[X86] Use ADD/SUB instead of INC/DEC for Silvermont

According to Intel Software Optimization Manual on Silvermont INC or DEC instructions require an additional uop to merge the flags. As a result, a branch instruction depending on an INC or a DEC instruction incurs a 1 cycle penalty. Differential Revision: http://reviews.llvm.org/D3990 llvm-svn: 210466
llvm · Jun 9, 2014 · 5260dba · 5260dba
1 parent b22b963
commit 5260dba
Show file tree

Hide file tree

Showing 6 changed files with 37 additions and 15 deletions.
diff --git a/llvm/lib/Target/X86/X86.td b/llvm/lib/Target/X86/X86.td
@@ -168,6 +168,8 @@ def FeatureLEAUsesAG : SubtargetFeature<"lea-uses-ag", "LEAUsesAG", "true",
                                    "LEA instruction needs inputs at AG stage">;
 def FeatureSlowLEA : SubtargetFeature<"slow-lea", "SlowLEA", "true",
                                    "LEA instruction with certain arguments is slow">;
+def FeatureSlowIncDec : SubtargetFeature<"slow-incdec", "SlowIncDec", "true",
+                                   "INC and DEC instructions are slower than ADD and SUB">;
 
 //===----------------------------------------------------------------------===//
 // X86 processors supported.
@@ -228,7 +230,7 @@ def : ProcessorModel<"slm",  SLMModel, [ProcIntelSLM,
                                FeaturePCLMUL, FeatureAES,
                                FeatureCallRegIndirect,
                                FeaturePRFCHW,
-                               FeatureSlowLEA,
+                               FeatureSlowLEA, FeatureSlowIncDec,
                                FeatureSlowBTMem, FeatureFastUAMem]>;
 // "Arrandale" along with corei3 and corei5
 def : ProcessorModel<"corei7", SandyBridgeModel,

diff --git a/llvm/lib/Target/X86/X86ISelLowering.cpp b/llvm/lib/Target/X86/X86ISelLowering.cpp
@@ -10139,14 +10139,14 @@ SDValue X86TargetLowering::EmitTest(SDValue Op, unsigned X86CC, SDLoc dl,
     if (ConstantSDNode *C =
         dyn_cast<ConstantSDNode>(ArithOp.getNode()->getOperand(1))) {
       // An add of one will be selected as an INC.
-      if (C->getAPIntValue() == 1) {
+      if (C->getAPIntValue() == 1 && !Subtarget->slowIncDec()) {
         Opcode = X86ISD::INC;
         NumOperands = 1;
         break;
       }
 
       // An add of negative one (subtract of one) will be selected as a DEC.
-      if (C->getAPIntValue().isAllOnesValue()) {
+      if (C->getAPIntValue().isAllOnesValue() && !Subtarget->slowIncDec()) {
         Opcode = X86ISD::DEC;
         NumOperands = 1;
         break;

diff --git a/llvm/lib/Target/X86/X86InstrCompiler.td b/llvm/lib/Target/X86/X86InstrCompiler.td
@@ -1696,20 +1696,34 @@ def : Pat<(mul (loadi64 addr:$src1), i64immSExt32:$src2),
           (IMUL64rmi32 addr:$src1, i64immSExt32:$src2)>;
 
 // Increment reg.
-def : Pat<(add GR8 :$src, 1), (INC8r     GR8 :$src)>;
-def : Pat<(add GR16:$src, 1), (INC16r    GR16:$src)>, Requires<[Not64BitMode]>;
-def : Pat<(add GR16:$src, 1), (INC64_16r GR16:$src)>, Requires<[In64BitMode]>;
-def : Pat<(add GR32:$src, 1), (INC32r    GR32:$src)>, Requires<[Not64BitMode]>;
-def : Pat<(add GR32:$src, 1), (INC64_32r GR32:$src)>, Requires<[In64BitMode]>;
-def : Pat<(add GR64:$src, 1), (INC64r    GR64:$src)>;
+// Do not make INC if it is slow
+def : Pat<(add GR8:$src, 1),
+          (INC8r GR8:$src)>, Requires<[NotSlowIncDec]>;
+def : Pat<(add GR16:$src, 1),
+          (INC16r GR16:$src)>, Requires<[NotSlowIncDec, Not64BitMode]>;
+def : Pat<(add GR16:$src, 1),
+          (INC64_16r GR16:$src)>, Requires<[NotSlowIncDec, In64BitMode]>;
+def : Pat<(add GR32:$src, 1),
+          (INC32r GR32:$src)>, Requires<[NotSlowIncDec, Not64BitMode]>;
+def : Pat<(add GR32:$src, 1),
+          (INC64_32r GR32:$src)>, Requires<[NotSlowIncDec, In64BitMode]>;
+def : Pat<(add GR64:$src, 1),
+          (INC64r GR64:$src)>, Requires<[NotSlowIncDec]>;
 
