[clang] [CIR][AArch64] Lower NEON vbsl builtins (PR #188449)

Andrzej WarzyƄski via cfe-commits cfe-commits at lists.llvm.org
Wed Mar 25 23:57:29 PDT 2026


================
@@ -982,3 +982,158 @@ int64_t test_vshld_u64(int64_t a,int64_t b) {
   return (int64_t)vshld_u64(a, b);
 }
 
+// LLVM-LABEL: @test_vbsl_s8(
----------------
banach-space wrote:

Please add a block comment, e.g.:

```suggestion
//===------------------------------------------------------===//
// 2.1.8.5 Bitwise select 
// https://arm-software.github.io/acle/neon_intrinsics/advsimd.html#markdown-toc-bitwise-select
//===------------------------------------------------------===//

// LLVM-LABEL: @test_vbsl_s8(
```

Btw, have you looked at the other variants in https://arm-software.github.io/acle/neon_intrinsics/advsimd.html#bitwise-select?

https://github.com/llvm/llvm-project/pull/188449


More information about the cfe-commits mailing list