[PATCH] D12665: [X86][AVX512] extend support in Scalar conversion

Asaf Badouh via llvm-commits llvm-commits at lists.llvm.org
Tue Sep 8 07:27:49 PDT 2015


AsafBadouh added a comment.

see the upcoming patch


================
Comment at: ../llvm/lib/Target/X86/X86ISelLowering.cpp:19331
@@ -19330,2 +19330,3 @@
   case X86ISD::VFPROUND:           return "X86ISD::VFPROUND";
+  case X86ISD::SFROUND:            return "X86ISD::SFROUND";
   case X86ISD::CVTDQ2PD:           return "X86ISD::CVTDQ2PD";
----------------
delena wrote:
> Please try to avoid adding new nodes. May be you can use the existing staff.
removed, used VFPROUND instead

================
Comment at: ../llvm/lib/Target/X86/X86InstrAVX512.td:4494
@@ +4493,3 @@
+                            X86VectorVTInfo _DstRC, SDNode OpNode, 
+                            SDNode OpNodeRnd, Intrinsic Int>{
+let Predicates = [HasAVX512] in {
----------------
delena wrote:
> You don't use the "Int" parameter here.
fixed and removed


http://reviews.llvm.org/D12665





More information about the llvm-commits mailing list