 // Decrement reg.
-def : Pat<(add GR8 :$src, -1), (DEC8r     GR8 :$src)>;
-def : Pat<(add GR16:$src, -1), (DEC16r    GR16:$src)>, Requires<[Not64BitMode]>;
-def : Pat<(add GR16:$src, -1), (DEC64_16r GR16:$src)>, Requires<[In64BitMode]>;
-def : Pat<(add GR32:$src, -1), (DEC32r    GR32:$src)>, Requires<[Not64BitMode]>;
-def : Pat<(add GR32:$src, -1), (DEC64_32r GR32:$src)>, Requires<[In64BitMode]>;
-def : Pat<(add GR64:$src, -1), (DEC64r    GR64:$src)>;
+// Do not make DEC if it is slow
+def : Pat<(add GR8:$src, -1),
+          (DEC8r GR8:$src)>, Requires<[NotSlowIncDec]>;
+def : Pat<(add GR16:$src, -1),
+          (DEC16r GR16:$src)>, Requires<[NotSlowIncDec, Not64BitMode]>;
+def : Pat<(add GR16:$src, -1),
+          (DEC64_16r GR16:$src)>, Requires<[NotSlowIncDec, In64BitMode]>;
+def : Pat<(add GR32:$src, -1),
+          (DEC32r GR32:$src)>, Requires<[NotSlowIncDec, Not64BitMode]>;
+def : Pat<(add GR32:$src, -1),
+          (DEC64_32r GR32:$src)>, Requires<[NotSlowIncDec, In64BitMode]>;
+def : Pat<(add GR64:$src, -1),
+          (DEC64r GR64:$src)>, Requires<[NotSlowIncDec]>;
 
 // or reg/reg.
 def : Pat<(or GR8 :$src1, GR8 :$src2), (OR8rr  GR8 :$src1, GR8 :$src2)>;

diff --git a/llvm/lib/Target/X86/X86InstrInfo.td b/llvm/lib/Target/X86/X86InstrInfo.td
@@ -795,6 +795,7 @@ def OptForSpeed  : Predicate<"!OptForSize">;
 def FastBTMem    : Predicate<"!Subtarget->isBTMemSlow()">;
 def CallImmAddr  : Predicate<"Subtarget->IsLegalToCallImmediateAddr(TM)">;
 def FavorMemIndirectCall  : Predicate<"!Subtarget->callRegIndirect()">;
+def NotSlowIncDec : Predicate<"!Subtarget->slowIncDec()">;
 
 //===----------------------------------------------------------------------===//
 // X86 Instruction Format Definitions.

diff --git a/llvm/lib/Target/X86/X86Subtarget.cpp b/llvm/lib/Target/X86/X86Subtarget.cpp
@@ -291,6 +291,7 @@ void X86Subtarget::initializeEnvironment() {
   CallRegIndirect = false;
   LEAUsesAG = false;
   SlowLEA = false;
+  SlowIncDec = false;
   stackAlignment = 4;
   // FIXME: this is a known good value for Yonah. How about others?
   MaxInlineSizeThreshold = 128;

diff --git a/llvm/lib/Target/X86/X86Subtarget.h b/llvm/lib/Target/X86/X86Subtarget.h
@@ -181,6 +181,9 @@ class X86Subtarget final : public X86GenSubtargetInfo {
   /// SlowLEA - True if the LEA instruction with certain arguments is slow
   bool SlowLEA;
 
+  /// SlowIncDec - True if INC and DEC instructions are slow when writing to flags
+  bool SlowIncDec;
+
   /// Processor has AVX-512 PreFetch Instructions
   bool HasPFI;
 
@@ -319,6 +322,7 @@ class X86Subtarget final : public X86GenSubtargetInfo {
   bool callRegIndirect() const { return CallRegIndirect; }
   bool LEAusesAG() const { return LEAUsesAG; }
   bool slowLEA() const { return SlowLEA; }
+  bool slowIncDec() const { return SlowIncDec; }
   bool hasCDI() const { return HasCDI; }
   bool hasPFI() const { return HasPFI; }
   bool hasERI() const { return HasERI; }