llvm/test/Transforms/VectorCombine/X86/permute-of-binops.ll

; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 4
; RUN: opt < %s -passes=vector-combine -S -mtriple=x86_64-- -mattr=sse2 | FileCheck %s
; RUN: opt < %s -passes=vector-combine -S -mtriple=x86_64-- -mattr=avx2 | FileCheck %s

; Fold "shuffle (binop (shuffle, shuffle)), undef" --> "binop (shuffle), (shuffle)"

declare void @use_v4f64(<4 x double>)

define <4 x double> @fadd_v4f64(<4 x double> %a, <4 x double> %b) {
; CHECK-LABEL: define <4 x double> @fadd_v4f64(
; CHECK-SAME: <4 x double> [[A:%.*]], <4 x double> [[B:%.*]]) #[[ATTR0:[0-9]+]] {
; CHECK-NEXT:    [[TMP1:%.*]] = shufflevector <4 x double> [[A]], <4 x double> poison, <4 x i32> <i32 2, i32 3, i32 0, i32 1>
; CHECK-NEXT:    [[TMP2:%.*]] = shufflevector <4 x double> [[B]], <4 x double> poison, <4 x i32> <i32 0, i32 1, i32 0, i32 1>
; CHECK-NEXT:    [[POST:%.*]] = fadd <4 x double> [[TMP1]], [[TMP2]]
; CHECK-NEXT:    ret <4 x double> [[POST]]
;
  %a1 = shufflevector <4 x double> %a, <4 x double> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
  %b1 = shufflevector <4 x double> %b, <4 x double> poison, <4 x i32> <i32 1, i32 0, i32 1, i32 0>
  %op = fadd <4 x double> %a1, %b1
  %post = shufflevector <4 x double> %op, <4 x double> poison, <4 x i32> <i32 1, i32 0, i32 3, i32 2>
  ret <4 x double> %post
}

define <4 x double> @fadd_v4f64_poison_idx(<4 x double> %a, <4 x double> %b) {
; CHECK-LABEL: define <4 x double> @fadd_v4f64_poison_idx(
; CHECK-SAME: <4 x double> [[A:%.*]], <4 x double> [[B:%.*]]) #[[ATTR0]] {
; CHECK-NEXT:    [[TMP1:%.*]] = shufflevector <4 x double> [[A]], <4 x double> poison, <4 x i32> <i32 2, i32 3, i32 0, i32 poison>
; CHECK-NEXT:    [[TMP2:%.*]] = shufflevector <4 x double> [[B]], <4 x double> poison, <4 x i32> <i32 0, i32 1, i32 0, i32 poison>
; CHECK-NEXT:    [[POST:%.*]] = fadd <4 x double> [[TMP1]], [[TMP2]]
; CHECK-NEXT:    ret <4 x double> [[POST]]
;
  %a1 = shufflevector <4 x double> %a, <4 x double> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
  %b1 = shufflevector <4 x double> %b, <4 x double> poison, <4 x i32> <i32 1, i32 0, i32 1, i32 0>
  %op = fadd <4 x double> %a1, %b1
  %post = shufflevector <4 x double> %op, <4 x double> poison, <4 x i32> <i32 1, i32 0, i32 3, i32 4>
  ret <4 x double> %post
}

define <4 x double> @fadd_v4f64_mixed_types(<4 x double> %a, <2 x double> %b) {
; CHECK-LABEL: define <4 x double> @fadd_v4f64_mixed_types(
; CHECK-SAME: <4 x double> [[A:%.*]], <2 x double> [[B:%.*]]) #[[ATTR0]] {
; CHECK-NEXT:    [[TMP1:%.*]] = shufflevector <4 x double> [[A]], <4 x double> poison, <4 x i32> <i32 2, i32 3, i32 0, i32 1>
; CHECK-NEXT:    [[TMP2:%.*]] = shufflevector <2 x double> [[B]], <2 x double> poison, <4 x i32> <i32 0, i32 1, i32 0, i32 1>
; CHECK-NEXT:    [[POST:%.*]] = fadd <4 x double> [[TMP1]], [[TMP2]]
; CHECK-NEXT:    ret <4 x double> [[POST]]
;
  %a1 = shufflevector <4 x double> %a, <4 x double> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
  %b1 = shufflevector <2 x double> %b, <2 x double> poison, <4 x i32> <i32 1, i32 0, i32 1, i32 0>
  %op = fadd <4 x double> %a1, %b1
  %post = shufflevector <4 x double> %op, <4 x double> poison, <4 x i32> <i32 1, i32 0, i32 3, i32 2>
  ret <4 x double> %post
}

