llvm/test/CodeGen/AArch64/sve-masked-gather.ll

; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -mtriple=aarch64--linux-gnu -mattr=+sve < %s | FileCheck %s

define <vscale x 2 x i64> @masked_gather_nxv2i8(<vscale x 2 x ptr> %ptrs, <vscale x 2 x i1> %mask) {
; CHECK-LABEL: masked_gather_nxv2i8:
; CHECK:       // %bb.0:
; CHECK-NEXT:    ld1b { z0.d }, p0/z, [z0.d]
; CHECK-NEXT:    ret
  %vals = call <vscale x 2 x i8> @llvm.masked.gather.nxv2i8(<vscale x 2 x ptr> %ptrs, i32 1, <vscale x 2 x i1> %mask, <vscale x 2 x i8> poison)
  %vals.zext = zext <vscale x 2 x i8> %vals to <vscale x 2 x i64>
  ret <vscale x 2 x i64> %vals.zext
}

define <vscale x 2 x i64> @masked_gather_nxv2i16(<vscale x 2 x ptr> %ptrs, <vscale x 2 x i1> %mask) {
; CHECK-LABEL: masked_gather_nxv2i16:
; CHECK:       // %bb.0:
; CHECK-NEXT:    ld1h { z0.d }, p0/z, [z0.d]
; CHECK-NEXT:    ret
  %vals = call <vscale x 2 x i16> @llvm.masked.gather.nxv2i16(<vscale x 2 x ptr> %ptrs, i32 2, <vscale x 2 x i1> %mask, <vscale x 2 x i16> poison)
  %vals.zext = zext <vscale x 2 x i16> %vals to <vscale x 2 x i64>
  ret <vscale x 2 x i64> %vals.zext
}

define <vscale x 2 x i64> @masked_gather_nxv2i32(<vscale x 2 x ptr> %ptrs, <vscale x 2 x i1> %mask) {
; CHECK-LABEL: masked_gather_nxv2i32:
; CHECK:       // %bb.0:
; CHECK-NEXT:    ld1w { z0.d }, p0/z, [z0.d]
; CHECK-NEXT:    ret
  %vals = call <vscale x 2 x i32> @llvm.masked.gather.nxv2i32(<vscale x 2 x ptr> %ptrs, i32 4, <vscale x 2 x i1> %mask, <vscale x 2 x i32> poison)
  %vals.zext = zext <vscale x 2 x i32> %vals to <vscale x 2 x i64>
  ret <vscale x 2 x i64> %vals.zext
}

define <vscale x 2 x i64> @masked_gather_nxv2i64(<vscale x 2 x ptr> %ptrs, <vscale x 2 x i1> %mask) {
; CHECK-LABEL: masked_gather_nxv2i64:
; CHECK:       // %bb.0:
; CHECK-NEXT:    ld1d { z0.d }, p0/z, [z0.d]
; CHECK-NEXT:    ret
  %vals = call <vscale x 2 x i64> @llvm.masked.gather.nxv2i64(<vscale x 2 x ptr> %ptrs, i32 8, <vscale x 2 x i1> %mask, <vscale x 2 x i64> poison)
  ret <vscale x 2 x i64> %vals
}

define <vscale x 2 x half> @masked_gather_nxv2f16(<vscale x 2 x ptr> %ptrs, <vscale x 2 x i1> %mask) {
; CHECK-LABEL: masked_gather_nxv2f16:
; CHECK:       // %bb.0:
; CHECK-NEXT:    ld1h { z0.d }, p0/z, [z0.d]
; CHECK-NEXT:    ret
  %vals = call <vscale x 2 x half> @llvm.masked.gather.nxv2f16(<vscale x 2 x ptr> %ptrs, i32 2, <vscale x 2 x i1> %mask, <vscale x 2 x half> poison)
  ret <vscale x 2 x half> %vals
}

define <vscale x 2 x bfloat> @masked_gather_nxv2bf16(<vscale x 2 x ptr> %ptrs, <vscale x 2 x i1> %mask) #0 {
; CHECK-LABEL: masked_gather_nxv2bf16:
; CHECK:       // %bb.0:
; CHECK-NEXT:    ld1h { z0.d }, p0/z, [z0.d]
; CHECK-NEXT:    ret
  %vals = call <vscale x 2 x bfloat> @llvm.masked.gather.nxv2bf16(<vscale x 2 x ptr> %ptrs, i32 2, <vscale x 2 x i1> %mask, <vscale x 2 x bfloat> poison)
  ret <vscale x 2 x bfloat> %vals
}

