHECK-NEXT:    beqz a3, .LBB19_7
 ; CHECK-NEXT:  .LBB19_5: # %for.body.preheader
-; CHECK-NEXT:    addi a2, a1, -1024
 ; CHECK-NEXT:    slli a1, a1, 2
-; CHECK-NEXT:    add a0, a0, a1
+; CHECK-NEXT:    add a1, a0, a1
+; CHECK-NEXT:    lui a2, 1
+; CHECK-NEXT:    add a0, a0, a2
 ; CHECK-NEXT:  .LBB19_6: # %for.body
 ; CHECK-NEXT:    # =>This Inner Loop Header: Depth=1
-; CHECK-NEXT:    lw a1, 0(a0)
-; CHECK-NEXT:    srli a1, a1, 2
-; CHECK-NEXT:    sw a1, 0(a0)
-; CHECK-NEXT:    addi a2, a2, 1
-; CHECK-NEXT:    addi a0, a0, 4
-; CHECK-NEXT:    bnez a2, .LBB19_6
+; CHECK-NEXT:    lw a2, 0(a1)
+; CHECK-NEXT:    srli a2, a2, 2
+; CHECK-NEXT:    sw a2, 0(a1)
+; CHECK-NEXT:    addi a1, a1, 4
+; CHECK-NEXT:    bne a1, a0, .LBB19_6
 ; CHECK-NEXT:  .LBB19_7: # %for.cond.cleanup
 ; CHECK-NEXT:    ret
 entry:
@@ -1536,17 +1536,17 @@
 ; CHECK-NEXT:  # %bb.4: # %middle.block
 ; CHECK-NEXT:    beqz a4, .LBB26_7
 ; CHECK-NEXT:  .LBB26_5: # %for.body.preheader
-; CHECK-NEXT:    addi a2, a1, -1024
 ; CHECK-NEXT:    slli a1, a1, 2
-; CHECK-NEXT:    add a0, a0, a1
+; CHECK-NEXT:    add a1, a0, a1
+; CHECK-NEXT:    lui a2, 1
+; CHECK-NEXT:    add a0, a0, a2
 ; CHECK-NEXT:  .LBB26_6: # %for.body
 ; CHECK-NEXT:    # =>This Inner Loop Header: Depth=1
-; CHECK-NEXT:    flw ft0, 0(a0)
+; CHECK-NEXT:    flw ft0, 0(a1)
 ; CHECK-NEXT:    fmul.s ft0, ft0, fa0
-; CHECK-NEXT:    fsw ft0, 0(a0)
-; CHECK-NEXT:    addi a2, a2, 1
-; CHECK-NEXT:    addi a0, a0, 4
-; CHECK-NEXT:    bnez a2, .LBB26_6
+; CHECK-NEXT:    fsw ft0, 0(a1)
+; CHECK-NEXT:    addi a1, a1, 4
+; CHECK-NEXT:    bne a1, a0, .LBB26_6
 ; CHECK-NEXT:  .LBB26_7: # %for.cond.cleanup
 ; CHECK-NEXT:    ret
 entry:
@@ -1628,17 +1628,17 @@
 ; CHECK-NEXT:  # %bb.4: # %middle.block
 ; CHECK-NEXT:    beqz a4, .LBB27_7
 ; CHECK-NEXT:  .LBB27_5: # %for.body.preheader
-; CHECK-NEXT:    addi a2, a1, -1024
 ; CHECK-NEXT:    slli a1, a1, 2
-; CHECK-NEXT:    add a0, a0, a1
+; CHECK-NEXT:    add a1, a0, a1
+; CHECK-NEXT:    lui a2, 1
+; CHECK-NEXT:    add a0, a0, a2
 ; CHECK-NEXT:  .LBB27_6: # %for.body
 ; CHECK-NEXT:    # =>This Inner Loop Header: Depth=1
-; CHECK-NEXT:    flw ft0, 0(a0)
+; CHECK-NEXT:    flw ft0, 0(a1)
 ; CHECK-NEXT:    fdiv.s ft0, ft0, fa0
-; CHECK-NEXT:    fsw ft0, 0(a0)
-; CHECK-NEXT:    addi a2, a2, 1
