[X86] Add abds/abdu lowering for scalar i8/i16/i32/i64 types

The next step will be to begin adding generic legalization/lowering support

diff --git a/llvm/lib/Target/X86/X86ISelLowering.cpp b/llvm/lib/Target/X86/X86ISelLowering.cpp
index 0b1b976..5f9aec6 100644 (file)
--- a/llvm/lib/Target/X86/X86ISelLowering.cpp
+++ b/llvm/lib/Target/X86/X86ISelLowering.cpp
@@ -229,6 +229,15 @@ X86TargetLowering::X86TargetLowering(const X86TargetMachine &TM,
       setOperationAction(ISD::ABS          , MVT::i64  , Custom);
   }
 
+  // Absolute difference.
+  for (auto Op : {ISD::ABDS, ISD::ABDU}) {
+    setOperationAction(Op                  , MVT::i8   , Custom);
+    setOperationAction(Op                  , MVT::i16  , Custom);
+    setOperationAction(Op                  , MVT::i32  , Custom);
+    if (Subtarget.is64Bit())
+     setOperationAction(Op                 , MVT::i64  , Custom);
+  }
+
   // Signed saturation subtraction.
   setOperationAction(ISD::SSUBSAT          , MVT::i8   , Custom);
   setOperationAction(ISD::SSUBSAT          , MVT::i16  , Custom);
@@ -30389,6 +30398,21 @@ static SDValue LowerABD(SDValue Op, const X86Subtarget &Subtarget,
                        DAG.getNode(ISD::USUBSAT, dl, VT, LHS, RHS),
                        DAG.getNode(ISD::USUBSAT, dl, VT, RHS, LHS));
 
+  if (VT.isScalarInteger()) {
+    unsigned WideBits = std::max<unsigned>(2 * VT.getScalarSizeInBits(), 32u);
+    MVT WideVT = MVT::getIntegerVT(WideBits);
+    if (TLI.isTypeLegal(WideVT)) {
+      // abds(lhs, rhs) -> trunc(abs(sub(sext(lhs), sext(rhs))))
+      // abdu(lhs, rhs) -> trunc(abs(sub(zext(lhs), zext(rhs))))
+      unsigned ExtOpc = IsSigned ? ISD::SIGN_EXTEND : ISD::ZERO_EXTEND;
+      LHS = DAG.getNode(ExtOpc, dl, WideVT, LHS);
+      RHS = DAG.getNode(ExtOpc, dl, WideVT, RHS);
+      SDValue Diff = DAG.getNode(ISD::SUB, dl, WideVT, LHS, RHS);
+      SDValue AbsDiff = DAG.getNode(ISD::ABS, dl, WideVT, Diff);
+      return DAG.getNode(ISD::TRUNCATE, dl, VT, AbsDiff);
+    }
+  }
+
   // abds(lhs, rhs) -> select(sgt(lhs,rhs), sub(lhs,rhs), sub(rhs,lhs))
   // abdu(lhs, rhs) -> select(ugt(lhs,rhs), sub(lhs,rhs), sub(rhs,lhs))
   EVT CCVT = TLI.getSetCCResultType(DAG.getDataLayout(), *DAG.getContext(), VT);

diff --git a/llvm/test/CodeGen/X86/abds.ll b/llvm/test/CodeGen/X86/abds.ll
index d5d5c8d..c527368 100644 (file)
--- a/llvm/test/CodeGen/X86/abds.ll
+++ b/llvm/test/CodeGen/X86/abds.ll
@@ -20,15 +20,13 @@ define i8 @abd_ext_i8(i8 %a, i8 %b) nounwind {
 ;
 ; X64-LABEL: abd_ext_i8:
 ; X64:       # %bb.0:
-; X64-NEXT:    # kill: def $esi killed $esi def $rsi
-; X64-NEXT:    # kill: def $edi killed $edi def $rdi
-; X64-NEXT:    movsbq %dil, %rcx
-; X64-NEXT:    movsbq %sil, %rax
-; X64-NEXT:    subq %rax, %rcx
-; X64-NEXT:    movq %rcx, %rax
-; X64-NEXT:    negq %rax
-; X64-NEXT:    cmovsq %rcx, %rax
-; X64-NEXT:    # kill: def $al killed $al killed $rax
