[X86][Btver2] Most RMW instructions don't require an additional uop

Remove uop on WriteRMW and move it into the few instructions that need it.

Match AMD Fam16h SOG + llvm-exegesis tests

llvm-svn: 343671

1 files changed, 5 insertions, 4 deletions

diff --git a/llvm/lib/Target/X86/X86ScheduleBtVer2.td b/llvm/lib/Target/X86/X86ScheduleBtVer2.td
index 3d73e38efb6..b807fa58b55 100644
--- a/llvm/lib/Target/X86/X86ScheduleBtVer2.td
+++ b/llvm/lib/Target/X86/X86ScheduleBtVer2.td
@@ -158,8 +158,9 @@ multiclass JWriteResYMMPair<X86FoldableSchedWrite SchedRW,
   }
 }
 
-// A folded store needs a cycle on the SAGU for the store data.
-def : WriteRes<WriteRMW, [JSAGU]>;
+// A folded store needs a cycle on the SAGU for the store data,
+// most RMW instructions don't need an extra uop.
+defm : X86WriteRes<WriteRMW, [JSAGU], 1, [1], 0>;
 
 ////////////////////////////////////////////////////////////////////////////////
 // Arithmetic.
@@ -208,8 +209,8 @@ defm : X86WriteRes<WriteBitTest,         [JALU01], 1, [1], 1>;
 defm : X86WriteRes<WriteBitTestImmLd,    [JALU01,JLAGU], 4, [1,1], 1>;
 defm : X86WriteRes<WriteBitTestRegLd,    [JALU01,JLAGU], 4, [1,1], 5>;
 defm : X86WriteRes<WriteBitTestSet,      [JALU01], 1, [1], 2>;
-defm : X86WriteRes<WriteBitTestSetImmLd, [JALU01,JLAGU], 4, [1,1], 3>;
-defm : X86WriteRes<WriteBitTestSetRegLd, [JALU01,JLAGU], 4, [1,1], 7>;
+defm : X86WriteRes<WriteBitTestSetImmLd, [JALU01,JLAGU], 4, [1,1], 4>;
+defm : X86WriteRes<WriteBitTestSetRegLd, [JALU01,JLAGU], 4, [1,1], 8>;
 
 // This is for simple LEAs with one or two input operands.
 def : WriteRes<WriteLEA, [JALU01]>;