[llvm] AMDGPU: Fix handling of -0 in round lowering (PR #65761)

[Jay Foad via llvm-commits]

================
@@ -41,6 +42,57 @@ define amdgpu_kernel void @v_fneg_add_f32(ptr addrspace(1) %out, ptr addrspace(1
 ; GCN-NEXT: flat_store_dword v{{\[[0-9]+:[0-9]+\]}}, [[ADD]]
 ; GCN-NEXT: s_waitcnt vmcnt(0)
 define amdgpu_kernel void @v_fneg_add_store_use_add_f32(ptr addrspace(1) %out, ptr addrspace(1) %a.ptr, ptr addrspace(1) %b.ptr) #0 {
----------------
jayfoad wrote:

Does this test pass? Looks like you should have deleted the manual checks above this line, now that there are autogenerated checks below this line.

https://github.com/llvm/llvm-project/pull/65761