[PATCH] D19310: X86 TRUNCATE (v16i32 to v16i8) cost change in SSE4.1 mode

Simon Pilgrim via llvm-commits llvm-commits at lists.llvm.org
Wed Apr 20 07:17:56 PDT 2016


RKSimon added inline comments.

================
Comment at: lib/Target/X86/X86TargetTransformInfo.cpp:734
@@ -733,3 +733,3 @@
     { ISD::TRUNCATE,    MVT::v4i16,  MVT::v4i32,  1 },
-    { ISD::TRUNCATE,    MVT::v16i8,  MVT::v16i32, 30 },
+    { ISD::TRUNCATE,    MVT::v16i8,  MVT::v16i32, 7 },
     { ISD::TRUNCATE,    MVT::v8i8,   MVT::v8i32,  3 },
----------------
To stop this happening in the future it may be better to just remove the entry from the SSE41 table to let it 'fall through' to the SSE2 entry. I think the MVT::v16i8/MVT::v16i16 entry can go as well.


Repository:
  rL LLVM

http://reviews.llvm.org/D19310





More information about the llvm-commits mailing list