define <vscale x 2 x float> @masked_gather_nxv2f32(<vscale x 2 x ptr> %ptrs, <vscale x 2 x i1> %mask) {
; CHECK-LABEL: masked_gather_nxv2f32:
; CHECK:       // %bb.0:
; CHECK-NEXT:    ld1w { z0.d }, p0/z, [z0.d]
; CHECK-NEXT:    ret
  %vals = call <vscale x 2 x float> @llvm.masked.gather.nxv2f32(<vscale x 2 x ptr> %ptrs, i32 4, <vscale x 2 x i1> %mask, <vscale x 2 x float> poison)
  ret <vscale x 2 x float> %vals
}

define <vscale x 2 x double> @masked_gather_nxv2f64(<vscale x 2 x ptr> %ptrs, <vscale x 2 x i1> %mask) {
; CHECK-LABEL: masked_gather_nxv2f64:
; CHECK:       // %bb.0:
; CHECK-NEXT:    ld1d { z0.d }, p0/z, [z0.d]
; CHECK-NEXT:    ret
  %vals = call <vscale x 2 x double> @llvm.masked.gather.nxv2f64(<vscale x 2 x ptr> %ptrs, i32 8, <vscale x 2 x i1> %mask, <vscale x 2 x double> poison)
  ret <vscale x 2 x double> %vals
}

define <vscale x 2 x i64> @masked_sgather_nxv2i8(<vscale x 2 x ptr> %ptrs, <vscale x 2 x i1> %mask) {
; CHECK-LABEL: masked_sgather_nxv2i8:
; CHECK:       // %bb.0:
; CHECK-NEXT:    ld1sb { z0.d }, p0/z, [z0.d]
; CHECK-NEXT:    ret
  %vals = call <vscale x 2 x i8> @llvm.masked.gather.nxv2i8(<vscale x 2 x ptr> %ptrs, i32 1, <vscale x 2 x i1> %mask, <vscale x 2 x i8> poison)
  %vals.sext = sext <vscale x 2 x i8> %vals to <vscale x 2 x i64>
  ret <vscale x 2 x i64> %vals.sext
}

define <vscale x 2 x i64> @masked_sgather_nxv2i16(<vscale x 2 x ptr> %ptrs, <vscale x 2 x i1> %mask) {
; CHECK-LABEL: masked_sgather_nxv2i16:
; CHECK:       // %bb.0:
; CHECK-NEXT:    ld1sh { z0.d }, p0/z, [z0.d]
; CHECK-NEXT:    ret
  %vals = call <vscale x 2 x i16> @llvm.masked.gather.nxv2i16(<vscale x 2 x ptr> %ptrs, i32 2, <vscale x 2 x i1> %mask, <vscale x 2 x i16> poison)
  %vals.sext = sext <vscale x 2 x i16> %vals to <vscale x 2 x i64>
  ret <vscale x 2 x i64> %vals.sext
}

define <vscale x 16 x i8> @masked_gather_nxv1i8(<vscale x 2 x ptr> %wide.ptrs, <vscale x 1 x i1> %mask) {
; CHECK-LABEL: masked_gather_nxv1i8:
; CHECK:       // %bb.0:
; CHECK-NEXT:    pfalse p1.b
; CHECK-NEXT:    uzp1 p0.d, p0.d, p1.d
; CHECK-NEXT:    ld1b { z0.d }, p0/z, [z0.d]
; CHECK-NEXT:    uzp1 z0.s, z0.s, z0.s
; CHECK-NEXT:    uzp1 z0.h, z0.h, z0.h
; CHECK-NEXT:    uzp1 z0.b, z0.b, z0.b
; CHECK-NEXT:    ret
  %ptrs = call <vscale x 1 x ptr> @llvm.vector.extract.nxv1p0.nxv2p0(<vscale x 2 x ptr> %wide.ptrs, i64 0)
  %r = call <vscale x 1 x i8> @llvm.masked.gather.nxv1i8(<vscale x 1 x ptr> align 1 %ptrs, <vscale x 1 x i1> %mask, <vscale x 1 x i8> poison)
  %r.legal = call <vscale x 16 x i8> @llvm.vector.insert.nxv16i8.nxv1i8(<vscale x 16 x i8> poison, <vscale x 1 x i8> %r, i64 0)
  ret <vscale x 16 x i8> %r.legal
}

