llvm/test/CodeGen/Thumb2/mve-complex-deinterleaving-i64-add.ll

; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s --mattr=+mve -o - --verify-machineinstrs | FileCheck %s

target triple = "thumbv8.1m.main-none-none-eabi"


; Expected to not transform
define arm_aapcs_vfpcc <2 x i64> @complex_add_v2i64(<2 x i64> %a, <2 x i64> %b) {
; CHECK-LABEL: complex_add_v2i64:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .save {r7, lr}
; CHECK-NEXT:    push {r7, lr}
; CHECK-NEXT:    vmov r0, r1, d0
; CHECK-NEXT:    vmov r2, r3, d3
; CHECK-NEXT:    adds.w lr, r2, r0
; CHECK-NEXT:    adc.w r12, r3, r1
; CHECK-NEXT:    vmov r2, r3, d1
; CHECK-NEXT:    vmov r1, r0, d2
; CHECK-NEXT:    subs r1, r1, r2
; CHECK-NEXT:    vmov q0[2], q0[0], r1, lr
; CHECK-NEXT:    sbcs r0, r3
; CHECK-NEXT:    vmov q0[3], q0[1], r0, r12
; CHECK-NEXT:    pop {r7, pc}
entry:
  %a.real = shufflevector <2 x i64> %a, <2 x i64> zeroinitializer, <1 x i32> <i32 0>
  %a.imag = shufflevector <2 x i64> %a, <2 x i64> zeroinitializer, <1 x i32> <i32 1>
  %b.real = shufflevector <2 x i64> %b, <2 x i64> zeroinitializer, <1 x i32> <i32 0>
  %b.imag = shufflevector <2 x i64> %b, <2 x i64> zeroinitializer, <1 x i32> <i32 1>
  %0 = sub <1 x i64> %b.real, %a.imag
  %1 = add <1 x i64> %b.imag, %a.real
  %interleaved.vec = shufflevector <1 x i64> %0, <1 x i64> %1, <2 x i32> <i32 0, i32 1>
  ret <2 x i64> %interleaved.vec
}

; Expected to not transform
define arm_aapcs_vfpcc <4 x i64> @complex_add_v4i64(<4 x i64> %a, <4 x i64> %b) {
; CHECK-LABEL: complex_add_v4i64:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .save {r7, lr}
; CHECK-NEXT:    push {r7, lr}
; CHECK-NEXT:    .vsave {d8, d9}
; CHECK-NEXT:    vpush {d8, d9}
; CHECK-NEXT:    vmov q4, q1
; CHECK-NEXT:    vmov r2, r3, d7
; CHECK-NEXT:    vmov r0, r1, d8
; CHECK-NEXT:    adds.w lr, r2, r0
; CHECK-NEXT:    adc.w r12, r3, r1
; CHECK-NEXT:    vmov r2, r3, d0
; CHECK-NEXT:    vmov r1, r0, d5
; CHECK-NEXT:    adds r1, r1, r2
; CHECK-NEXT:    adcs r0, r3
; CHECK-NEXT:    vmov q1[2], q1[0], r1, lr
; CHECK-NEXT:    vmov q1[3], q1[1], r0, r12
; CHECK-NEXT:    vmov r0, r1, d9
; CHECK-NEXT:    vmov r2, r3, d6
; CHECK-NEXT:    subs.w lr, r2, r0
; CHECK-NEXT:    sbc.w r12, r3, r1
; CHECK-NEXT:    vmov r2, r3, d1
; CHECK-NEXT:    vmov r1, r0, d4
; CHECK-NEXT:    vmov.f32 s2, s4
; CHECK-NEXT:    vmov.f32 s3, s5
; CHECK-NEXT:    subs r1, r1, r2
; CHECK-NEXT:    vmov q2[2], q2[0], r1, lr
; CHECK-NEXT:    sbcs r0, r3
; CHECK-NEXT:    vmov q2[3], q2[1], r0, r12
; CHECK-NEXT:    vmov.f32 s0, s8
; CHECK-NEXT:    vmov.f32 s4, s10
; CHECK-NEXT:    vmov.f32 s1, s9
; CHECK-NEXT:    vmov.f32 s5, s11
; CHECK-NEXT:    vpop {d8, d9}
; CHECK-NEXT:    pop {r7, pc}
entry:
  %a.real = shufflevector <4 x i64> %a, <4 x i64> zeroinitializer, <2 x i32> <i32 0, i32 2>
  %a.imag = shufflevector <4 x i64> %a, <4 x i64> zeroinitializer, <2 x i32> <i32 1, i32 3>
  %b.real = shufflevector <4 x i64> %b, <4 x i64> zeroinitializer, <2 x i32> <i32 0, i32 2>
  %b.imag = shufflevector <4 x i64> %b, <4 x i64> zeroinitializer, <2 x i32> <i32 1, i32 3>
  %0 = sub <2 x i64> %b.real, %a.imag
  %1 = add <2 x i64> %b.imag, %a.real
  %interleaved.vec = shufflevector <2 x i64> %0, <2 x i64> %1, <4 x i32> <i32 0, i32 2, i32 1, i32 3>
  ret <4 x i64> %interleaved.vec
}