define <4 x float> @fadd_v4f32_mixed_types(<4 x float> %a0) {
; CHECK-LABEL: define <4 x float> @fadd_v4f32_mixed_types(
; CHECK-SAME: <4 x float> [[A0:%.*]]) #[[ATTR0]] {
; CHECK-NEXT:    [[TMP1:%.*]] = shufflevector <4 x float> [[A0]], <4 x float> zeroinitializer, <4 x i32> <i32 1, i32 5, i32 poison, i32 poison>
; CHECK-NEXT:    [[POST:%.*]] = fmul <4 x float> [[TMP1]], <float 0.000000e+00, float 0.000000e+00, float undef, float undef>
; CHECK-NEXT:    ret <4 x float> [[POST]]
;
  %pre = shufflevector <4 x float> %a0, <4 x float> zeroinitializer, <2 x i32> <i32 1, i32 5>
  %op = fmul <2 x float> %pre, zeroinitializer
  %post = shufflevector <2 x float> %op, <2 x float> poison, <4 x i32> <i32 0, i32 1, i32 poison, i32 poison>
  ret <4 x float> %post
}

; Negative test - multiple use of fadd
define <4 x double> @fadd_v4f64_multiuse_op(<4 x double> %a, <4 x double> %b) {
; CHECK-LABEL: define <4 x double> @fadd_v4f64_multiuse_op(
; CHECK-SAME: <4 x double> [[A:%.*]], <4 x double> [[B:%.*]]) #[[ATTR0]] {
; CHECK-NEXT:    [[A1:%.*]] = shufflevector <4 x double> [[A]], <4 x double> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
; CHECK-NEXT:    [[B1:%.*]] = shufflevector <4 x double> [[B]], <4 x double> poison, <4 x i32> <i32 1, i32 0, i32 1, i32 0>
; CHECK-NEXT:    [[OP:%.*]] = fadd <4 x double> [[A1]], [[B1]]
; CHECK-NEXT:    [[POST:%.*]] = shufflevector <4 x double> [[OP]], <4 x double> poison, <4 x i32> <i32 1, i32 0, i32 3, i32 2>
; CHECK-NEXT:    call void @use_v4f64(<4 x double> [[OP]])
; CHECK-NEXT:    ret <4 x double> [[POST]]
;
  %a1 = shufflevector <4 x double> %a, <4 x double> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
  %b1 = shufflevector <4 x double> %b, <4 x double> poison, <4 x i32> <i32 1, i32 0, i32 1, i32 0>
  %op = fadd <4 x double> %a1, %b1
  %post = shufflevector <4 x double> %op, <4 x double> poison, <4 x i32> <i32 1, i32 0, i32 3, i32 2>
  call void @use_v4f64(<4 x double> %op)
  ret <4 x double> %post
}

define <4 x double> @fadd_v4f64_multiuse_shuffle(<4 x double> %a, <4 x double> %b) {
; CHECK-LABEL: define <4 x double> @fadd_v4f64_multiuse_shuffle(
; CHECK-SAME: <4 x double> [[A:%.*]], <4 x double> [[B:%.*]]) #[[ATTR0]] {
; CHECK-NEXT:    [[A1:%.*]] = shufflevector <4 x double> [[A]], <4 x double> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
; CHECK-NEXT:    [[TMP1:%.*]] = shufflevector <4 x double> [[A]], <4 x double> poison, <4 x i32> <i32 2, i32 3, i32 0, i32 1>
; CHECK-NEXT:    [[TMP2:%.*]] = shufflevector <4 x double> [[B]], <4 x double> poison, <4 x i32> <i32 0, i32 1, i32 0, i32 1>
; CHECK-NEXT:    [[POST:%.*]] = fadd <4 x double> [[TMP1]], [[TMP2]]
; CHECK-NEXT:    call void @use_v4f64(<4 x double> [[A1]])
; CHECK-NEXT:    ret <4 x double> [[POST]]
;
  %a1 = shufflevector <4 x double> %a, <4 x double> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
  %b1 = shufflevector <4 x double> %b, <4 x double> poison, <4 x i32> <i32 1, i32 0, i32 1, i32 0>
  %op = fadd <4 x double> %a1, %b1
  %post = shufflevector <4 x double> %op, <4 x double> poison, <4 x i32> <i32 1, i32 0, i32 3, i32 2>
  call void @use_v4f64(<4 x double> %a1)
  ret <4 x double> %post
}