define <vscale x 8 x i16> @masked_gather_nxv1i16(<vscale x 2 x ptr> %wide.ptrs, <vscale x 1 x i1> %mask) {
; CHECK-LABEL: masked_gather_nxv1i16:
; CHECK:       // %bb.0:
; CHECK-NEXT:    pfalse p1.b
; CHECK-NEXT:    uzp1 p0.d, p0.d, p1.d
; CHECK-NEXT:    ld1h { z0.d }, p0/z, [z0.d]
; CHECK-NEXT:    uzp1 z0.s, z0.s, z0.s
; CHECK-NEXT:    uzp1 z0.h, z0.h, z0.h
; CHECK-NEXT:    ret
  %ptrs = call <vscale x 1 x ptr> @llvm.vector.extract.nxv1p0.nxv2p0(<vscale x 2 x ptr> %wide.ptrs, i64 0)
  %r = call <vscale x 1 x i16> @llvm.masked.gather.nxv1i16(<vscale x 1 x ptr> align 2 %ptrs, <vscale x 1 x i1> %mask, <vscale x 1 x i16> poison)
  %r.legal = call <vscale x 8 x i16> @llvm.vector.insert.nxv8i16.nxv1i16(<vscale x 8 x i16> poison, <vscale x 1 x i16> %r, i64 0)
  ret <vscale x 8 x i16> %r.legal
}

define <vscale x 4 x i32> @masked_gather_nxv1i32(<vscale x 2 x ptr> %wide.ptrs, <vscale x 1 x i1> %mask) {
; CHECK-LABEL: masked_gather_nxv1i32:
; CHECK:       // %bb.0:
; CHECK-NEXT:    pfalse p1.b
; CHECK-NEXT:    uzp1 p0.d, p0.d, p1.d
; CHECK-NEXT:    ld1w { z0.d }, p0/z, [z0.d]
; CHECK-NEXT:    uzp1 z0.s, z0.s, z0.s
; CHECK-NEXT:    ret
  %ptrs = call <vscale x 1 x ptr> @llvm.vector.extract.nxv1p0.nxv2p0(<vscale x 2 x ptr> %wide.ptrs, i64 0)
  %r = call <vscale x 1 x i32> @llvm.masked.gather.nxv1i32(<vscale x 1 x ptr> align 4 %ptrs, <vscale x 1 x i1> %mask, <vscale x 1 x i32> poison)
  %r.legal = call <vscale x 4 x i32> @llvm.vector.insert.nxv4i32.nxv1i32(<vscale x 4 x i32> poison, <vscale x 1 x i32> %r, i64 0)
  ret <vscale x 4 x i32> %r.legal
}

define <vscale x 2 x i64> @masked_gather_nxv1i64(<vscale x 2 x ptr> %wide.ptrs, <vscale x 1 x i1> %mask) {
; CHECK-LABEL: masked_gather_nxv1i64:
; CHECK:       // %bb.0:
; CHECK-NEXT:    pfalse p1.b
; CHECK-NEXT:    uzp1 p0.d, p0.d, p1.d
; CHECK-NEXT:    ld1d { z0.d }, p0/z, [z0.d]
; CHECK-NEXT:    ret
  %ptrs = call <vscale x 1 x ptr> @llvm.vector.extract.nxv1p0.nxv2p0(<vscale x 2 x ptr> %wide.ptrs, i64 0)
  %r = call <vscale x 1 x i64> @llvm.masked.gather.nxv1i64(<vscale x 1 x ptr> align 8 %ptrs, <vscale x 1 x i1> %mask, <vscale x 1 x i64> poison)
  %r.legal = call <vscale x 2 x i64> @llvm.vector.insert.nxv2i64.nxv1i64(<vscale x 2 x i64> poison, <vscale x 1 x i64> %r, i64 0)
  ret <vscale x 2 x i64> %r.legal
}

define <vscale x 2 x i64> @masked_sgather_nxv2i32(<vscale x 2 x ptr> %ptrs, <vscale x 2 x i1> %mask) {
; CHECK-LABEL: masked_sgather_nxv2i32:
; CHECK:       // %bb.0:
; CHECK-NEXT:    ld1sw { z0.d }, p0/z, [z0.d]
; CHECK-NEXT:    ret
  %vals = call <vscale x 2 x i32> @llvm.masked.gather.nxv2i32(<vscale x 2 x ptr> %ptrs, i32 4, <vscale x 2 x i1> %mask, <vscale x 2 x i32> poison)
  %vals.sext = sext <vscale x 2 x i32> %vals to <vscale x 2 x i64>
  ret <vscale x 2 x i64> %vals.sext
}

define <vscale x 2 x i64> @masked_gather_passthru(<vscale x 2 x ptr> %ptrs, <vscale x 2 x i1> %mask, <vscale x 2 x i32> %passthru) {
; CHECK-LABEL: masked_gather_passthru:
; CHECK:       // %bb.0:
; CHECK-NEXT:    ld1sw { z0.d }, p0/z, [z0.d]
; CHECK-NEXT:    sel z0.d, p0, z0.d, z1.d
; CHECK-NEXT:    ret
  %vals = call <vscale x 2 x i32> @llvm.masked.gather.nxv2i32(<vscale x 2 x ptr> %ptrs, i32 4, <vscale x 2 x i1> %mask, <vscale x 2 x i32> %passthru)
  %vals.sext = sext <vscale x 2 x i32> %vals to <vscale x 2 x i64>
  ret <vscale x 2 x i64> %vals.sext
}