-; CHECK-NEXT:    addi a0, a0, 4
-; CHECK-NEXT:    bnez a2, .LBB27_6
+; CHECK-NEXT:    fsw ft0, 0(a1)
+; CHECK-NEXT:    addi a1, a1, 4
+; CHECK-NEXT:    bne a1, a0, .LBB27_6
 ; CHECK-NEXT:  .LBB27_7: # %for.cond.cleanup
 ; CHECK-NEXT:    ret
 entry:
@@ -1720,17 +1720,17 @@
 ; CHECK-NEXT:  # %bb.4: # %middle.block
 ; CHECK-NEXT:    beqz a4, .LBB28_7
 ; CHECK-NEXT:  .LBB28_5: # %for.body.preheader
-; CHECK-NEXT:    addi a2, a1, -1024
 ; CHECK-NEXT:    slli a1, a1, 2
-; CHECK-NEXT:    add a0, a0, a1
+; CHECK-NEXT:    add a1, a0, a1
+; CHECK-NEXT:    lui a2, 1
+; CHECK-NEXT:    add a0, a0, a2
 ; CHECK-NEXT:  .LBB28_6: # %for.body
 ; CHECK-NEXT:    # =>This Inner Loop Header: Depth=1
-; CHECK-NEXT:    flw ft0, 0(a0)
+; CHECK-NEXT:    flw ft0, 0(a1)
 ; CHECK-NEXT:    fdiv.s ft0, fa0, ft0
-; CHECK-NEXT:    fsw ft0, 0(a0)
-; CHECK-NEXT:    addi a2, a2, 1
-; CHECK-NEXT:    addi a0, a0, 4
-; CHECK-NEXT:    bnez a2, .LBB28_6
+; CHECK-NEXT:    fsw ft0, 0(a1)
+; CHECK-NEXT:    addi a1, a1, 4
+; CHECK-NEXT:    bne a1, a0, .LBB28_6
 ; CHECK-NEXT:  .LBB28_7: # %for.cond.cleanup
 ; CHECK-NEXT:    ret
 entry:
@@ -1812,17 +1812,17 @@
 ; CHECK-NEXT:  # %bb.4: # %middle.block
 ; CHECK-NEXT:    beqz a4, .LBB29_7
 ; CHECK-NEXT:  .LBB29_5: # %for.body.preheader
-; CHECK-NEXT:    addi a2, a1, -1024
 ; CHECK-NEXT:    slli a1, a1, 2
-; CHECK-NEXT:    add a0, a0, a1
+; CHECK-NEXT:    add a1, a0, a1
+; CHECK-NEXT:    lui a2, 1
+; CHECK-NEXT:    add a0, a0, a2
 ; CHECK-NEXT:  .LBB29_6: # %for.body
 ; CHECK-NEXT:    # =>This Inner Loop Header: Depth=1
-; CHECK-NEXT:    flw ft0, 0(a0)
+; CHECK-NEXT:    flw ft0, 0(a1)
 ; CHECK-NEXT:    fadd.s ft0, ft0, fa0
-; CHECK-NEXT:    fsw ft0, 0(a0)
-; CHECK-NEXT:    addi a2, a2, 1
-; CHECK-NEXT:    addi a0, a0, 4
-; CHECK-NEXT:    bnez a2, .LBB29_6
+; CHECK-NEXT:    fsw ft0, 0(a1)
+; CHECK-NEXT:    addi a1, a1, 4
+; CHECK-NEXT:    bne a1, a0, .LBB29_6
 ; CHECK-NEXT:  .LBB29_7: # %for.cond.cleanup
 ; CHECK-NEXT:    ret
 entry:
@@ -1904,17 +1904,17 @@
 ; CHECK-NEXT:  # %bb.4: # %middle.block
 ; CHECK-NEXT:    beqz a4, .LBB30_7
 ; CHECK-NEXT:  .LBB30_5: # %for.body.preheader
-; CHECK-NEXT:    addi a2, a1, -1024
 ; CHECK-NEXT:    slli a1, a1, 2
-; CHECK-NEXT:    add a0, a0, a1
+; CHECK-NEXT:    add a1, a0, a1
