; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mtriple=i686-apple-darwin -mcpu=core-avx2 -mattr=+avx2 | FileCheck %s --check-prefix=X32
; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=core-avx2 -mattr=+avx2 | FileCheck %s --check-prefix=X64

declare <4 x float> @llvm.x86.avx2.gather.d.ps(<4 x float>, i8*,
                      <4 x i32>, <4 x float>, i8) nounwind readonly

define <4 x float> @test_x86_avx2_gather_d_ps(i8* %a1, <4 x i32> %idx, <4 x float> %mask) {
; X32-LABEL: test_x86_avx2_gather_d_ps:
; X32:       ## BB#0:
; X32-NEXT:    movl {{[0-9]+}}(%esp), %eax
; X32-NEXT:    vgatherdps %xmm1, (%eax,%xmm0,2), %xmm2
; X32-NEXT:    vmovaps %xmm2, %xmm0
; X32-NEXT:    retl
;
; X64-LABEL: test_x86_avx2_gather_d_ps:
; X64:       ## BB#0:
; X64-NEXT:    vgatherdps %xmm1, (%rdi,%xmm0,2), %xmm2
; X64-NEXT:    vmovaps %xmm2, %xmm0
; X64-NEXT:    retq
  %res = call <4 x float> @llvm.x86.avx2.gather.d.ps(<4 x float> undef,
                            i8* %a1, <4 x i32> %idx, <4 x float> %mask, i8 2) ;
  ret <4 x float> %res
}

declare <2 x double> @llvm.x86.avx2.gather.d.pd(<2 x double>, i8*,
                      <4 x i32>, <2 x double>, i8) nounwind readonly

define <2 x double> @test_x86_avx2_gather_d_pd(i8* %a1, <4 x i32> %idx, <2 x double> %mask) {
; X32-LABEL: test_x86_avx2_gather_d_pd:
; X32:       ## BB#0:
; X32-NEXT:    movl {{[0-9]+}}(%esp), %eax
; X32-NEXT:    vgatherdpd %xmm1, (%eax,%xmm0,2), %xmm2
; X32-NEXT:    vmovapd %xmm2, %xmm0
; X32-NEXT:    retl
;
; X64-LABEL: test_x86_avx2_gather_d_pd:
; X64:       ## BB#0:
; X64-NEXT:    vgatherdpd %xmm1, (%rdi,%xmm0,2), %xmm2
; X64-NEXT:    vmovapd %xmm2, %xmm0
; X64-NEXT:    retq
  %res = call <2 x double> @llvm.x86.avx2.gather.d.pd(<2 x double> undef,
                            i8* %a1, <4 x i32> %idx, <2 x double> %mask, i8 2) ;
  ret <2 x double> %res
}

declare <8 x float> @llvm.x86.avx2.gather.d.ps.256(<8 x float>, i8*,
                      <8 x i32>, <8 x float>, i8) nounwind readonly

define <8 x float> @test_x86_avx2_gather_d_ps_256(i8* %a1, <8 x i32> %idx, <8 x float> %mask) {
; X32-LABEL: test_x86_avx2_gather_d_ps_256:
; X32:       ## BB#0:
; X32-NEXT:    movl {{[0-9]+}}(%esp), %eax
; X32-NEXT:    vgatherdps %ymm1, (%eax,%ymm0,4), %ymm2
; X32-NEXT:    vmovaps %ymm2, %ymm0
; X32-NEXT:    retl
;
; X64-LABEL: test_x86_avx2_gather_d_ps_256:
; X64:       ## BB#0:
; X64-NEXT:    vgatherdps %ymm1, (%rdi,%ymm0,4), %ymm2
; X64-NEXT:    vmovaps %ymm2, %ymm0
; X64-NEXT:    retq
  %res = call <8 x float> @llvm.x86.avx2.gather.d.ps.256(<8 x float> undef,
                            i8* %a1, <8 x i32> %idx, <8 x float> %mask, i8 4) ;
  ret <8 x float> %res
}

declare <4 x double> @llvm.x86.avx2.gather.d.pd.256(<4 x double>, i8*,
                      <4 x i32>, <4 x double>, i8) nounwind readonly

define <4 x double> @test_x86_avx2_gather_d_pd_256(i8* %a1, <4 x i32> %idx, <4 x double> %mask) {
; X32-LABEL: test_x86_avx2_gather_d_pd_256:
; X32:       ## BB#0:
; X32-NEXT:    movl {{[0-9]+}}(%esp), %eax
; X32-NEXT:    vgatherdpd %ymm1, (%eax,%xmm0,8), %ymm2
; X32-NEXT:    vmovapd %ymm2, %ymm0
; X32-NEXT:    retl
;
; X64-LABEL: test_x86_avx2_gather_d_pd_256:
; X64:       ## BB#0:
; X64-NEXT:    vgatherdpd %ymm1, (%rdi,%xmm0,8), %ymm2
; X64-NEXT:    vmovapd %ymm2, %ymm0
; X64-NEXT:    retq
  %res = call <4 x double> @llvm.x86.avx2.gather.d.pd.256(<4 x double> undef,
                            i8* %a1, <4 x i32> %idx, <4 x double> %mask, i8 8) ;
  ret <4 x double> %res
}