[llvm] [AMDGPU] Add wave reduce intrinsics for float types - 2 (PR #161815)

[Matt Arsenault via llvm-commits]

================
@@ -5480,11 +5480,15 @@ static uint32_t getIdentityValueFor32BitWaveReduction(unsigned Opc) {
     return std::numeric_limits<uint32_t>::min();
   case AMDGPU::S_MAX_I32:
     return std::numeric_limits<int32_t>::min();
+  case AMDGPU::V_SUB_F32_e64: // +0.0
+    return __builtin_bit_cast(uint32_t, +0.0f);
----------------
arsenm wrote:

Should not be using a raw builtin like this. It's easier to just write the raw bit value 

https://github.com/llvm/llvm-project/pull/161815