llvm-for-llvmta/test/CodeGen/X86/2011-10-19-widen_vselect.ll

128 lines
4.2 KiB
LLVM

; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mtriple=i686-unknown-linux-gnu -mcpu=corei7 | FileCheck %s --check-prefix=X86
; RUN: llc < %s -mtriple=x86_64-unknown-linux-gnu -mcpu=corei7 | FileCheck %s --check-prefix=X64
; Make sure that we don't crash when legalizing vselect and vsetcc and that
; we are able to generate vector blend instructions.
define void @simple_widen(<2 x float> %a, <2 x float> %b) {
; X86-LABEL: simple_widen:
; X86: # %bb.0: # %entry
; X86-NEXT: movlps %xmm1, (%eax)
; X86-NEXT: retl
;
; X64-LABEL: simple_widen:
; X64: # %bb.0: # %entry
; X64-NEXT: movlps %xmm1, (%rax)
; X64-NEXT: retq
entry:
%0 = select <2 x i1> undef, <2 x float> %a, <2 x float> %b
store <2 x float> %0, <2 x float>* undef
ret void
}
define void @complex_inreg_work(<2 x float> %a, <2 x float> %b, <2 x float> %c) {
; X86-LABEL: complex_inreg_work:
; X86: # %bb.0: # %entry
; X86-NEXT: movaps %xmm0, %xmm3
; X86-NEXT: cmpordps %xmm2, %xmm2
; X86-NEXT: movaps %xmm2, %xmm0
; X86-NEXT: blendvps %xmm0, %xmm3, %xmm1
; X86-NEXT: movlps %xmm1, (%eax)
; X86-NEXT: retl
;
; X64-LABEL: complex_inreg_work:
; X64: # %bb.0: # %entry
; X64-NEXT: movaps %xmm0, %xmm3
; X64-NEXT: cmpordps %xmm2, %xmm2
; X64-NEXT: movaps %xmm2, %xmm0
; X64-NEXT: blendvps %xmm0, %xmm3, %xmm1
; X64-NEXT: movlps %xmm1, (%rax)
; X64-NEXT: retq
entry:
%0 = fcmp oeq <2 x float> %c, %c
%1 = select <2 x i1> %0, <2 x float> %a, <2 x float> %b
store <2 x float> %1, <2 x float>* undef
ret void
}
define void @zero_test() {
; X86-LABEL: zero_test:
; X86: # %bb.0: # %entry
; X86-NEXT: xorps %xmm0, %xmm0
; X86-NEXT: movlps %xmm0, (%eax)
; X86-NEXT: retl
;
; X64-LABEL: zero_test:
; X64: # %bb.0: # %entry
; X64-NEXT: xorps %xmm0, %xmm0
; X64-NEXT: movlps %xmm0, (%rax)
; X64-NEXT: retq
entry:
%0 = select <2 x i1> undef, <2 x float> undef, <2 x float> zeroinitializer
store <2 x float> %0, <2 x float>* undef
ret void
}
define void @full_test() {
; X86-LABEL: full_test:
; X86: # %bb.0: # %entry
; X86-NEXT: subl $60, %esp
; X86-NEXT: .cfi_def_cfa_offset 64
; X86-NEXT: movsd {{.*#+}} xmm2 = mem[0],zero
; X86-NEXT: cvttps2dq %xmm2, %xmm0
; X86-NEXT: cvtdq2ps %xmm0, %xmm1
; X86-NEXT: xorps %xmm0, %xmm0
; X86-NEXT: cmpltps %xmm2, %xmm0
; X86-NEXT: movaps {{.*#+}} xmm3 = <1.0E+0,1.0E+0,u,u>
; X86-NEXT: addps %xmm1, %xmm3
; X86-NEXT: movaps %xmm1, %xmm4
; X86-NEXT: blendvps %xmm0, %xmm3, %xmm4
; X86-NEXT: cmpeqps %xmm2, %xmm1
; X86-NEXT: movaps %xmm1, %xmm0
; X86-NEXT: blendvps %xmm0, %xmm2, %xmm4
; X86-NEXT: movlps %xmm4, {{[0-9]+}}(%esp)
; X86-NEXT: movlps %xmm4, {{[0-9]+}}(%esp)
; X86-NEXT: addl $60, %esp
; X86-NEXT: .cfi_def_cfa_offset 4
; X86-NEXT: retl
;
; X64-LABEL: full_test:
; X64: # %bb.0: # %entry
; X64-NEXT: movsd {{.*#+}} xmm2 = mem[0],zero
; X64-NEXT: cvttps2dq %xmm2, %xmm0
; X64-NEXT: cvtdq2ps %xmm0, %xmm1
; X64-NEXT: xorps %xmm0, %xmm0
; X64-NEXT: cmpltps %xmm2, %xmm0
; X64-NEXT: movaps {{.*#+}} xmm3 = <1.0E+0,1.0E+0,u,u>
; X64-NEXT: addps %xmm1, %xmm3
; X64-NEXT: movaps %xmm1, %xmm4
; X64-NEXT: blendvps %xmm0, %xmm3, %xmm4
; X64-NEXT: cmpeqps %xmm2, %xmm1
; X64-NEXT: movaps %xmm1, %xmm0
; X64-NEXT: blendvps %xmm0, %xmm2, %xmm4
; X64-NEXT: movlps %xmm4, -{{[0-9]+}}(%rsp)
; X64-NEXT: movlps %xmm4, -{{[0-9]+}}(%rsp)
; X64-NEXT: retq
entry:
%Cy300 = alloca <4 x float>
%Cy11a = alloca <2 x float>
%Cy118 = alloca <2 x float>
%Cy119 = alloca <2 x float>
br label %B1
B1: ; preds = %entry
%0 = load <2 x float>, <2 x float>* %Cy119
%1 = fptosi <2 x float> %0 to <2 x i32>
%2 = sitofp <2 x i32> %1 to <2 x float>
%3 = fcmp ogt <2 x float> %0, zeroinitializer
%4 = fadd <2 x float> %2, <float 1.000000e+00, float 1.000000e+00>
%5 = select <2 x i1> %3, <2 x float> %4, <2 x float> %2
%6 = fcmp oeq <2 x float> %2, %0
%7 = select <2 x i1> %6, <2 x float> %0, <2 x float> %5
store <2 x float> %7, <2 x float>* %Cy118
%8 = load <2 x float>, <2 x float>* %Cy118
store <2 x float> %8, <2 x float>* %Cy11a
ret void
}