1
0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2025-01-31 20:51:52 +01:00
llvm-mirror/test/CodeGen/X86/lsr-static-addr.ll
Craig Topper b18c061910 [X86] Initial cleanups on the FixupLEAs pass. Separate Atom LEA creation from other LEA optimizations.
This removes some of the class variables. Merge basic block processing into
runOnMachineFunction to keep the flags local.

Pass MachineBasicBlock around instead of an iterator. We can get the iterator in
the few places that need it. Allows a range-based outer for loop.

Separate the Atom optimization from the rest of the optimizations. This allows
fixupIncDec to create INC/DEC and still allow Atom to turn it back into LEA
when profitable by its heuristics.

I'd like to improve fixupIncDec to turn LEAs into ADD any time the base or index
register is equal to the destination register. This is profitable regardless of
the various slow flags. But again we would want Atom to be able to undo that.

Differential Revision: https://reviews.llvm.org/D60993

llvm-svn: 359581
2019-04-30 17:56:28 +00:00

66 lines
2.2 KiB
LLVM

; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -mcpu=generic -mtriple=x86_64-unknown-linux-gnu -relocation-model=static < %s | FileCheck %s --check-prefix=CHECK
; RUN: llc -mcpu=atom -mtriple=x86_64-unknown-linux-gnu -relocation-model=static < %s | FileCheck %s --check-prefix=ATOM
@A = external global [0 x double]
define void @foo(i64 %n) nounwind {
; CHECK-LABEL: foo:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: testq %rdi, %rdi
; CHECK-NEXT: jle .LBB0_3
; CHECK-NEXT: # %bb.1: # %for.body.preheader
; CHECK-NEXT: xorl %eax, %eax
; CHECK-NEXT: movsd {{.*#+}} xmm0 = mem[0],zero
; CHECK-NEXT: .p2align 4, 0x90
; CHECK-NEXT: .LBB0_2: # %for.body
; CHECK-NEXT: # =>This Inner Loop Header: Depth=1
; CHECK-NEXT: movsd {{.*#+}} xmm1 = mem[0],zero
; CHECK-NEXT: mulsd %xmm0, %xmm1
; CHECK-NEXT: movsd %xmm1, A(,%rax,8)
; CHECK-NEXT: incq %rax
; CHECK-NEXT: cmpq %rax, %rdi
; CHECK-NEXT: jne .LBB0_2
; CHECK-NEXT: .LBB0_3: # %for.end
; CHECK-NEXT: retq
;
; ATOM-LABEL: foo:
; ATOM: # %bb.0: # %entry
; ATOM-NEXT: testq %rdi, %rdi
; ATOM-NEXT: jle .LBB0_3
; ATOM-NEXT: # %bb.1: # %for.body.preheader
; ATOM-NEXT: xorl %eax, %eax
; ATOM-NEXT: movsd {{.*#+}} xmm0 = mem[0],zero
; ATOM-NEXT: .p2align 4, 0x90
; ATOM-NEXT: .LBB0_2: # %for.body
; ATOM-NEXT: # =>This Inner Loop Header: Depth=1
; ATOM-NEXT: movsd {{.*#+}} xmm1 = mem[0],zero
; ATOM-NEXT: mulsd %xmm0, %xmm1
; ATOM-NEXT: movsd %xmm1, A(,%rax,8)
; ATOM-NEXT: leaq 1(%rax), %rax
; ATOM-NEXT: cmpq %rax, %rdi
; ATOM-NEXT: jne .LBB0_2
; ATOM-NEXT: .LBB0_3: # %for.end
; ATOM-NEXT: nop
; ATOM-NEXT: nop
; ATOM-NEXT: nop
; ATOM-NEXT: nop
; ATOM-NEXT: retq
entry:
%cmp5 = icmp sgt i64 %n, 0
br i1 %cmp5, label %for.body, label %for.end
for.body:
%i.06 = phi i64 [ %inc, %for.body ], [ 0, %entry ]
%arrayidx = getelementptr [0 x double], [0 x double]* @A, i64 0, i64 %i.06
%tmp3 = load double, double* %arrayidx, align 8
%mul = fmul double %tmp3, 2.300000e+00
store double %mul, double* %arrayidx, align 8
%inc = add nsw i64 %i.06, 1
%exitcond = icmp eq i64 %inc, %n
br i1 %exitcond, label %for.end, label %for.body
for.end:
ret void
}