; Expected to not transform
define arm_aapcs_vfpcc <8 x i64> @complex_add_v8i64(<8 x i64> %a, <8 x i64> %b) {
; CHECK-LABEL: complex_add_v8i64:
; CHECK:       @ %bb.0: @ %entry
; CHECK-NEXT:    .save {r7, lr}
; CHECK-NEXT:    push {r7, lr}
; CHECK-NEXT:    .vsave {d8, d9, d10, d11, d12, d13}
; CHECK-NEXT:    vpush {d8, d9, d10, d11, d12, d13}
; CHECK-NEXT:    add r2, sp, #72
; CHECK-NEXT:    vmov q4, q1
; CHECK-NEXT:    vldrw.u32 q5, [r2]
; CHECK-NEXT:    vmov r0, r1, d8
; CHECK-NEXT:    vmov r2, r3, d11
; CHECK-NEXT:    adds.w lr, r2, r0
; CHECK-NEXT:    adc.w r12, r3, r1
; CHECK-NEXT:    add r1, sp, #56
; CHECK-NEXT:    vldrw.u32 q6, [r1]
; CHECK-NEXT:    vmov r2, r3, d0
; CHECK-NEXT:    vmov r1, r0, d13
; CHECK-NEXT:    adds r1, r1, r2
; CHECK-NEXT:    adcs r0, r3
; CHECK-NEXT:    vmov q1[2], q1[0], r1, lr
; CHECK-NEXT:    vmov q1[3], q1[1], r0, r12
; CHECK-NEXT:    vmov r0, r1, d9
; CHECK-NEXT:    vmov r2, r3, d10
; CHECK-NEXT:    subs.w lr, r2, r0
; CHECK-NEXT:    sbc.w r12, r3, r1
; CHECK-NEXT:    vmov r2, r3, d1
; CHECK-NEXT:    vmov r1, r0, d12
; CHECK-NEXT:    vmov.f32 s2, s4
; CHECK-NEXT:    vmov.f32 s3, s5
; CHECK-NEXT:    subs r1, r1, r2
; CHECK-NEXT:    add r2, sp, #104
; CHECK-NEXT:    vldrw.u32 q5, [r2]
; CHECK-NEXT:    sbcs r0, r3
; CHECK-NEXT:    vmov q4[2], q4[0], r1, lr
; CHECK-NEXT:    vmov q4[3], q4[1], r0, r12
; CHECK-NEXT:    vmov r0, r1, d6
; CHECK-NEXT:    vmov r2, r3, d11
; CHECK-NEXT:    vmov.f32 s0, s16
; CHECK-NEXT:    vmov.f32 s4, s18
; CHECK-NEXT:    vmov.f32 s1, s17
; CHECK-NEXT:    vmov.f32 s5, s19
; CHECK-NEXT:    adds.w lr, r2, r0
; CHECK-NEXT:    adc.w r12, r3, r1
; CHECK-NEXT:    add r1, sp, #88
; CHECK-NEXT:    vldrw.u32 q6, [r1]
; CHECK-NEXT:    vmov r2, r3, d4
; CHECK-NEXT:    vmov r1, r0, d13
; CHECK-NEXT:    adds r1, r1, r2
; CHECK-NEXT:    adcs r0, r3
; CHECK-NEXT:    vmov q4[2], q4[0], r1, lr
; CHECK-NEXT:    vmov q4[3], q4[1], r0, r12
; CHECK-NEXT:    vmov r0, r1, d7
; CHECK-NEXT:    vmov r2, r3, d10
; CHECK-NEXT:    subs.w lr, r2, r0
; CHECK-NEXT:    sbc.w r12, r3, r1
; CHECK-NEXT:    vmov r2, r3, d5
; CHECK-NEXT:    vmov r1, r0, d12
; CHECK-NEXT:    vmov.f32 s10, s16
; CHECK-NEXT:    vmov.f32 s11, s17
; CHECK-NEXT:    subs r1, r1, r2
; CHECK-NEXT:    vmov q3[2], q3[0], r1, lr
; CHECK-NEXT:    sbcs r0, r3
; CHECK-NEXT:    vmov q3[3], q3[1], r0, r12
; CHECK-NEXT:    vmov.f32 s16, s14
; CHECK-NEXT:    vmov.f32 s8, s12
; CHECK-NEXT:    vmov.f32 s17, s15
; CHECK-NEXT:    vmov.f32 s9, s13
; CHECK-NEXT:    vmov q3, q4
; CHECK-NEXT:    vpop {d8, d9, d10, d11, d12, d13}
; CHECK-NEXT:    pop {r7, pc}
entry:
  %a.real = shufflevector <8 x i64> %a, <8 x i64> zeroinitializer, <4 x i32> <i32 0, i32 2, i32 4, i32 6>
  %a.imag = shufflevector <8 x i64> %a, <8 x i64> zeroinitializer, <4 x i32> <i32 1, i32 3, i32 5, i32 7>
  %b.real = shufflevector <8 x i64> %b, <8 x i64> zeroinitializer, <4 x i32> <i32 0, i32 2, i32 4, i32 6>
  %b.imag = shufflevector <8 x i64> %b, <8 x i64> zeroinitializer, <4 x i32> <i32 1, i32 3, i32 5, i32 7>
  %0 = sub <4 x i64> %b.real, %a.imag
  %1 = add <4 x i64> %b.imag, %a.real
  %interleaved.vec = shufflevector <4 x i64> %0, <4 x i64> %1, <8 x i32> <i32 0, i32 4, i32 1, i32 5, i32 2, i32 6, i32 3, i32 7>
  ret <8 x i64> %interleaved.vec
}