+; X64-NEXT:    movsbl %sil, %eax
+; X64-NEXT:    movsbl %dil, %ecx
+; X64-NEXT:    subl %eax, %ecx
+; X64-NEXT:    movl %ecx, %eax
+; X64-NEXT:    negl %eax
+; X64-NEXT:    cmovsl %ecx, %eax
+; X64-NEXT:    # kill: def $al killed $al killed $eax
 ; X64-NEXT:    retq
   %aext = sext i8 %a to i64
   %bext = sext i8 %b to i64
@@ -52,15 +50,13 @@ define i8 @abd_ext_i8_undef(i8 %a, i8 %b) nounwind {
 ;
 ; X64-LABEL: abd_ext_i8_undef:
 ; X64:       # %bb.0:
-; X64-NEXT:    # kill: def $esi killed $esi def $rsi
-; X64-NEXT:    # kill: def $edi killed $edi def $rdi
-; X64-NEXT:    movsbq %dil, %rcx
-; X64-NEXT:    movsbq %sil, %rax
-; X64-NEXT:    subq %rax, %rcx
-; X64-NEXT:    movq %rcx, %rax
-; X64-NEXT:    negq %rax
-; X64-NEXT:    cmovsq %rcx, %rax
-; X64-NEXT:    # kill: def $al killed $al killed $rax
+; X64-NEXT:    movsbl %sil, %eax
+; X64-NEXT:    movsbl %dil, %ecx
+; X64-NEXT:    subl %eax, %ecx
+; X64-NEXT:    movl %ecx, %eax
+; X64-NEXT:    negl %eax
+; X64-NEXT:    cmovsl %ecx, %eax
+; X64-NEXT:    # kill: def $al killed $al killed $eax
 ; X64-NEXT:    retq
   %aext = sext i8 %a to i64
   %bext = sext i8 %b to i64
@@ -84,15 +80,13 @@ define i16 @abd_ext_i16(i16 %a, i16 %b) nounwind {
 ;
 ; X64-LABEL: abd_ext_i16:
 ; X64:       # %bb.0:
-; X64-NEXT:    # kill: def $esi killed $esi def $rsi
-; X64-NEXT:    # kill: def $edi killed $edi def $rdi
-; X64-NEXT:    movswq %di, %rcx
-; X64-NEXT:    movswq %si, %rax
-; X64-NEXT:    subq %rax, %rcx
-; X64-NEXT:    movq %rcx, %rax
-; X64-NEXT:    negq %rax
-; X64-NEXT:    cmovsq %rcx, %rax
-; X64-NEXT:    # kill: def $ax killed $ax killed $rax
+; X64-NEXT:    movswl %si, %eax
+; X64-NEXT:    movswl %di, %ecx
+; X64-NEXT:    subl %eax, %ecx
+; X64-NEXT:    movl %ecx, %eax
+; X64-NEXT:    negl %eax
+; X64-NEXT:    cmovsl %ecx, %eax
+; X64-NEXT:    # kill: def $ax killed $ax killed $eax
 ; X64-NEXT:    retq
   %aext = sext i16 %a to i64
   %bext = sext i16 %b to i64
@@ -116,15 +110,13 @@ define i16 @abd_ext_i16_undef(i16 %a, i16 %b) nounwind {
 ;
 ; X64-LABEL: abd_ext_i16_undef:
 ; X64:       # %bb.0:
-; X64-NEXT:    # kill: def $esi killed $esi def $rsi
-; X64-NEXT:    # kill: def $edi killed $edi def $rdi
-; X64-NEXT:    movswq %di, %rcx
-; X64-NEXT:    movswq %si, %rax
-; X64-NEXT:    subq %rax, %rcx
-; X64-NEXT:    movq %rcx, %rax
-; X64-NEXT:    negq %rax
-; X64-NEXT:    cmovsq %rcx, %rax
-; X64-NEXT:    # kill: def $ax killed $ax killed $rax
+; X64-NEXT:    movswl %si, %eax
+; X64-NEXT:    movswl %di, %ecx
+; X64-NEXT:    subl %eax, %ecx
+; X64-NEXT:    movl %ecx, %eax
+; X64-NEXT:    negl %eax
+; X64-NEXT:    cmovsl %ecx, %eax
+; X64-NEXT:    # kill: def $ax killed $ax killed $eax
 ; X64-NEXT:    retq
   %aext = sext i16 %a to i64
   %bext = sext i16 %b to i64
@@ -137,25 +129,19 @@ define i16 @abd_ext_i16_undef(i16 %a, i16 %b) nounwind {
 define i32 @abd_ext_i32(i32 %a, i32 %b) nounwind {
 ; X86-LABEL: abd_ext_i32:
