[llvm] AMDGPU: Simplify synthesis of nextdown(1.0) constant (PR #189039)

[via llvm-commits]

llvmbot wrote:


<!--LLVM PR SUMMARY COMMENT-->

@llvm/pr-subscribers-backend-amdgpu

Author: Matt Arsenault (arsenm)

<details>
<summary>Changes</summary>



---
Full diff: https://github.com/llvm/llvm-project/pull/189039.diff


1 Files Affected:

- (modified) llvm/lib/Target/AMDGPU/AMDGPUCodeGenPrepare.cpp (+3-5) 


``````````diff
diff --git a/llvm/lib/Target/AMDGPU/AMDGPUCodeGenPrepare.cpp b/llvm/lib/Target/AMDGPU/AMDGPUCodeGenPrepare.cpp
index 79c6a50d77aec..fe5624d824f94 100644
--- a/llvm/lib/Target/AMDGPU/AMDGPUCodeGenPrepare.cpp
+++ b/llvm/lib/Target/AMDGPU/AMDGPUCodeGenPrepare.cpp
@@ -2052,13 +2052,11 @@ Value *AMDGPUCodeGenPrepareImpl::matchFractPat(IntrinsicInst &I) {
   if (!match(Arg1, m_APFloatAllowPoison(C)))
     return nullptr;
 
-  APFloat One(1.0);
-  bool LosesInfo;
-  One.convert(C->getSemantics(), APFloat::rmNearestTiesToEven, &LosesInfo);
+  APFloat OneNextDown = APFloat::getOne(C->getSemantics());
+  OneNextDown.next(true);
 
   // Match nextafter(1.0, -1)
-  One.next(true);
-  if (One != *C)
+  if (OneNextDown != *C)
     return nullptr;
 
   Value *FloorSrc;

``````````

</details>


https://github.com/llvm/llvm-project/pull/189039