[Feature][CPU Backend]: Optimize ARM vectorization backend (#30329)

Signed-off-by: Radu Salavat <radu.salavat@arm.com>

csrc/cpu/mla_decode.cpp

@@ -38,9 +38,7 @@ struct KernelVecType<c10::BFloat16> {
   using qk_vec_type = vec_op::BF16Vec32;
   using v_load_vec_type = vec_op::BF16Vec16;
 };
-#elif defined(__aarch64__) && !defined(ARM_BF16_SUPPORT)
-// pass
-#else
+#elif defined(__aarch64__)
 template <>
 struct KernelVecType<c10::BFloat16> {
   using qk_load_vec_type = vec_op::BF16Vec16;