[X86] For lzcnt/tzcnt intrinsics use cttz/ctlz intrinsics with zero_undef flag set to false.

Previously we used a select and the zero_undef=true intrinsic. In -O2 this pattern will get optimized to zero_undef=false. But in -O0 this optimization won't happen. This results in a compare and cmov being wrapped around a tzcnt/lzcnt instruction.

By using the zero_undef=false intrinsic directly without the select, we can improve the -O0 codegen to just an lzcnt/tzcnt instruction.

Differential Revision: https://reviews.llvm.org/D52392

llvm-svn: 343126

3 files changed, 22 insertions, 10 deletions

diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index 23c760207b8..df71dbb4b4a 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -9164,6 +9164,18 @@ Value *CodeGenFunction::EmitX86BuiltinExpr(unsigned BuiltinID,
                                       Ops[0]);
     return Builder.CreateExtractValue(Call, 0);
   }
+  case X86::BI__builtin_ia32_lzcnt_u16:
+  case X86::BI__builtin_ia32_lzcnt_u32:
+  case X86::BI__builtin_ia32_lzcnt_u64: {
+    Value *F = CGM.getIntrinsic(Intrinsic::ctlz, Ops[0]->getType());
+    return Builder.CreateCall(F, {Ops[0], Builder.getInt1(false)});
+  }
+  case X86::BI__builtin_ia32_tzcnt_u16:
+  case X86::BI__builtin_ia32_tzcnt_u32:
+  case X86::BI__builtin_ia32_tzcnt_u64: {
+    Value *F = CGM.getIntrinsic(Intrinsic::cttz, Ops[0]->getType());
+    return Builder.CreateCall(F, {Ops[0], Builder.getInt1(false)});
+  }
   case X86::BI__builtin_ia32_undef128:
   case X86::BI__builtin_ia32_undef256:
   case X86::BI__builtin_ia32_undef512:
diff --git a/clang/lib/Headers/bmiintrin.h b/clang/lib/Headers/bmiintrin.h
index d03bef442a2..56c20b78d34 100644
--- a/clang/lib/Headers/bmiintrin.h
+++ b/clang/lib/Headers/bmiintrin.h
@@ -62,7 +62,7 @@
 static __inline__ unsigned short __RELAXED_FN_ATTRS
 __tzcnt_u16(unsigned short __X)
 {
-  return __X ? __builtin_ctzs(__X) : 16;
+  return __builtin_ia32_tzcnt_u16(__X);
 }
 
 /// Performs a bitwise AND of the second operand with the one's
@@ -196,7 +196,7 @@ __blsr_u32(unsigned int __X)
 static __inline__ unsigned int __RELAXED_FN_ATTRS
 __tzcnt_u32(unsigned int __X)
 {
-  return __X ? __builtin_ctz(__X) : 32;
+  return __builtin_ia32_tzcnt_u32(__X);
 }
 
 /// Counts the number of trailing zero bits in the operand.
@@ -212,7 +212,7 @@ __tzcnt_u32(unsigned int __X)
 static __inline__ int __RELAXED_FN_ATTRS
 _mm_tzcnt_32(unsigned int __X)
 {
-  return __X ? __builtin_ctz(__X) : 32;
+  return __builtin_ia32_tzcnt_u32(__X);
 }
 
 #ifdef __x86_64__
@@ -359,7 +359,7 @@ __blsr_u64(unsigned long long __X)
 static __inline__ unsigned long long __RELAXED_FN_ATTRS
 __tzcnt_u64(unsigned long long __X)
 {
-  return __X ? __builtin_ctzll(__X) : 64;
+  return __builtin_ia32_tzcnt_u64(__X);
 }
 
 /// Counts the number of trailing zero bits in the operand.
@@ -375,7 +375,7 @@ __tzcnt_u64(unsigned long long __X)
 static __inline__ long long __RELAXED_FN_ATTRS
 _mm_tzcnt_64(unsigned long long __X)
 {
-  return __X ? __builtin_ctzll(__X) : 64;
+  return __builtin_ia32_tzcnt_u64(__X);
 }
 
 #endif /* __x86_64__ */
diff --git a/clang/lib/Headers/lzcntintrin.h b/clang/lib/Headers/lzcntintrin.h
index 558f1828f0e..f2674d2a5ab 100644
--- a/clang/lib/Headers/lzcntintrin.h
+++ b/clang/lib/Headers/lzcntintrin.h
@@ -44,7 +44,7 @@
 static __inline__ unsigned short __DEFAULT_FN_ATTRS
 __lzcnt16(unsigned short __X)
 {
-  return __X ? __builtin_clzs(__X) : 16;
+  return __builtin_ia32_lzcnt_u16(__X);
 }
 
 /// Counts the number of leading zero bits in the operand.
@@ -61,7 +61,7 @@ __lzcnt16(unsigned short __X)
 static __inline__ unsigned int __DEFAULT_FN_ATTRS
 __lzcnt32(unsigned int __X)
 {
-  return __X ? __builtin_clz(__X) : 32;
+  return __builtin_ia32_lzcnt_u32(__X);
 }
 
 /// Counts the number of leading zero bits in the operand.
@@ -78,7 +78,7 @@ __lzcnt32(unsigned int __X)
 static __inline__ unsigned int __DEFAULT_FN_ATTRS
 _lzcnt_u32(unsigned int __X)
 {
-  return __X ? __builtin_clz(__X) : 32;
+  return __builtin_ia32_lzcnt_u32(__X);
 }
 
 #ifdef __x86_64__
@@ -96,7 +96,7 @@ _lzcnt_u32(unsigned int __X)
 static __inline__ unsigned long long __DEFAULT_FN_ATTRS
 __lzcnt64(unsigned long long __X)
 {
-  return __X ? __builtin_clzll(__X) : 64;
+  return __builtin_ia32_lzcnt_u64(__X);
 }
 
 /// Counts the number of leading zero bits in the operand.
@@ -113,7 +113,7 @@ __lzcnt64(unsigned long long __X)
 static __inline__ unsigned long long __DEFAULT_FN_ATTRS
 _lzcnt_u64(unsigned long long __X)
 {
-  return __X ? __builtin_clzll(__X) : 64;
+  return __builtin_ia32_lzcnt_u64(__X);
 }
 #endif