[llvm] [X86] Optimize v4i16->v4i8 truncating stores via v4i32 widening (PR #186676)

[Simon Pilgrim via llvm-commits]

================
@@ -1538,3 +1538,29 @@ define void @test_masked_store_unaligned_v8i64(<8 x i64> %data, ptr %ptr, <8 x i
   store <8 x i64> %sel, ptr %ptr_vec, align 1
   ret void
 }
+
+define void @cast_i16x4_to_u8x4(ptr %a0, ptr %a1) {
----------------
RKSimon wrote:

Would avx512-trunc.ll be a better file to add this test to?

https://github.com/llvm/llvm-project/pull/186676