+; CHECK-NEXT:    lui a2, 1
+; CHECK-NEXT:    add a0, a0, a2
 ; CHECK-NEXT:  .LBB30_6: # %for.body
 ; CHECK-NEXT:    # =>This Inner Loop Header: Depth=1
-; CHECK-NEXT:    flw ft0, 0(a0)
+; CHECK-NEXT:    flw ft0, 0(a1)
 ; CHECK-NEXT:    fsub.s ft0, ft0, fa0
-; CHECK-NEXT:    fsw ft0, 0(a0)
-; CHECK-NEXT:    addi a2, a2, 1
-; CHECK-NEXT:    addi a0, a0, 4
-; CHECK-NEXT:    bnez a2, .LBB30_6
+; CHECK-NEXT:    fsw ft0, 0(a1)
+; CHECK-NEXT:    addi a1, a1, 4
+; CHECK-NEXT:    bne a1, a0, .LBB30_6
 ; CHECK-NEXT:  .LBB30_7: # %for.cond.cleanup
 ; CHECK-NEXT:    ret
 entry:
@@ -1996,17 +1996,17 @@
 ; CHECK-NEXT:  # %bb.4: # %middle.block
 ; CHECK-NEXT:    beqz a4, .LBB31_7
 ; CHECK-NEXT:  .LBB31_5: # %for.body.preheader
-; CHECK-NEXT:    addi a2, a1, -1024
 ; CHECK-NEXT:    slli a1, a1, 2
-; CHECK-NEXT:    add a0, a0, a1
+; CHECK-NEXT:    add a1, a0, a1
+; CHECK-NEXT:    lui a2, 1
+; CHECK-NEXT:    add a0, a0, a2
 ; CHECK-NEXT:  .LBB31_6: # %for.body
 ; CHECK-NEXT:    # =>This Inner Loop Header: Depth=1
-; CHECK-NEXT:    flw ft0, 0(a0)
+; CHECK-NEXT:    flw ft0, 0(a1)
 ; CHECK-NEXT:    fsub.s ft0, fa0, ft0
-; CHECK-NEXT:    fsw ft0, 0(a0)
-; CHECK-NEXT:    addi a2, a2, 1
-; CHECK-NEXT:    addi a0, a0, 4
-; CHECK-NEXT:    bnez a2, .LBB31_6
+; CHECK-NEXT:    fsw ft0, 0(a1)
+; CHECK-NEXT:    addi a1, a1, 4
+; CHECK-NEXT:    bne a1, a0, .LBB31_6
 ; CHECK-NEXT:  .LBB31_7: # %for.cond.cleanup
 ; CHECK-NEXT:    ret
 entry:
@@ -2173,20 +2173,20 @@
 ; CHECK-NEXT:  # %bb.4: # %middle.block
 ; CHECK-NEXT:    beqz a5, .LBB34_7
 ; CHECK-NEXT:  .LBB34_5: # %for.body.preheader
-; CHECK-NEXT:    addi a2, a4, -1024
 ; CHECK-NEXT:    slli a3, a4, 2
-; CHECK-NEXT:    add a1, a1, a3
+; CHECK-NEXT:    add a2, a1, a3
 ; CHECK-NEXT:    add a0, a0, a3
+; CHECK-NEXT:    lui a3, 1
+; CHECK-NEXT:    add a1, a1, a3
 ; CHECK-NEXT:  .LBB34_6: # %for.body
 ; CHECK-NEXT:    # =>This Inner Loop Header: Depth=1
 ; CHECK-NEXT:    flw ft0, 0(a0)
-; CHECK-NEXT:    flw ft1, 0(a1)
+; CHECK-NEXT:    flw ft1, 0(a2)
 ; CHECK-NEXT:    fmadd.s ft0, ft0, fa0, ft1
 ; CHECK-NEXT:    fsw ft0, 0(a0)
-; CHECK-NEXT:    addi a2, a2, 1
-; CHECK-NEXT:    addi a1, a1, 4
+; CHECK-NEXT:    addi a2, a2, 4
 ; CHECK-NEXT:    addi a0, a0, 4
-; CHECK-NEXT:    bnez a2, .LBB34_6
+; CHECK-NEXT:    bne a2, a1, .LBB34_6
