[llvm] [LLVM][SVE] Implement isel for bfloat fptoi and itofp operations. (PR #129713)

[David Sherwood via llvm-commits]

================
@@ -5465,6 +5465,14 @@ multiclass sve_int_dup_fpimm_pred<string asm> {
             (!cast<Instruction>(NAME # _S) $zd, $pg, fpimm32:$imm8)>;
   def : Pat<(nxv2f64 (vselect nxv2i1:$pg, (splat_vector fpimm64:$imm8), nxv2f64:$zd)),
             (!cast<Instruction>(NAME # _D) $zd, $pg, fpimm64:$imm8)>;
+
+  // Some half precision immediates alias with bfloat (e.g. f16(1.875) == bf16(1.0)).
----------------
david-arm wrote:

This comment implies that some don't, so what happens if fpimmbf16 matches a value that the fp16 variant doesn't have? Or should the comment actually be something like `All fpimmbf16 immediates alias with a FP16 immediate`?

https://github.com/llvm/llvm-project/pull/129713