[clang] [llvm] Match bitsin(typeof(x)) - popcnt(x) to s_bcnt0_i32 on AMDGPU (PR #164847)

[Matt Arsenault via llvm-commits]

================
@@ -2359,6 +2359,12 @@ def int_amdgcn_mbcnt_hi :
   DefaultAttrsIntrinsic<[llvm_i32_ty], [llvm_i32_ty, llvm_i32_ty],
             [IntrNoMem]>;
 
+def int_amdgcn_bcnt32_lo :
----------------
arsenm wrote:

This does not need an intrinsic. This doesn't help match the pattern in any cases, and introduces new support burdens to every optimization 

https://github.com/llvm/llvm-project/pull/164847