 ; CHECK-NEXT:  .LBB34_7: # %for.cond.cleanup
 ; CHECK-NEXT:    ret
 entry:
@@ -2276,20 +2276,20 @@
 ; CHECK-NEXT:  # %bb.4: # %middle.block
 ; CHECK-NEXT:    beqz a5, .LBB35_7
 ; CHECK-NEXT:  .LBB35_5: # %for.body.preheader
-; CHECK-NEXT:    addi a2, a4, -1024
 ; CHECK-NEXT:    slli a3, a4, 2
-; CHECK-NEXT:    add a1, a1, a3
+; CHECK-NEXT:    add a2, a1, a3
 ; CHECK-NEXT:    add a0, a0, a3
+; CHECK-NEXT:    lui a3, 1
+; CHECK-NEXT:    add a1, a1, a3
 ; CHECK-NEXT:  .LBB35_6: # %for.body
 ; CHECK-NEXT:    # =>This Inner Loop Header: Depth=1
 ; CHECK-NEXT:    flw ft0, 0(a0)
-; CHECK-NEXT:    flw ft1, 0(a1)
+; CHECK-NEXT:    flw ft1, 0(a2)
 ; CHECK-NEXT:    fmadd.s ft0, fa0, ft0, ft1
 ; CHECK-NEXT:    fsw ft0, 0(a0)
-; CHECK-NEXT:    addi a2, a2, 1
-; CHECK-NEXT:    addi a1, a1, 4
+; CHECK-NEXT:    addi a2, a2, 4
 ; CHECK-NEXT:    addi a0, a0, 4
-; CHECK-NEXT:    bnez a2, .LBB35_6
+; CHECK-NEXT:    bne a2, a1, .LBB35_6
 ; CHECK-NEXT:  .LBB35_7: # %for.cond.cleanup
 ; CHECK-NEXT:    ret
 entry:
@@ -2602,17 +2602,17 @@
 ; CHECK-NEXT:  # %bb.4: # %middle.block
 ; CHECK-NEXT:    beqz a4, .LBB42_7
 ; CHECK-NEXT:  .LBB42_5: # %for.body.preheader
-; CHECK-NEXT:    addi a3, a2, -1024
 ; CHECK-NEXT:    slli a2, a2, 2
-; CHECK-NEXT:    add a0, a0, a2
+; CHECK-NEXT:    add a2, a0, a2
+; CHECK-NEXT:    lui a3, 1
+; CHECK-NEXT:    add a0, a0, a3
 ; CHECK-NEXT:  .LBB42_6: # %for.body
 ; CHECK-NEXT:    # =>This Inner Loop Header: Depth=1
-; CHECK-NEXT:    lw a2, 0(a0)
-; CHECK-NEXT:    divuw a2, a2, a1
-; CHECK-NEXT:    sw a2, 0(a0)
-; CHECK-NEXT:    addi a3, a3, 1
-; CHECK-NEXT:    addi a0, a0, 4
-; CHECK-NEXT:    bnez a3, .LBB42_6
+; CHECK-NEXT:    lw a3, 0(a2)
+; CHECK-NEXT:    divuw a3, a3, a1
+; CHECK-NEXT:    sw a3, 0(a2)
+; CHECK-NEXT:    addi a2, a2, 4
+; CHECK-NEXT:    bne a2, a0, .LBB42_6
 ; CHECK-NEXT:  .LBB42_7: # %for.cond.cleanup
 ; CHECK-NEXT:    ret
 entry:
@@ -2695,17 +2695,17 @@
 ; CHECK-NEXT:  # %bb.4: # %middle.block
 ; CHECK-NEXT:    beqz a4, .LBB43_7
 ; CHECK-NEXT:  .LBB43_5: # %for.body.preheader
-; CHECK-NEXT:    addi a3, a2, -1024
 ; CHECK-NEXT:    slli a2, a2, 2
-; CHECK-NEXT:    add a0, a0, a2
+; CHECK-NEXT:    add a2, a0, a2
+; CHECK-NEXT:    lui a3, 1
+; CHECK-NEXT:    add a0, a0, a3
 ; CHECK-NEXT:  .LBB43_6: # %for.body
