[PATCH] D35700: DAGCombiner: Extend reduceBuildVecToTrunc to handle non-zero offset

[Guy Blank via Phabricator via llvm-commits]

guyblank added inline comments.


================
Comment at: test/CodeGen/X86/shuffle-strided-with-offset-256.ll:39
+; AVX512F-NEXT:    vpmovsxwd %ymm0, %zmm0
+; AVX512F-NEXT:    vpmovdb %zmm0, %xmm0
 ; AVX512F-NEXT:    vmovdqa %xmm0, (%rsi)
----------------
should this be 

```
vpmovdb %zmm0, (%rsi)
```

same for AVX512VL


https://reviews.llvm.org/D35700