 ; X86:       # %bb.0:
-; X86-NEXT:    pushl %esi
 ; X86-NEXT:    movl {{[0-9]+}}(%esp), %ecx
-; X86-NEXT:    movl %ecx, %edx
-; X86-NEXT:    sarl $31, %edx
 ; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax
-; X86-NEXT:    movl %eax, %esi
-; X86-NEXT:    sarl $31, %esi
+; X86-NEXT:    movl %eax, %edx
+; X86-NEXT:    subl %ecx, %edx
+; X86-NEXT:    negl %edx
 ; X86-NEXT:    subl %ecx, %eax
-; X86-NEXT:    sbbl %edx, %esi
-; X86-NEXT:    sarl $31, %esi
-; X86-NEXT:    xorl %esi, %eax
-; X86-NEXT:    subl %esi, %eax
-; X86-NEXT:    popl %esi
+; X86-NEXT:    cmovlel %edx, %eax
 ; X86-NEXT:    retl
 ;
 ; X64-LABEL: abd_ext_i32:
 ; X64:       # %bb.0:
-; X64-NEXT:    movslq %edi, %rcx
 ; X64-NEXT:    movslq %esi, %rax
+; X64-NEXT:    movslq %edi, %rcx
 ; X64-NEXT:    subq %rax, %rcx
 ; X64-NEXT:    movq %rcx, %rax
 ; X64-NEXT:    negq %rax
@@ -173,25 +159,19 @@ define i32 @abd_ext_i32(i32 %a, i32 %b) nounwind {
 define i32 @abd_ext_i32_undef(i32 %a, i32 %b) nounwind {
 ; X86-LABEL: abd_ext_i32_undef:
 ; X86:       # %bb.0:
-; X86-NEXT:    pushl %esi
 ; X86-NEXT:    movl {{[0-9]+}}(%esp), %ecx
-; X86-NEXT:    movl %ecx, %edx
-; X86-NEXT:    sarl $31, %edx
 ; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax
-; X86-NEXT:    movl %eax, %esi
-; X86-NEXT:    sarl $31, %esi
+; X86-NEXT:    movl %eax, %edx
+; X86-NEXT:    subl %ecx, %edx
+; X86-NEXT:    negl %edx
 ; X86-NEXT:    subl %ecx, %eax
-; X86-NEXT:    sbbl %edx, %esi
-; X86-NEXT:    sarl $31, %esi
-; X86-NEXT:    xorl %esi, %eax
-; X86-NEXT:    subl %esi, %eax
-; X86-NEXT:    popl %esi
+; X86-NEXT:    cmovlel %edx, %eax
 ; X86-NEXT:    retl
 ;
 ; X64-LABEL: abd_ext_i32_undef:
 ; X64:       # %bb.0:
-; X64-NEXT:    movslq %edi, %rcx
 ; X64-NEXT:    movslq %esi, %rax
+; X64-NEXT:    movslq %edi, %rcx
 ; X64-NEXT:    subq %rax, %rcx
 ; X64-NEXT:    movq %rcx, %rax
 ; X64-NEXT:    negq %rax
@@ -235,15 +215,10 @@ define i64 @abd_ext_i64(i64 %a, i64 %b) nounwind {
 ; X64-LABEL: abd_ext_i64:
 ; X64:       # %bb.0:
 ; X64-NEXT:    movq %rdi, %rax
-; X64-NEXT:    movq %rdi, %rcx
-; X64-NEXT:    sarq $63, %rcx
-; X64-NEXT:    movq %rsi, %rdx
-; X64-NEXT:    sarq $63, %rdx
 ; X64-NEXT:    subq %rsi, %rax
-; X64-NEXT:    sbbq %rdx, %rcx
-; X64-NEXT:    sarq $63, %rcx
-; X64-NEXT:    xorq %rcx, %rax
-; X64-NEXT:    subq %rcx, %rax
+; X64-NEXT:    negq %rax
+; X64-NEXT:    subq %rsi, %rdi
+; X64-NEXT:    cmovgq %rdi, %rax
 ; X64-NEXT:    retq
   %aext = sext i64 %a to i128
   %bext = sext i64 %b to i128
@@ -282,15 +257,10 @@ define i64 @abd_ext_i64_undef(i64 %a, i64 %b) nounwind {
 ; X64-LABEL: abd_ext_i64_undef:
 ; X64:       # %bb.0:
 ; X64-NEXT:    movq %rdi, %rax
-; X64-NEXT:    movq %rdi, %rcx
-; X64-NEXT:    sarq $63, %rcx
-; X64-NEXT:    movq %rsi, %rdx
-; X64-NEXT:    sarq $63, %rdx
 ; X64-NEXT:    subq %rsi, %rax
-; X64-NEXT:    sbbq %rdx, %rcx
-; X64-NEXT:    sarq $63, %rcx
-; X64-NEXT:    xorq %rcx, %rax
-; X64-NEXT:    subq %rcx, %rax
+; X64-NEXT:    negq %rax
+; X64-NEXT:    subq %rsi, %rdi
+; X64-NEXT:    cmovgq %rdi, %rax
 ; X64-NEXT:    retq
   %aext = sext i64 %a to i128
   %bext = sext i64 %b to i128
@@ -307,24 +277,23 @@ define i64 @abd_ext_i64_undef(i64 %a, i64 %b) nounwind {
 define i8 @abd_minmax_i8(i8 %a, i8 %b) nounwind {
 ; X86-LABEL: abd_minmax_i8:
 ; X86:       # %bb.0:
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %ecx
-; X86-NEXT:    cmpb %al, %cl
-; X86-NEXT:    movl %eax, %edx
-; X86-NEXT:    cmovll %ecx, %edx
-; X86-NEXT:    cmovgl %ecx, %eax
-; X86-NEXT:    subb %dl, %al
+; X86-NEXT:    movsbl {{[0-9]+}}(%esp), %eax
+; X86-NEXT:    movsbl {{[0-9]+}}(%esp), %ecx
+; X86-NEXT:    subl %eax, %ecx
+; X86-NEXT:    movl %ecx, %eax
+; X86-NEXT:    negl %eax
+; X86-NEXT:    cmovsl %ecx, %eax
 ; X86-NEXT:    # kill: def $al killed $al killed $eax
 ; X86-NEXT:    retl
 ;
 ; X64-LABEL: abd_minmax_i8:
 ; X64:       # %bb.0:
-; X64-NEXT:    movl %esi, %eax
-; X64-NEXT:    cmpb %al, %dil
-; X64-NEXT:    movl %esi, %ecx
-; X64-NEXT:    cmovll %edi, %ecx
-; X64-NEXT:    cmovgl %edi, %eax
-; X64-NEXT:    subb %cl, %al
+; X64-NEXT:    movsbl %sil, %eax
+; X64-NEXT:    movsbl %dil, %ecx
+; X64-NEXT:    subl %eax, %ecx
+; X64-NEXT:    movl %ecx, %eax
+; X64-NEXT:    negl %eax
+; X64-NEXT:    cmovsl %ecx, %eax
 ; X64-NEXT:    # kill: def $al killed $al killed $eax
 ; X64-NEXT:    retq
   %min = call i8 @llvm.smin.i8(i8 %a, i8 %b)
@@ -336,24 +305,23 @@ define i8 @abd_minmax_i8(i8 %a, i8 %b) nounwind {
 define i16 @abd_minmax_i16(i16 %a, i16 %b) nounwind {
 ; X86-LABEL: abd_minmax_i16:
 ; X86:       # %bb.0:
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %ecx
-; X86-NEXT:    cmpw %ax, %cx
-; X86-NEXT:    movl %eax, %edx
-; X86-NEXT:    cmovll %ecx, %edx
-; X86-NEXT:    cmovgl %ecx, %eax
-; X86-NEXT:    subl %edx, %eax
+; X86-NEXT:    movswl {{[0-9]+}}(%esp), %eax
+; X86-NEXT:    movswl {{[0-9]+}}(%esp), %ecx
+; X86-NEXT:    subl %eax, %ecx
+; X86-NEXT:    movl %ecx, %eax
+; X86-NEXT:    negl %eax
+; X86-NEXT:    cmovsl %ecx, %eax
 ; X86-NEXT:    # kill: def $ax killed $ax killed $eax
 ; X86-NEXT:    retl
 ;
 ; X64-LABEL: abd_minmax_i16:
 ; X64:       # %bb.0:
-; X64-NEXT:    movl %esi, %eax
-; X64-NEXT:    cmpw %ax, %di
-; X64-NEXT:    movl %esi, %ecx
-; X64-NEXT:    cmovll %edi, %ecx
-; X64-NEXT:    cmovgl %edi, %eax
-; X64-NEXT:    subl %ecx, %eax
+; X64-NEXT:    movswl %si, %eax
+; X64-NEXT:    movswl %di, %ecx
+; X64-NEXT:    subl %eax, %ecx
+; X64-NEXT:    movl %ecx, %eax
+; X64-NEXT:    negl %eax
+; X64-NEXT:    cmovsl %ecx, %eax
 ; X64-NEXT:    # kill: def $ax killed $ax killed $eax
 ; X64-NEXT:    retq
   %min = call i16 @llvm.smin.i16(i16 %a, i16 %b)
@@ -365,23 +333,24 @@ define i16 @abd_minmax_i16(i16 %a, i16 %b) nounwind {
 define i32 @abd_minmax_i32(i32 %a, i32 %b) nounwind {
 ; X86-LABEL: abd_minmax_i32:
 ; X86:       # %bb.0:
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax
 ; X86-NEXT:    movl {{[0-9]+}}(%esp), %ecx
-; X86-NEXT:    cmpl %eax, %ecx
+; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax
 ; X86-NEXT:    movl %eax, %edx
-; X86-NEXT:    cmovll %ecx, %edx
-; X86-NEXT:    cmovgl %ecx, %eax
-; X86-NEXT:    subl %edx, %eax
+; X86-NEXT:    subl %ecx, %edx
+; X86-NEXT:    negl %edx
+; X86-NEXT:    subl %ecx, %eax
+; X86-NEXT:    cmovlel %edx, %eax
 ; X86-NEXT:    retl
 ;
 ; X64-LABEL: abd_minmax_i32:
 ; X64:       # %bb.0:
-; X64-NEXT:    movl %esi, %eax
-; X64-NEXT:    cmpl %esi, %edi
-; X64-NEXT:    movl %esi, %ecx
-; X64-NEXT:    cmovll %edi, %ecx
-; X64-NEXT:    cmovgl %edi, %eax
-; X64-NEXT:    subl %ecx, %eax
+; X64-NEXT:    movslq %esi, %rax
+; X64-NEXT:    movslq %edi, %rcx
+; X64-NEXT:    subq %rax, %rcx
+; X64-NEXT:    movq %rcx, %rax
+; X64-NEXT:    negq %rax
+; X64-NEXT:    cmovsq %rcx, %rax
+; X64-NEXT:    # kill: def $eax killed $eax killed $rax
 ; X64-NEXT:    retq
   %min = call i32 @llvm.smin.i32(i32 %a, i32 %b)
   %max = call i32 @llvm.smax.i32(i32 %a, i32 %b)
@@ -424,12 +393,11 @@ define i64 @abd_minmax_i64(i64 %a, i64 %b) nounwind {
 ;
 ; X64-LABEL: abd_minmax_i64:
 ; X64:       # %bb.0:
-; X64-NEXT:    movq %rsi, %rax
-; X64-NEXT:    cmpq %rsi, %rdi
-; X64-NEXT:    movq %rsi, %rcx
-; X64-NEXT:    cmovlq %rdi, %rcx
+; X64-NEXT:    movq %rdi, %rax
+; X64-NEXT:    subq %rsi, %rax
+; X64-NEXT:    negq %rax
+; X64-NEXT:    subq %rsi, %rdi
 ; X64-NEXT:    cmovgq %rdi, %rax
-; X64-NEXT:    subq %rcx, %rax
 ; X64-NEXT:    retq
   %min = call i64 @llvm.smin.i64(i64 %a, i64 %b)
   %max = call i64 @llvm.smax.i64(i64 %a, i64 %b)

diff --git a/llvm/test/CodeGen/X86/abdu.ll b/llvm/test/CodeGen/X86/abdu.ll
index 7095f77..f98daa1 100644 (file)
--- a/llvm/test/CodeGen/X86/abdu.ll
+++ b/llvm/test/CodeGen/X86/abdu.ll
@@ -20,13 +20,13 @@ define i8 @abd_ext_i8(i8 %a, i8 %b) nounwind {
 ;
 ; X64-LABEL: abd_ext_i8:
 ; X64:       # %bb.0:
-; X64-NEXT:    movzbl %dil, %ecx
 ; X64-NEXT:    movzbl %sil, %eax
-; X64-NEXT:    subq %rax, %rcx
-; X64-NEXT:    movq %rcx, %rax
-; X64-NEXT:    negq %rax
-; X64-NEXT:    cmovsq %rcx, %rax
-; X64-NEXT:    # kill: def $al killed $al killed $rax
+; X64-NEXT:    movzbl %dil, %ecx
+; X64-NEXT:    subl %eax, %ecx
+; X64-NEXT:    movl %ecx, %eax
+; X64-NEXT:    negl %eax
+; X64-NEXT:    cmovsl %ecx, %eax
+; X64-NEXT:    # kill: def $al killed $al killed $eax
 ; X64-NEXT:    retq
   %aext = zext i8 %a to i64
   %bext = zext i8 %b to i64
@@ -50,13 +50,13 @@ define i8 @abd_ext_i8_undef(i8 %a, i8 %b) nounwind {
 ;
 ; X64-LABEL: abd_ext_i8_undef:
 ; X64:       # %bb.0:
-; X64-NEXT:    movzbl %dil, %ecx
 ; X64-NEXT:    movzbl %sil, %eax
-; X64-NEXT:    subq %rax, %rcx
-; X64-NEXT:    movq %rcx, %rax
-; X64-NEXT:    negq %rax
-; X64-NEXT:    cmovsq %rcx, %rax
-; X64-NEXT:    # kill: def $al killed $al killed $rax
+; X64-NEXT:    movzbl %dil, %ecx
+; X64-NEXT:    subl %eax, %ecx
+; X64-NEXT:    movl %ecx, %eax
+; X64-NEXT:    negl %eax
+; X64-NEXT:    cmovsl %ecx, %eax
+; X64-NEXT:    # kill: def $al killed $al killed $eax
 ; X64-NEXT:    retq
   %aext = zext i8 %a to i64
   %bext = zext i8 %b to i64
@@ -80,13 +80,13 @@ define i16 @abd_ext_i16(i16 %a, i16 %b) nounwind {
 ;
 ; X64-LABEL: abd_ext_i16:
 ; X64:       # %bb.0:
-; X64-NEXT:    movzwl %di, %ecx
 ; X64-NEXT:    movzwl %si, %eax
-; X64-NEXT:    subq %rax, %rcx
-; X64-NEXT:    movq %rcx, %rax
-; X64-NEXT:    negq %rax
-; X64-NEXT:    cmovsq %rcx, %rax
-; X64-NEXT:    # kill: def $ax killed $ax killed $rax
+; X64-NEXT:    movzwl %di, %ecx
+; X64-NEXT:    subl %eax, %ecx
+; X64-NEXT:    movl %ecx, %eax
+; X64-NEXT:    negl %eax
+; X64-NEXT:    cmovsl %ecx, %eax
+; X64-NEXT:    # kill: def $ax killed $ax killed $eax
 ; X64-NEXT:    retq
   %aext = zext i16 %a to i64
   %bext = zext i16 %b to i64
@@ -110,13 +110,13 @@ define i16 @abd_ext_i16_undef(i16 %a, i16 %b) nounwind {
 ;
 ; X64-LABEL: abd_ext_i16_undef:
 ; X64:       # %bb.0:
-; X64-NEXT:    movzwl %di, %ecx
 ; X64-NEXT:    movzwl %si, %eax
-; X64-NEXT:    subq %rax, %rcx
-; X64-NEXT:    movq %rcx, %rax
-; X64-NEXT:    negq %rax
-; X64-NEXT:    cmovsq %rcx, %rax
-; X64-NEXT:    # kill: def $ax killed $ax killed $rax
+; X64-NEXT:    movzwl %di, %ecx
+; X64-NEXT:    subl %eax, %ecx
+; X64-NEXT:    movl %ecx, %eax
+; X64-NEXT:    negl %eax
+; X64-NEXT:    cmovsl %ecx, %eax
+; X64-NEXT:    # kill: def $ax killed $ax killed $eax
 ; X64-NEXT:    retq
   %aext = zext i16 %a to i64
   %bext = zext i16 %b to i64
@@ -129,19 +129,19 @@ define i16 @abd_ext_i16_undef(i16 %a, i16 %b) nounwind {
 define i32 @abd_ext_i32(i32 %a, i32 %b) nounwind {
 ; X86-LABEL: abd_ext_i32:
 ; X86:       # %bb.0:
+; X86-NEXT:    movl {{[0-9]+}}(%esp), %ecx
 ; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax
-; X86-NEXT:    xorl %ecx, %ecx
-; X86-NEXT:    subl {{[0-9]+}}(%esp), %eax
-; X86-NEXT:    sbbl %ecx, %ecx
-; X86-NEXT:    sarl $31, %ecx
-; X86-NEXT:    xorl %ecx, %eax
+; X86-NEXT:    movl %eax, %edx
+; X86-NEXT:    subl %ecx, %edx
+; X86-NEXT:    negl %edx
 ; X86-NEXT:    subl %ecx, %eax
+; X86-NEXT:    cmovbel %edx, %eax
 ; X86-NEXT:    retl
 ;
 ; X64-LABEL: abd_ext_i32:
 ; X64:       # %bb.0:
-; X64-NEXT:    movl %edi, %ecx
 ; X64-NEXT:    movl %esi, %eax
+; X64-NEXT:    movl %edi, %ecx
 ; X64-NEXT:    subq %rax, %rcx
 ; X64-NEXT:    movq %rcx, %rax
 ; X64-NEXT:    negq %rax
@@ -159,19 +159,19 @@ define i32 @abd_ext_i32(i32 %a, i32 %b) nounwind {
 define i32 @abd_ext_i32_undef(i32 %a, i32 %b) nounwind {
 ; X86-LABEL: abd_ext_i32_undef:
 ; X86:       # %bb.0:
+; X86-NEXT:    movl {{[0-9]+}}(%esp), %ecx
 ; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax
-; X86-NEXT:    xorl %ecx, %ecx
-; X86-NEXT:    subl {{[0-9]+}}(%esp), %eax
-; X86-NEXT:    sbbl %ecx, %ecx
-; X86-NEXT:    sarl $31, %ecx
-; X86-NEXT:    xorl %ecx, %eax
+; X86-NEXT:    movl %eax, %edx
+; X86-NEXT:    subl %ecx, %edx
+; X86-NEXT:    negl %edx
 ; X86-NEXT:    subl %ecx, %eax
+; X86-NEXT:    cmovbel %edx, %eax
 ; X86-NEXT:    retl
 ;
 ; X64-LABEL: abd_ext_i32_undef:
 ; X64:       # %bb.0:
-; X64-NEXT:    movl %edi, %ecx
 ; X64-NEXT:    movl %esi, %eax
+; X64-NEXT:    movl %edi, %ecx
 ; X64-NEXT:    subq %rax, %rcx
 ; X64-NEXT:    movq %rcx, %rax
 ; X64-NEXT:    negq %rax
@@ -209,12 +209,10 @@ define i64 @abd_ext_i64(i64 %a, i64 %b) nounwind {
 ; X64-LABEL: abd_ext_i64:
 ; X64:       # %bb.0:
 ; X64-NEXT:    movq %rdi, %rax
-; X64-NEXT:    xorl %ecx, %ecx
 ; X64-NEXT:    subq %rsi, %rax
-; X64-NEXT:    sbbq %rcx, %rcx
-; X64-NEXT:    sarq $63, %rcx
-; X64-NEXT:    xorq %rcx, %rax
-; X64-NEXT:    subq %rcx, %rax
+; X64-NEXT:    negq %rax
+; X64-NEXT:    subq %rsi, %rdi
+; X64-NEXT:    cmovaq %rdi, %rax
 ; X64-NEXT:    retq
   %aext = zext i64 %a to i128
   %bext = zext i64 %b to i128
@@ -247,12 +245,10 @@ define i64 @abd_ext_i64_undef(i64 %a, i64 %b) nounwind {
 ; X64-LABEL: abd_ext_i64_undef:
 ; X64:       # %bb.0:
 ; X64-NEXT:    movq %rdi, %rax
-; X64-NEXT:    xorl %ecx, %ecx
 ; X64-NEXT:    subq %rsi, %rax
-; X64-NEXT:    sbbq %rcx, %rcx
-; X64-NEXT:    sarq $63, %rcx
-; X64-NEXT:    xorq %rcx, %rax
-; X64-NEXT:    subq %rcx, %rax
+; X64-NEXT:    negq %rax
+; X64-NEXT:    subq %rsi, %rdi
+; X64-NEXT:    cmovaq %rdi, %rax
 ; X64-NEXT:    retq
   %aext = zext i64 %a to i128
   %bext = zext i64 %b to i128
@@ -269,24 +265,23 @@ define i64 @abd_ext_i64_undef(i64 %a, i64 %b) nounwind {
 define i8 @abd_minmax_i8(i8 %a, i8 %b) nounwind {
 ; X86-LABEL: abd_minmax_i8:
 ; X86:       # %bb.0:
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %ecx
-; X86-NEXT:    cmpb %al, %cl
-; X86-NEXT:    movl %eax, %edx
-; X86-NEXT:    cmovbl %ecx, %edx
-; X86-NEXT:    cmoval %ecx, %eax
-; X86-NEXT:    subb %dl, %al
+; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %eax
+; X86-NEXT:    movzbl {{[0-9]+}}(%esp), %ecx
+; X86-NEXT:    subl %eax, %ecx
+; X86-NEXT:    movl %ecx, %eax
+; X86-NEXT:    negl %eax
+; X86-NEXT:    cmovsl %ecx, %eax
 ; X86-NEXT:    # kill: def $al killed $al killed $eax
 ; X86-NEXT:    retl
 ;
 ; X64-LABEL: abd_minmax_i8:
 ; X64:       # %bb.0:
-; X64-NEXT:    movl %esi, %eax
-; X64-NEXT:    cmpb %al, %dil
-; X64-NEXT:    movl %esi, %ecx
-; X64-NEXT:    cmovbl %edi, %ecx
-; X64-NEXT:    cmoval %edi, %eax
-; X64-NEXT:    subb %cl, %al
+; X64-NEXT:    movzbl %sil, %eax
+; X64-NEXT:    movzbl %dil, %ecx
+; X64-NEXT:    subl %eax, %ecx
+; X64-NEXT:    movl %ecx, %eax
+; X64-NEXT:    negl %eax
+; X64-NEXT:    cmovsl %ecx, %eax
 ; X64-NEXT:    # kill: def $al killed $al killed $eax
 ; X64-NEXT:    retq
   %min = call i8 @llvm.umin.i8(i8 %a, i8 %b)
@@ -298,24 +293,23 @@ define i8 @abd_minmax_i8(i8 %a, i8 %b) nounwind {
 define i16 @abd_minmax_i16(i16 %a, i16 %b) nounwind {
 ; X86-LABEL: abd_minmax_i16:
 ; X86:       # %bb.0:
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %ecx
-; X86-NEXT:    cmpw %ax, %cx
-; X86-NEXT:    movl %eax, %edx
-; X86-NEXT:    cmovbl %ecx, %edx
-; X86-NEXT:    cmoval %ecx, %eax
-; X86-NEXT:    subl %edx, %eax
+; X86-NEXT:    movzwl {{[0-9]+}}(%esp), %eax
+; X86-NEXT:    movzwl {{[0-9]+}}(%esp), %ecx
+; X86-NEXT:    subl %eax, %ecx
+; X86-NEXT:    movl %ecx, %eax
+; X86-NEXT:    negl %eax
+; X86-NEXT:    cmovsl %ecx, %eax
 ; X86-NEXT:    # kill: def $ax killed $ax killed $eax
 ; X86-NEXT:    retl
 ;
 ; X64-LABEL: abd_minmax_i16:
 ; X64:       # %bb.0:
-; X64-NEXT:    movl %esi, %eax
-; X64-NEXT:    cmpw %ax, %di
-; X64-NEXT:    movl %esi, %ecx
-; X64-NEXT:    cmovbl %edi, %ecx
-; X64-NEXT:    cmoval %edi, %eax
-; X64-NEXT:    subl %ecx, %eax
+; X64-NEXT:    movzwl %si, %eax
+; X64-NEXT:    movzwl %di, %ecx
+; X64-NEXT:    subl %eax, %ecx
+; X64-NEXT:    movl %ecx, %eax
+; X64-NEXT:    negl %eax
+; X64-NEXT:    cmovsl %ecx, %eax
 ; X64-NEXT:    # kill: def $ax killed $ax killed $eax
 ; X64-NEXT:    retq
   %min = call i16 @llvm.umin.i16(i16 %a, i16 %b)
@@ -327,23 +321,24 @@ define i16 @abd_minmax_i16(i16 %a, i16 %b) nounwind {
 define i32 @abd_minmax_i32(i32 %a, i32 %b) nounwind {
 ; X86-LABEL: abd_minmax_i32:
 ; X86:       # %bb.0:
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax
 ; X86-NEXT:    movl {{[0-9]+}}(%esp), %ecx
-; X86-NEXT:    cmpl %eax, %ecx
+; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax
 ; X86-NEXT:    movl %eax, %edx
-; X86-NEXT:    cmovbl %ecx, %edx
-; X86-NEXT:    cmoval %ecx, %eax
-; X86-NEXT:    subl %edx, %eax
+; X86-NEXT:    subl %ecx, %edx
+; X86-NEXT:    negl %edx
+; X86-NEXT:    subl %ecx, %eax
+; X86-NEXT:    cmovbel %edx, %eax
 ; X86-NEXT:    retl
 ;
 ; X64-LABEL: abd_minmax_i32:
 ; X64:       # %bb.0:
 ; X64-NEXT:    movl %esi, %eax
-; X64-NEXT:    cmpl %esi, %edi
-; X64-NEXT:    movl %esi, %ecx
-; X64-NEXT:    cmovbl %edi, %ecx
-; X64-NEXT:    cmoval %edi, %eax
-; X64-NEXT:    subl %ecx, %eax
+; X64-NEXT:    movl %edi, %ecx
+; X64-NEXT:    subq %rax, %rcx
+; X64-NEXT:    movq %rcx, %rax
+; X64-NEXT:    negq %rax
+; X64-NEXT:    cmovsq %rcx, %rax
+; X64-NEXT:    # kill: def $eax killed $eax killed $rax
 ; X64-NEXT:    retq
   %min = call i32 @llvm.umin.i32(i32 %a, i32 %b)
   %max = call i32 @llvm.umax.i32(i32 %a, i32 %b)
@@ -386,12 +381,11 @@ define i64 @abd_minmax_i64(i64 %a, i64 %b) nounwind {
 ;
 ; X64-LABEL: abd_minmax_i64:
 ; X64:       # %bb.0:
-; X64-NEXT:    movq %rsi, %rax
-; X64-NEXT:    cmpq %rsi, %rdi
-; X64-NEXT:    movq %rsi, %rcx
-; X64-NEXT:    cmovbq %rdi, %rcx
+; X64-NEXT:    movq %rdi, %rax
+; X64-NEXT:    subq %rsi, %rax
+; X64-NEXT:    negq %rax
+; X64-NEXT:    subq %rsi, %rdi
 ; X64-NEXT:    cmovaq %rdi, %rax
-; X64-NEXT:    subq %rcx, %rax
 ; X64-NEXT:    retq
   %min = call i64 @llvm.umin.i64(i64 %a, i64 %b)
   %max = call i64 @llvm.umax.i64(i64 %a, i64 %b)