define <4 x i32> @sdiv_v4i32(<4 x i32> %a, <4 x i32> %b) {
; CHECK-LABEL: define <4 x i32> @sdiv_v4i32(
; CHECK-SAME: <4 x i32> [[A:%.*]], <4 x i32> [[B:%.*]]) #[[ATTR0]] {
; CHECK-NEXT:    [[TMP1:%.*]] = shufflevector <4 x i32> [[A]], <4 x i32> poison, <4 x i32> <i32 2, i32 3, i32 0, i32 3>
; CHECK-NEXT:    [[TMP2:%.*]] = shufflevector <4 x i32> [[B]], <4 x i32> poison, <4 x i32> <i32 0, i32 1, i32 0, i32 1>
; CHECK-NEXT:    [[POST:%.*]] = sdiv <4 x i32> [[TMP1]], [[TMP2]]
; CHECK-NEXT:    ret <4 x i32> [[POST]]
;
  %a1 = shufflevector <4 x i32> %a, <4 x i32> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
  %b1 = shufflevector <4 x i32> %b, <4 x i32> poison, <4 x i32> <i32 1, i32 0, i32 1, i32 0>
  %op = sdiv <4 x i32> %a1, %b1
  %post = shufflevector <4 x i32> %op, <4 x i32> poison, <4 x i32> <i32 1, i32 0, i32 3, i32 0>
  ret <4 x i32> %post
}

; Negative test - don't introduce poison element into div/rem instruction
define <4 x i32> @sdiv_v4i32_poison(<4 x i32> %a, <4 x i32> %b) {
; CHECK-LABEL: define <4 x i32> @sdiv_v4i32_poison(
; CHECK-SAME: <4 x i32> [[A:%.*]], <4 x i32> [[B:%.*]]) #[[ATTR0]] {
; CHECK-NEXT:    [[A1:%.*]] = shufflevector <4 x i32> [[A]], <4 x i32> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
; CHECK-NEXT:    [[B1:%.*]] = shufflevector <4 x i32> [[B]], <4 x i32> poison, <4 x i32> <i32 1, i32 0, i32 1, i32 0>
; CHECK-NEXT:    [[OP:%.*]] = sdiv <4 x i32> [[A1]], [[B1]]
; CHECK-NEXT:    [[POST:%.*]] = shufflevector <4 x i32> [[OP]], <4 x i32> poison, <4 x i32> <i32 1, i32 0, i32 3, i32 poison>
; CHECK-NEXT:    ret <4 x i32> [[POST]]
;
  %a1 = shufflevector <4 x i32> %a, <4 x i32> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
  %b1 = shufflevector <4 x i32> %b, <4 x i32> poison, <4 x i32> <i32 1, i32 0, i32 1, i32 0>
  %op = sdiv <4 x i32> %a1, %b1
  %post = shufflevector <4 x i32> %op, <4 x i32> poison, <4 x i32> <i32 1, i32 0, i32 3, i32 poison>
  ret <4 x i32> %post
}

; Negative test - don't introduce poison element into div/rem instruction
define <4 x i32> @sdiv_v4i32_poison_idx(<4 x i32> %a, <4 x i32> %b) {
; CHECK-LABEL: define <4 x i32> @sdiv_v4i32_poison_idx(
; CHECK-SAME: <4 x i32> [[A:%.*]], <4 x i32> [[B:%.*]]) #[[ATTR0]] {
; CHECK-NEXT:    [[A1:%.*]] = shufflevector <4 x i32> [[A]], <4 x i32> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
; CHECK-NEXT:    [[B1:%.*]] = shufflevector <4 x i32> [[B]], <4 x i32> poison, <4 x i32> <i32 1, i32 0, i32 1, i32 0>
; CHECK-NEXT:    [[OP:%.*]] = sdiv <4 x i32> [[A1]], [[B1]]
; CHECK-NEXT:    [[POST:%.*]] = shufflevector <4 x i32> [[OP]], <4 x i32> poison, <4 x i32> <i32 1, i32 0, i32 3, i32 4>
; CHECK-NEXT:    ret <4 x i32> [[POST]]
;
  %a1 = shufflevector <4 x i32> %a, <4 x i32> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
  %b1 = shufflevector <4 x i32> %b, <4 x i32> poison, <4 x i32> <i32 1, i32 0, i32 1, i32 0>
  %op = sdiv <4 x i32> %a1, %b1
  %post = shufflevector <4 x i32> %op, <4 x i32> poison, <4 x i32> <i32 1, i32 0, i32 3, i32 4>
  ret <4 x i32> %post
}