[PATCH] D68717: [Codegen] More add_sat and sub_sat promotion

[Dave Green via Phabricator via llvm-commits]

dmgreen marked an inline comment as done.
dmgreen added inline comments.


================
Comment at: llvm/test/CodeGen/X86/sadd_sat_vec.ll:496
+; SSE2-NEXT:    psubb %xmm1, %xmm0
+; SSE2-NEXT:    retq
+;
----------------
If we always went though the min/max pair, this would look like:
```
+; SSE2-NEXT:    paddb %xmm1, %xmm0
+; SSE2-NEXT:    movdqa {{.*#+}} xmm1 = [7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7]
+; SSE2-NEXT:    movdqa %xmm1, %xmm2
+; SSE2-NEXT:    pcmpgtb %xmm0, %xmm2
+; SSE2-NEXT:    pand %xmm2, %xmm0
+; SSE2-NEXT:    pandn %xmm1, %xmm2
+; SSE2-NEXT:    por %xmm0, %xmm2
+; SSE2-NEXT:    movdqa {{.*#+}} xmm1 = [248,248,248,248,248,248,248,248,248,248,248,248,248,
+; SSE2-NEXT:    movdqa %xmm2, %xmm0
+; SSE2-NEXT:    pcmpgtb %xmm1, %xmm0
+; SSE2-NEXT:    pand %xmm0, %xmm2
+; SSE2-NEXT:    pandn %xmm1, %xmm0
+; SSE2-NEXT:    por %xmm2, %xmm0
+; SSE2-NEXT:    retq
```
Which is a little larger.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D68717/new/

https://reviews.llvm.org/D68717