define <vscale x 2 x i64> @masked_gather_passthru_0(<vscale x 2 x ptr> %ptrs, <vscale x 2 x i1> %mask) {
; CHECK-LABEL: masked_gather_passthru_0:
; CHECK:       // %bb.0:
; CHECK-NEXT:    ld1sw { z0.d }, p0/z, [z0.d]
; CHECK-NEXT:    ret
  %vals = call <vscale x 2 x i32> @llvm.masked.gather.nxv2i32(<vscale x 2 x ptr> %ptrs, i32 4, <vscale x 2 x i1> %mask, <vscale x 2 x i32> zeroinitializer)
  %vals.sext = sext <vscale x 2 x i32> %vals to <vscale x 2 x i64>
  ret <vscale x 2 x i64> %vals.sext
}

%i64_x3 = type { i64, i64, i64}
define <vscale x 2 x i64> @masked_gather_non_power_of_two_based_scaling(ptr %base, <vscale x 2 x i64> %offsets, <vscale x 2 x i1> %mask) {
; CHECK-LABEL: masked_gather_non_power_of_two_based_scaling:
; CHECK:       // %bb.0:
; CHECK-NEXT:    mul z0.d, z0.d, #24
; CHECK-NEXT:    ld1d { z0.d }, p0/z, [x0, z0.d]
; CHECK-NEXT:    ret
  %ptrs = getelementptr inbounds %i64_x3, ptr %base, <vscale x 2 x i64> %offsets
  %vals = call <vscale x 2 x i64> @llvm.masked.gather.nxv2i64(<vscale x 2 x ptr> %ptrs, i32 8, <vscale x 2 x i1> %mask, <vscale x 2 x i64> poison)
  ret <vscale x 2 x i64> %vals
}

%i64_x4 = type { i64, i64, i64, i64}
define <vscale x 2 x i64> @masked_gather_non_element_type_based_scaling(ptr %base, <vscale x 2 x i64> %offsets, <vscale x 2 x i1> %mask) {
; CHECK-LABEL: masked_gather_non_element_type_based_scaling:
; CHECK:       // %bb.0:
; CHECK-NEXT:    lsl z0.d, z0.d, #5
; CHECK-NEXT:    ld1d { z0.d }, p0/z, [x0, z0.d]
; CHECK-NEXT:    ret
  %ptrs = getelementptr inbounds %i64_x4, ptr %base, <vscale x 2 x i64> %offsets
  %vals = call <vscale x 2 x i64> @llvm.masked.gather.nxv2i64(<vscale x 2 x ptr> %ptrs, i32 8, <vscale x 2 x i1> %mask, <vscale x 2 x i64> poison)
  ret <vscale x 2 x i64> %vals
}

declare <vscale x 2 x i8> @llvm.masked.gather.nxv2i8(<vscale x 2 x ptr>, i32, <vscale x 2 x i1>, <vscale x 2 x i8>)
declare <vscale x 2 x i16> @llvm.masked.gather.nxv2i16(<vscale x 2 x ptr>, i32, <vscale x 2 x i1>, <vscale x 2 x i16>)
declare <vscale x 2 x i32> @llvm.masked.gather.nxv2i32(<vscale x 2 x ptr>, i32, <vscale x 2 x i1>, <vscale x 2 x i32>)
declare <vscale x 2 x i64> @llvm.masked.gather.nxv2i64(<vscale x 2 x ptr>, i32, <vscale x 2 x i1>, <vscale x 2 x i64>)
declare <vscale x 2 x half> @llvm.masked.gather.nxv2f16(<vscale x 2 x ptr>, i32, <vscale x 2 x i1>, <vscale x 2 x half>)
declare <vscale x 2 x bfloat> @llvm.masked.gather.nxv2bf16(<vscale x 2 x ptr>, i32, <vscale x 2 x i1>, <vscale x 2 x bfloat>)
declare <vscale x 2 x float> @llvm.masked.gather.nxv2f32(<vscale x 2 x ptr>, i32, <vscale x 2 x i1>, <vscale x 2 x float>)
declare <vscale x 2 x double> @llvm.masked.gather.nxv2f64(<vscale x 2 x ptr>, i32, <vscale x 2 x i1>, <vscale x 2 x double>)
attributes #0 = { "target-features"="+sve,+bf16" }