[PATCH] D41599: [X86] Lowering X86 avx512 sqrt intrinsics to IR - LLVM

[Craig Topper via Phabricator via llvm-commits]

craig.topper added a comment.

Looking at clang's CGBuiltin.cpp we do have precedent for using Intrinsic::sqrt for builtins for AArch64, PowerPC, and SystemZ.



================
Comment at: include/llvm/IR/IntrinsicsX86.td:4497
 
-  def int_x86_avx512_mask_sqrt_ss : GCCBuiltin<"__builtin_ia32_sqrtss_round_mask">,
+  def int_x86_avx512_sqrt_ss_mask : GCCBuiltin<"__builtin_ia32_sqrtss_mask">,
         Intrinsic<[llvm_v4f32_ty], [llvm_v4f32_ty, llvm_v4f32_ty, llvm_v4f32_ty,
----------------
Why are we renaming intrinsics here? Is this done to purposely exclude the AVX512 intrinsics? Why are we doing that?


https://reviews.llvm.org/D41599