 ; CHECK-NEXT:    # =>This Inner Loop Header: Depth=1
-; CHECK-NEXT:    lw a2, 0(a0)
-; CHECK-NEXT:    divw a2, a2, a1
-; CHECK-NEXT:    sw a2, 0(a0)
-; CHECK-NEXT:    addi a3, a3, 1
-; CHECK-NEXT:    addi a0, a0, 4
-; CHECK-NEXT:    bnez a3, .LBB43_6
+; CHECK-NEXT:    lw a3, 0(a2)
+; CHECK-NEXT:    divw a3, a3, a1
+; CHECK-NEXT:    sw a3, 0(a2)
+; CHECK-NEXT:    addi a2, a2, 4
+; CHECK-NEXT:    bne a2, a0, .LBB43_6
 ; CHECK-NEXT:  .LBB43_7: # %for.cond.cleanup
 ; CHECK-NEXT:    ret
 entry:
@@ -2788,17 +2788,17 @@
 ; CHECK-NEXT:  # %bb.4: # %middle.block
 ; CHECK-NEXT:    beqz a4, .LBB44_7
 ; CHECK-NEXT:  .LBB44_5: # %for.body.preheader
-; CHECK-NEXT:    addi a3, a2, -1024
 ; CHECK-NEXT:    slli a2, a2, 2
-; CHECK-NEXT:    add a0, a0, a2
+; CHECK-NEXT:    add a2, a0, a2
+; CHECK-NEXT:    lui a3, 1
+; CHECK-NEXT:    add a0, a0, a3
 ; CHECK-NEXT:  .LBB44_6: # %for.body
 ; CHECK-NEXT:    # =>This Inner Loop Header: Depth=1
-; CHECK-NEXT:    lw a2, 0(a0)
-; CHECK-NEXT:    remuw a2, a2, a1
-; CHECK-NEXT:    sw a2, 0(a0)
-; CHECK-NEXT:    addi a3, a3, 1
-; CHECK-NEXT:    addi a0, a0, 4
-; CHECK-NEXT:    bnez a3, .LBB44_6
+; CHECK-NEXT:    lw a3, 0(a2)
+; CHECK-NEXT:    remuw a3, a3, a1
+; CHECK-NEXT:    sw a3, 0(a2)
+; CHECK-NEXT:    addi a2, a2, 4
+; CHECK-NEXT:    bne a2, a0, .LBB44_6
 ; CHECK-NEXT:  .LBB44_7: # %for.cond.cleanup
 ; CHECK-NEXT:    ret
 entry:
@@ -2881,17 +2881,17 @@
 ; CHECK-NEXT:  # %bb.4: # %middle.block
 ; CHECK-NEXT:    beqz a4, .LBB45_7
 ; CHECK-NEXT:  .LBB45_5: # %for.body.preheader
-; CHECK-NEXT:    addi a3, a2, -1024
 ; CHECK-NEXT:    slli a2, a2, 2
-; CHECK-NEXT:    add a0, a0, a2
+; CHECK-NEXT:    add a2, a0, a2
+; CHECK-NEXT:    lui a3, 1
+; CHECK-NEXT:    add a0, a0, a3
 ; CHECK-NEXT:  .LBB45_6: # %for.body
 ; CHECK-NEXT:    # =>This Inner Loop Header: Depth=1
-; CHECK-NEXT:    lw a2, 0(a0)
-; CHECK-NEXT:    remw a2, a2, a1
-; CHECK-NEXT:    sw a2, 0(a0)
-; CHECK-NEXT:    addi a3, a3, 1
-; CHECK-NEXT:    addi a0, a0, 4
-; CHECK-NEXT:    bnez a3, .LBB45_6
+; CHECK-NEXT:    lw a3, 0(a2)
+; CHECK-NEXT:    remw a3, a3, a1
+; CHECK-NEXT:    sw a3, 0(a2)
+; CHECK-NEXT:    addi a2, a2, 4
+; CHECK-NEXT:    bne a2, a0, .LBB45_6
 ; CHECK-NEXT:  .LBB45_7: # %for.cond.cleanup
 ; CHECK-NEXT:    ret
 entry: