[llvm] r337856 - [X86] Generalize the multiply by 30 lowering to generic multipy by power 2 minus 2.
Craig Topper via llvm-commits
llvm-commits at lists.llvm.org
Tue Jul 24 14:15:41 PDT 2018
Author: ctopper
Date: Tue Jul 24 14:15:41 2018
New Revision: 337856
URL: http://llvm.org/viewvc/llvm-project?rev=337856&view=rev
Log:
[X86] Generalize the multiply by 30 lowering to generic multipy by power 2 minus 2.
Use a left shift and 2 subtracts like we do for 30. Move this out from behind the slow lea check since it doesn't even use an LEA.
Use this for multiply by 14 as well.
Modified:
llvm/trunk/lib/Target/X86/X86ISelLowering.cpp
llvm/trunk/test/CodeGen/X86/mul-constant-i16.ll
llvm/trunk/test/CodeGen/X86/mul-constant-i32.ll
llvm/trunk/test/CodeGen/X86/mul-constant-i64.ll
llvm/trunk/test/CodeGen/X86/mul-constant-result.ll
Modified: llvm/trunk/lib/Target/X86/X86ISelLowering.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/X86/X86ISelLowering.cpp?rev=337856&r1=337855&r2=337856&view=diff
==============================================================================
--- llvm/trunk/lib/Target/X86/X86ISelLowering.cpp (original)
+++ llvm/trunk/lib/Target/X86/X86ISelLowering.cpp Tue Jul 24 14:15:41 2018
@@ -33727,12 +33727,8 @@ static SDValue combineMulSpecial(uint64_
// mul x, 13 => add ((shl (mul x, 3), 2), x)
return combineMulShlAddOrSub(3, 2, /*isAdd*/ true);
case 23:
- // mul x, 13 => sub ((shl (mul x, 3), 3), x)
+ // mul x, 23 => sub ((shl (mul x, 3), 3), x)
return combineMulShlAddOrSub(3, 3, /*isAdd*/ false);
- case 14:
- // mul x, 14 => add (add ((shl (mul x, 3), 2), x), x)
- return DAG.getNode(ISD::ADD, DL, VT, N->getOperand(0),
- combineMulShlAddOrSub(3, 2, /*isAdd*/ true));
case 26:
// mul x, 26 => sub ((mul (mul x, 9), 3), x)
return combineMulMulAddOrSub(/*isAdd*/ false);
@@ -33743,15 +33739,6 @@ static SDValue combineMulSpecial(uint64_
// mul x, 29 => add (add ((mul (mul x, 9), 3), x), x)
return DAG.getNode(ISD::ADD, DL, VT, N->getOperand(0),
combineMulMulAddOrSub(/*isAdd*/ true));
- case 30:
- // mul x, 30 => sub (sub ((shl x, 5), x), x)
- return DAG.getNode(
- ISD::SUB, DL, VT,
- DAG.getNode(ISD::SUB, DL, VT,
- DAG.getNode(ISD::SHL, DL, VT, N->getOperand(0),
- DAG.getConstant(5, DL, MVT::i8)),
- N->getOperand(0)),
- N->getOperand(0));
}
return SDValue();
}
@@ -33927,6 +33914,7 @@ static SDValue combineMul(SDNode *N, Sel
int NumSign = SignMulAmt > 0 ? 1 : -1;
bool IsPowerOf2_64PlusOne = isPowerOf2_64(NumSign * SignMulAmt - 1);
bool IsPowerOf2_64MinusOne = isPowerOf2_64(NumSign * SignMulAmt + 1);
+ bool IsPowerOf2_64MinusTwo = isPowerOf2_64(NumSign * SignMulAmt + 2);
if (IsPowerOf2_64PlusOne) {
// (mul x, 2^N + 1) => (add (shl x, N), x)
NewMul = DAG.getNode(
@@ -33942,9 +33930,16 @@ static SDValue combineMul(SDNode *N, Sel
DAG.getConstant(Log2_64(NumSign * SignMulAmt + 1), DL,
MVT::i8)),
N->getOperand(0));
+ } else if (IsPowerOf2_64MinusTwo && NumSign == 1) {
+ // (mul x, 2^N - 1) => (sub (shl x, N), x)
+ NewMul = DAG.getNode(ISD::SHL, DL, VT, N->getOperand(0),
+ DAG.getConstant(Log2_64(NumSign * SignMulAmt + 2),
+ DL, MVT::i8));
+ NewMul = DAG.getNode(ISD::SUB, DL, VT, NewMul, N->getOperand(0));
+ NewMul = DAG.getNode(ISD::SUB, DL, VT, NewMul, N->getOperand(0));
}
// To negate, subtract the number from zero
- if ((IsPowerOf2_64PlusOne || IsPowerOf2_64MinusOne) && NumSign == -1)
+ if (NewMul && NumSign == -1)
NewMul =
DAG.getNode(ISD::SUB, DL, VT, DAG.getConstant(0, DL, VT), NewMul);
}
Modified: llvm/trunk/test/CodeGen/X86/mul-constant-i16.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/X86/mul-constant-i16.ll?rev=337856&r1=337855&r2=337856&view=diff
==============================================================================
--- llvm/trunk/test/CodeGen/X86/mul-constant-i16.ll (original)
+++ llvm/trunk/test/CodeGen/X86/mul-constant-i16.ll Tue Jul 24 14:15:41 2018
@@ -248,18 +248,19 @@ define i16 @test_mul_by_14(i16 %x) {
; X86-LABEL: test_mul_by_14:
; X86: # %bb.0:
; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx
-; X86-NEXT: leal (%ecx,%ecx,2), %eax
-; X86-NEXT: leal (%ecx,%eax,4), %eax
-; X86-NEXT: addl %ecx, %eax
+; X86-NEXT: movl %ecx, %eax
+; X86-NEXT: shll $4, %eax
+; X86-NEXT: subl %ecx, %eax
+; X86-NEXT: subl %ecx, %eax
; X86-NEXT: # kill: def $ax killed $ax killed $eax
; X86-NEXT: retl
;
; X64-LABEL: test_mul_by_14:
; X64: # %bb.0:
-; X64-NEXT: # kill: def $edi killed $edi def $rdi
-; X64-NEXT: leal (%rdi,%rdi,2), %eax
-; X64-NEXT: leal (%rdi,%rax,4), %eax
-; X64-NEXT: addl %edi, %eax
+; X64-NEXT: movl %edi, %eax
+; X64-NEXT: shll $4, %eax
+; X64-NEXT: subl %edi, %eax
+; X64-NEXT: subl %edi, %eax
; X64-NEXT: # kill: def $ax killed $ax killed $eax
; X64-NEXT: retq
%mul = mul nsw i16 %x, 14
@@ -638,6 +639,29 @@ define i16 @test_mul_by_32(i16 %x) {
ret i16 %mul
}
+define i16 @test_mul_by_62(i16 %x) {
+; X86-LABEL: test_mul_by_62:
+; X86: # %bb.0:
+; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx
+; X86-NEXT: movl %ecx, %eax
+; X86-NEXT: shll $6, %eax
+; X86-NEXT: subl %ecx, %eax
+; X86-NEXT: subl %ecx, %eax
+; X86-NEXT: # kill: def $ax killed $ax killed $eax
+; X86-NEXT: retl
+;
+; X64-LABEL: test_mul_by_62:
+; X64: # %bb.0:
+; X64-NEXT: movl %edi, %eax
+; X64-NEXT: shll $6, %eax
+; X64-NEXT: subl %edi, %eax
+; X64-NEXT: subl %edi, %eax
+; X64-NEXT: # kill: def $ax killed $ax killed $eax
+; X64-NEXT: retq
+ %mul = mul nsw i16 %x, 62
+ ret i16 %mul
+}
+
; (x*9+42)*(x*5+2)
define i16 @test_mul_spec(i16 %x) nounwind {
; X86-LABEL: test_mul_spec:
Modified: llvm/trunk/test/CodeGen/X86/mul-constant-i32.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/X86/mul-constant-i32.ll?rev=337856&r1=337855&r2=337856&view=diff
==============================================================================
--- llvm/trunk/test/CodeGen/X86/mul-constant-i32.ll (original)
+++ llvm/trunk/test/CodeGen/X86/mul-constant-i32.ll Tue Jul 24 14:15:41 2018
@@ -673,25 +673,26 @@ define i32 @test_mul_by_14(i32 %x) {
; X86-LABEL: test_mul_by_14:
; X86: # %bb.0:
; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx
-; X86-NEXT: leal (%ecx,%ecx,2), %eax
-; X86-NEXT: leal (%ecx,%eax,4), %eax
-; X86-NEXT: addl %ecx, %eax
+; X86-NEXT: movl %ecx, %eax
+; X86-NEXT: shll $4, %eax
+; X86-NEXT: subl %ecx, %eax
+; X86-NEXT: subl %ecx, %eax
; X86-NEXT: retl
;
; X64-HSW-LABEL: test_mul_by_14:
; X64-HSW: # %bb.0:
-; X64-HSW-NEXT: # kill: def $edi killed $edi def $rdi
-; X64-HSW-NEXT: leal (%rdi,%rdi,2), %eax # sched: [1:0.50]
-; X64-HSW-NEXT: leal (%rdi,%rax,4), %eax # sched: [1:0.50]
-; X64-HSW-NEXT: addl %edi, %eax # sched: [1:0.25]
+; X64-HSW-NEXT: movl %edi, %eax # sched: [1:0.25]
+; X64-HSW-NEXT: shll $4, %eax # sched: [1:0.50]
+; X64-HSW-NEXT: subl %edi, %eax # sched: [1:0.25]
+; X64-HSW-NEXT: subl %edi, %eax # sched: [1:0.25]
; X64-HSW-NEXT: retq # sched: [7:1.00]
;
; X64-JAG-LABEL: test_mul_by_14:
; X64-JAG: # %bb.0:
-; X64-JAG-NEXT: # kill: def $edi killed $edi def $rdi
-; X64-JAG-NEXT: leal (%rdi,%rdi,2), %eax # sched: [2:1.00]
-; X64-JAG-NEXT: leal (%rdi,%rax,4), %eax # sched: [2:1.00]
-; X64-JAG-NEXT: addl %edi, %eax # sched: [1:0.50]
+; X64-JAG-NEXT: movl %edi, %eax # sched: [1:0.50]
+; X64-JAG-NEXT: shll $4, %eax # sched: [1:0.50]
+; X64-JAG-NEXT: subl %edi, %eax # sched: [1:0.50]
+; X64-JAG-NEXT: subl %edi, %eax # sched: [1:0.50]
; X64-JAG-NEXT: retq # sched: [4:1.00]
;
; X86-NOOPT-LABEL: test_mul_by_14:
@@ -711,7 +712,10 @@ define i32 @test_mul_by_14(i32 %x) {
;
; X64-SLM-LABEL: test_mul_by_14:
; X64-SLM: # %bb.0:
-; X64-SLM-NEXT: imull $14, %edi, %eax # sched: [3:1.00]
+; X64-SLM-NEXT: movl %edi, %eax # sched: [1:0.50]
+; X64-SLM-NEXT: shll $4, %eax # sched: [1:1.00]
+; X64-SLM-NEXT: subl %edi, %eax # sched: [1:0.50]
+; X64-SLM-NEXT: subl %edi, %eax # sched: [1:0.50]
; X64-SLM-NEXT: retq # sched: [4:1.00]
;
; SLM-NOOPT-LABEL: test_mul_by_14:
@@ -1553,7 +1557,10 @@ define i32 @test_mul_by_30(i32 %x) {
;
; X64-SLM-LABEL: test_mul_by_30:
; X64-SLM: # %bb.0:
-; X64-SLM-NEXT: imull $30, %edi, %eax # sched: [3:1.00]
+; X64-SLM-NEXT: movl %edi, %eax # sched: [1:0.50]
+; X64-SLM-NEXT: shll $5, %eax # sched: [1:1.00]
+; X64-SLM-NEXT: subl %edi, %eax # sched: [1:0.50]
+; X64-SLM-NEXT: subl %edi, %eax # sched: [1:0.50]
; X64-SLM-NEXT: retq # sched: [4:1.00]
;
; SLM-NOOPT-LABEL: test_mul_by_30:
@@ -1669,6 +1676,63 @@ define i32 @test_mul_by_32(i32 %x) {
ret i32 %mul
}
+define i32 @test_mul_by_62(i32 %x) {
+; X86-LABEL: test_mul_by_62:
+; X86: # %bb.0:
+; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx
+; X86-NEXT: movl %ecx, %eax
+; X86-NEXT: shll $6, %eax
+; X86-NEXT: subl %ecx, %eax
+; X86-NEXT: subl %ecx, %eax
+; X86-NEXT: retl
+;
+; X64-HSW-LABEL: test_mul_by_62:
+; X64-HSW: # %bb.0:
+; X64-HSW-NEXT: movl %edi, %eax # sched: [1:0.25]
+; X64-HSW-NEXT: shll $6, %eax # sched: [1:0.50]
+; X64-HSW-NEXT: subl %edi, %eax # sched: [1:0.25]
+; X64-HSW-NEXT: subl %edi, %eax # sched: [1:0.25]
+; X64-HSW-NEXT: retq # sched: [7:1.00]
+;
+; X64-JAG-LABEL: test_mul_by_62:
+; X64-JAG: # %bb.0:
+; X64-JAG-NEXT: movl %edi, %eax # sched: [1:0.50]
+; X64-JAG-NEXT: shll $6, %eax # sched: [1:0.50]
+; X64-JAG-NEXT: subl %edi, %eax # sched: [1:0.50]
+; X64-JAG-NEXT: subl %edi, %eax # sched: [1:0.50]
+; X64-JAG-NEXT: retq # sched: [4:1.00]
+;
+; X86-NOOPT-LABEL: test_mul_by_62:
+; X86-NOOPT: # %bb.0:
+; X86-NOOPT-NEXT: imull $62, {{[0-9]+}}(%esp), %eax
+; X86-NOOPT-NEXT: retl
+;
+; HSW-NOOPT-LABEL: test_mul_by_62:
+; HSW-NOOPT: # %bb.0:
+; HSW-NOOPT-NEXT: imull $62, %edi, %eax # sched: [3:1.00]
+; HSW-NOOPT-NEXT: retq # sched: [7:1.00]
+;
+; JAG-NOOPT-LABEL: test_mul_by_62:
+; JAG-NOOPT: # %bb.0:
+; JAG-NOOPT-NEXT: imull $62, %edi, %eax # sched: [3:1.00]
+; JAG-NOOPT-NEXT: retq # sched: [4:1.00]
+;
+; X64-SLM-LABEL: test_mul_by_62:
+; X64-SLM: # %bb.0:
+; X64-SLM-NEXT: movl %edi, %eax # sched: [1:0.50]
+; X64-SLM-NEXT: shll $6, %eax # sched: [1:1.00]
+; X64-SLM-NEXT: subl %edi, %eax # sched: [1:0.50]
+; X64-SLM-NEXT: subl %edi, %eax # sched: [1:0.50]
+; X64-SLM-NEXT: retq # sched: [4:1.00]
+;
+; SLM-NOOPT-LABEL: test_mul_by_62:
+; SLM-NOOPT: # %bb.0:
+; SLM-NOOPT-NEXT: imull $62, %edi, %eax # sched: [3:1.00]
+; SLM-NOOPT-NEXT: retq # sched: [4:1.00]
+ %mul = mul nsw i32 %x, 62
+ ret i32 %mul
+}
+
; (x*9+42)*(x*5+2)
define i32 @test_mul_spec(i32 %x) nounwind {
; X86-LABEL: test_mul_spec:
Modified: llvm/trunk/test/CodeGen/X86/mul-constant-i64.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/X86/mul-constant-i64.ll?rev=337856&r1=337855&r2=337856&view=diff
==============================================================================
--- llvm/trunk/test/CodeGen/X86/mul-constant-i64.ll (original)
+++ llvm/trunk/test/CodeGen/X86/mul-constant-i64.ll Tue Jul 24 14:15:41 2018
@@ -686,9 +686,10 @@ define i64 @test_mul_by_14(i64 %x) {
; X86-LABEL: test_mul_by_14:
; X86: # %bb.0:
; X86-NEXT: movl {{[0-9]+}}(%esp), %eax
-; X86-NEXT: leal (%eax,%eax,2), %ecx
-; X86-NEXT: leal (%eax,%ecx,4), %ecx
-; X86-NEXT: addl %eax, %ecx
+; X86-NEXT: movl %eax, %ecx
+; X86-NEXT: shll $4, %ecx
+; X86-NEXT: subl %eax, %ecx
+; X86-NEXT: subl %eax, %ecx
; X86-NEXT: movl $14, %eax
; X86-NEXT: mull {{[0-9]+}}(%esp)
; X86-NEXT: addl %ecx, %edx
@@ -696,16 +697,18 @@ define i64 @test_mul_by_14(i64 %x) {
;
; X64-HSW-LABEL: test_mul_by_14:
; X64-HSW: # %bb.0:
-; X64-HSW-NEXT: leaq (%rdi,%rdi,2), %rax # sched: [1:0.50]
-; X64-HSW-NEXT: leaq (%rdi,%rax,4), %rax # sched: [1:0.50]
-; X64-HSW-NEXT: addq %rdi, %rax # sched: [1:0.25]
+; X64-HSW-NEXT: movq %rdi, %rax # sched: [1:0.25]
+; X64-HSW-NEXT: shlq $4, %rax # sched: [1:0.50]
+; X64-HSW-NEXT: subq %rdi, %rax # sched: [1:0.25]
+; X64-HSW-NEXT: subq %rdi, %rax # sched: [1:0.25]
; X64-HSW-NEXT: retq # sched: [7:1.00]
;
; X64-JAG-LABEL: test_mul_by_14:
; X64-JAG: # %bb.0:
-; X64-JAG-NEXT: leaq (%rdi,%rdi,2), %rax # sched: [2:1.00]
-; X64-JAG-NEXT: leaq (%rdi,%rax,4), %rax # sched: [2:1.00]
-; X64-JAG-NEXT: addq %rdi, %rax # sched: [1:0.50]
+; X64-JAG-NEXT: movq %rdi, %rax # sched: [1:0.50]
+; X64-JAG-NEXT: shlq $4, %rax # sched: [1:0.50]
+; X64-JAG-NEXT: subq %rdi, %rax # sched: [1:0.50]
+; X64-JAG-NEXT: subq %rdi, %rax # sched: [1:0.50]
; X64-JAG-NEXT: retq # sched: [4:1.00]
;
; X86-NOOPT-LABEL: test_mul_by_14:
@@ -728,7 +731,10 @@ define i64 @test_mul_by_14(i64 %x) {
;
; X64-SLM-LABEL: test_mul_by_14:
; X64-SLM: # %bb.0:
-; X64-SLM-NEXT: imulq $14, %rdi, %rax # sched: [3:1.00]
+; X64-SLM-NEXT: movq %rdi, %rax # sched: [1:0.50]
+; X64-SLM-NEXT: shlq $4, %rax # sched: [1:1.00]
+; X64-SLM-NEXT: subq %rdi, %rax # sched: [1:0.50]
+; X64-SLM-NEXT: subq %rdi, %rax # sched: [1:0.50]
; X64-SLM-NEXT: retq # sched: [4:1.00]
;
; SLM-NOOPT-LABEL: test_mul_by_14:
@@ -1626,7 +1632,10 @@ define i64 @test_mul_by_30(i64 %x) {
;
; X64-SLM-LABEL: test_mul_by_30:
; X64-SLM: # %bb.0:
-; X64-SLM-NEXT: imulq $30, %rdi, %rax # sched: [3:1.00]
+; X64-SLM-NEXT: movq %rdi, %rax # sched: [1:0.50]
+; X64-SLM-NEXT: shlq $5, %rax # sched: [1:1.00]
+; X64-SLM-NEXT: subq %rdi, %rax # sched: [1:0.50]
+; X64-SLM-NEXT: subq %rdi, %rax # sched: [1:0.50]
; X64-SLM-NEXT: retq # sched: [4:1.00]
;
; SLM-NOOPT-LABEL: test_mul_by_30:
@@ -1752,6 +1761,69 @@ define i64 @test_mul_by_32(i64 %x) {
ret i64 %mul
}
+define i64 @test_mul_by_62(i64 %x) {
+; X86-LABEL: test_mul_by_62:
+; X86: # %bb.0:
+; X86-NEXT: movl {{[0-9]+}}(%esp), %eax
+; X86-NEXT: movl %eax, %ecx
+; X86-NEXT: shll $6, %ecx
+; X86-NEXT: subl %eax, %ecx
+; X86-NEXT: subl %eax, %ecx
+; X86-NEXT: movl $62, %eax
+; X86-NEXT: mull {{[0-9]+}}(%esp)
+; X86-NEXT: addl %ecx, %edx
+; X86-NEXT: retl
+;
+; X64-HSW-LABEL: test_mul_by_62:
+; X64-HSW: # %bb.0:
+; X64-HSW-NEXT: movq %rdi, %rax # sched: [1:0.25]
+; X64-HSW-NEXT: shlq $6, %rax # sched: [1:0.50]
+; X64-HSW-NEXT: subq %rdi, %rax # sched: [1:0.25]
+; X64-HSW-NEXT: subq %rdi, %rax # sched: [1:0.25]
+; X64-HSW-NEXT: retq # sched: [7:1.00]
+;
+; X64-JAG-LABEL: test_mul_by_62:
+; X64-JAG: # %bb.0:
+; X64-JAG-NEXT: movq %rdi, %rax # sched: [1:0.50]
+; X64-JAG-NEXT: shlq $6, %rax # sched: [1:0.50]
+; X64-JAG-NEXT: subq %rdi, %rax # sched: [1:0.50]
+; X64-JAG-NEXT: subq %rdi, %rax # sched: [1:0.50]
+; X64-JAG-NEXT: retq # sched: [4:1.00]
+;
+; X86-NOOPT-LABEL: test_mul_by_62:
+; X86-NOOPT: # %bb.0:
+; X86-NOOPT-NEXT: movl $62, %eax
+; X86-NOOPT-NEXT: mull {{[0-9]+}}(%esp)
+; X86-NOOPT-NEXT: imull $62, {{[0-9]+}}(%esp), %ecx
+; X86-NOOPT-NEXT: addl %ecx, %edx
+; X86-NOOPT-NEXT: retl
+;
+; HSW-NOOPT-LABEL: test_mul_by_62:
+; HSW-NOOPT: # %bb.0:
+; HSW-NOOPT-NEXT: imulq $62, %rdi, %rax # sched: [3:1.00]
+; HSW-NOOPT-NEXT: retq # sched: [7:1.00]
+;
+; JAG-NOOPT-LABEL: test_mul_by_62:
+; JAG-NOOPT: # %bb.0:
+; JAG-NOOPT-NEXT: imulq $62, %rdi, %rax # sched: [6:4.00]
+; JAG-NOOPT-NEXT: retq # sched: [4:1.00]
+;
+; X64-SLM-LABEL: test_mul_by_62:
+; X64-SLM: # %bb.0:
+; X64-SLM-NEXT: movq %rdi, %rax # sched: [1:0.50]
+; X64-SLM-NEXT: shlq $6, %rax # sched: [1:1.00]
+; X64-SLM-NEXT: subq %rdi, %rax # sched: [1:0.50]
+; X64-SLM-NEXT: subq %rdi, %rax # sched: [1:0.50]
+; X64-SLM-NEXT: retq # sched: [4:1.00]
+;
+; SLM-NOOPT-LABEL: test_mul_by_62:
+; SLM-NOOPT: # %bb.0:
+; SLM-NOOPT-NEXT: imulq $62, %rdi, %rax # sched: [3:1.00]
+; SLM-NOOPT-NEXT: retq # sched: [4:1.00]
+ %mul = mul nsw i64 %x, 62
+ ret i64 %mul
+}
+
; (x*9+42)*(x*5+2)
define i64 @test_mul_spec(i64 %x) nounwind {
; X86-LABEL: test_mul_spec:
Modified: llvm/trunk/test/CodeGen/X86/mul-constant-result.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/X86/mul-constant-result.ll?rev=337856&r1=337855&r2=337856&view=diff
==============================================================================
--- llvm/trunk/test/CodeGen/X86/mul-constant-result.ll (original)
+++ llvm/trunk/test/CodeGen/X86/mul-constant-result.ll Tue Jul 24 14:15:41 2018
@@ -113,8 +113,10 @@ define i32 @mult(i32, i32) local_unnamed
; X86-NEXT: retl
; X86-NEXT: .LBB0_19:
; X86-NEXT: .cfi_def_cfa_offset 8
-; X86-NEXT: leal (%eax,%eax,2), %ecx
-; X86-NEXT: jmp .LBB0_20
+; X86-NEXT: movl %eax, %ecx
+; X86-NEXT: shll $4, %ecx
+; X86-NEXT: subl %eax, %ecx
+; X86-NEXT: jmp .LBB0_12
; X86-NEXT: .LBB0_21:
; X86-NEXT: leal (%eax,%eax,4), %eax
; X86-NEXT: leal (%eax,%eax,2), %eax
@@ -166,7 +168,6 @@ define i32 @mult(i32, i32) local_unnamed
; X86-NEXT: .LBB0_28:
; X86-NEXT: .cfi_def_cfa_offset 8
; X86-NEXT: leal (%eax,%eax,4), %ecx
-; X86-NEXT: .LBB0_20:
; X86-NEXT: leal (%eax,%ecx,4), %ecx
; X86-NEXT: addl %ecx, %eax
; X86-NEXT: popl %esi
@@ -312,21 +313,23 @@ define i32 @mult(i32, i32) local_unnamed
; X64-HSW-NEXT: # kill: def $eax killed $eax killed $rax
; X64-HSW-NEXT: retq
; X64-HSW-NEXT: .LBB0_15:
-; X64-HSW-NEXT: leal (%rax,%rax,2), %ecx
-; X64-HSW-NEXT: jmp .LBB0_16
-; X64-HSW-NEXT: .LBB0_18:
+; X64-HSW-NEXT: movl %eax, %ecx
+; X64-HSW-NEXT: shll $4, %ecx
+; X64-HSW-NEXT: subl %eax, %ecx
+; X64-HSW-NEXT: jmp .LBB0_8
+; X64-HSW-NEXT: .LBB0_17:
; X64-HSW-NEXT: leal (%rax,%rax,4), %eax
; X64-HSW-NEXT: leal (%rax,%rax,2), %eax
; X64-HSW-NEXT: # kill: def $eax killed $eax killed $rax
; X64-HSW-NEXT: retq
-; X64-HSW-NEXT: .LBB0_19:
+; X64-HSW-NEXT: .LBB0_18:
; X64-HSW-NEXT: shll $4, %eax
; X64-HSW-NEXT: # kill: def $eax killed $eax killed $rax
; X64-HSW-NEXT: retq
-; X64-HSW-NEXT: .LBB0_20:
+; X64-HSW-NEXT: .LBB0_19:
; X64-HSW-NEXT: movl %eax, %ecx
; X64-HSW-NEXT: shll $4, %ecx
-; X64-HSW-NEXT: jmp .LBB0_17
+; X64-HSW-NEXT: jmp .LBB0_20
; X64-HSW-NEXT: .LBB0_21:
; X64-HSW-NEXT: addl %eax, %eax
; X64-HSW-NEXT: leal (%rax,%rax,8), %eax
@@ -349,9 +352,8 @@ define i32 @mult(i32, i32) local_unnamed
; X64-HSW-NEXT: retq
; X64-HSW-NEXT: .LBB0_25:
; X64-HSW-NEXT: leal (%rax,%rax,4), %ecx
-; X64-HSW-NEXT: .LBB0_16:
; X64-HSW-NEXT: leal (%rax,%rcx,4), %ecx
-; X64-HSW-NEXT: jmp .LBB0_17
+; X64-HSW-NEXT: jmp .LBB0_20
; X64-HSW-NEXT: .LBB0_26:
; X64-HSW-NEXT: leal (%rax,%rax,2), %ecx
; X64-HSW-NEXT: shll $3, %ecx
@@ -378,12 +380,12 @@ define i32 @mult(i32, i32) local_unnamed
; X64-HSW-NEXT: .LBB0_31:
; X64-HSW-NEXT: leal (%rax,%rax,8), %ecx
; X64-HSW-NEXT: leal (%rcx,%rcx,2), %ecx
-; X64-HSW-NEXT: jmp .LBB0_17
+; X64-HSW-NEXT: jmp .LBB0_20
; X64-HSW-NEXT: .LBB0_32:
; X64-HSW-NEXT: leal (%rax,%rax,8), %ecx
; X64-HSW-NEXT: leal (%rcx,%rcx,2), %ecx
; X64-HSW-NEXT: addl %eax, %ecx
-; X64-HSW-NEXT: .LBB0_17:
+; X64-HSW-NEXT: .LBB0_20:
; X64-HSW-NEXT: addl %eax, %ecx
; X64-HSW-NEXT: movl %ecx, %eax
; X64-HSW-NEXT: # kill: def $eax killed $eax killed $rax
More information about the llvm-commits
mailing list