; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 2
; RUN: llc -mtriple=aarch64-linux-gnu -mattr=+sme2 -mattr=+fp8 -force-streaming -verify-machineinstrs < %s | FileCheck %s

; FSCALE (Single, x2)

define { <vscale x 8 x half>, <vscale x 8 x half> } @multi_vec_scale_single_x2_half( <vscale x 8 x half> %zdn1, <vscale x 8 x half> %zdn2, <vscale x 8 x i16> %zm) {
; CHECK-LABEL: multi_vec_scale_single_x2_half:
; CHECK:       // %bb.0:
; CHECK-NEXT:    // kill: def $z1 killed $z1 killed $z0_z1 def $z0_z1
; CHECK-NEXT:    // kill: def $z0 killed $z0 killed $z0_z1 def $z0_z1
; CHECK-NEXT:    fscale { z0.h, z1.h }, { z0.h, z1.h }, z2.h
; CHECK-NEXT:    ret
  %res = call { <vscale x 8 x half>, <vscale x 8 x half> } @llvm.aarch64.sme.fp8.scale.single.x2.nxv8f16(<vscale x 8 x half> %zdn1, <vscale x 8 x half> %zdn2, <vscale x 8 x i16> %zm)
  ret { <vscale x 8 x half>, <vscale x 8 x half> } %res
}

define { <vscale x 4 x float>, <vscale x 4 x float> } @multi_vec_scale_single_x2_float( <vscale x 4 x float> %zdn1, <vscale x 4 x float> %zdn2, <vscale x 4 x i32> %zm) {
; CHECK-LABEL: multi_vec_scale_single_x2_float:
; CHECK:       // %bb.0:
; CHECK-NEXT:    // kill: def $z1 killed $z1 killed $z0_z1 def $z0_z1
; CHECK-NEXT:    // kill: def $z0 killed $z0 killed $z0_z1 def $z0_z1
; CHECK-NEXT:    fscale { z0.s, z1.s }, { z0.s, z1.s }, z2.s
; CHECK-NEXT:    ret
  %res = call { <vscale x 4 x float>, <vscale x 4 x float> } @llvm.aarch64.sme.fp8.scale.single.x2.nxv4f32(<vscale x 4 x float> %zdn1, <vscale x 4 x float> %zdn2, <vscale x 4 x i32> %zm)
  ret { <vscale x 4 x float>, <vscale x 4 x float> } %res
}

define { <vscale x 2 x double>, <vscale x 2 x double> } @multi_vec_scale_single_x2_double( <vscale x 2 x double> %zdn1, <vscale x 2 x double> %zdn2, <vscale x 2 x i64> %zm) {
; CHECK-LABEL: multi_vec_scale_single_x2_double:
; CHECK:       // %bb.0:
; CHECK-NEXT:    // kill: def $z1 killed $z1 killed $z0_z1 def $z0_z1
; CHECK-NEXT:    // kill: def $z0 killed $z0 killed $z0_z1 def $z0_z1
; CHECK-NEXT:    fscale { z0.d, z1.d }, { z0.d, z1.d }, z2.d
; CHECK-NEXT:    ret
  %res = call { <vscale x 2 x double>, <vscale x 2 x double> } @llvm.aarch64.sme.fp8.scale.single.x2.nxv2f64(<vscale x 2 x double> %zdn1, <vscale x 2 x double> %zdn2, <vscale x 2 x i64> %zm)
  ret { <vscale x 2 x double>, <vscale x 2 x double> } %res
}

; FSCALE (Single, x4)

define { <vscale x 8 x half>, <vscale x 8 x half>, <vscale x 8 x half>, <vscale x 8 x half> } @multi_vec_scale_single_x4_half( <vscale x 8 x half> %zdn1, <vscale x 8 x half> %zdn2, <vscale x 8 x half> %zdn3, <vscale x 8 x half> %zdn4, <vscale x 8 x i16> %zm) {
; CHECK-LABEL: multi_vec_scale_single_x4_half:
; CHECK:       // %bb.0:
; CHECK-NEXT:    // kill: def $z3 killed $z3 killed $z0_z1_z2_z3 def $z0_z1_z2_z3
; CHECK-NEXT:    // kill: def $z2 killed $z2 killed $z0_z1_z2_z3 def $z0_z1_z2_z3
; CHECK-NEXT:    // kill: def $z1 killed $z1 killed $z0_z1_z2_z3 def $z0_z1_z2_z3
; CHECK-NEXT:    // kill: def $z0 killed $z0 killed $z0_z1_z2_z3 def $z0_z1_z2_z3
; CHECK-NEXT:    fscale { z0.h - z3.h }, { z0.h - z3.h }, z4.h
; CHECK-NEXT:    ret
  %res = call { <vscale x 8 x half>, <vscale x 8 x half>, <vscale x 8 x half>, <vscale x 8 x half> } @llvm.aarch64.sme.fp8.scale.single.x4.nxv8f16(<vscale x 8 x half> %zdn1, <vscale x 8 x half> %zdn2, <vscale x 8 x half> %zdn3, <vscale x 8 x half> %zdn4, <vscale x 8 x i16> %zm)
  ret { <vscale x 8 x half>, <vscale x 8 x half>, <vscale x 8 x half>, <vscale x 8 x half> } %res
}

define { <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float> } @multi_vec_scale_single_x4_float( <vscale x 4 x float> %zdn1, <vscale x 4 x float> %zdn2,  <vscale x 4 x float> %zdn3, <vscale x 4 x float> %zdn4, <vscale x 4 x i32> %zm) {
; CHECK-LABEL: multi_vec_scale_single_x4_float:
; CHECK:       // %bb.0:
; CHECK-NEXT:    // kill: def $z3 killed $z3 killed $z0_z1_z2_z3 def $z0_z1_z2_z3
; CHECK-NEXT:    // kill: def $z2 killed $z2 killed $z0_z1_z2_z3 def $z0_z1_z2_z3
; CHECK-NEXT:    // kill: def $z1 killed $z1 killed $z0_z1_z2_z3 def $z0_z1_z2_z3
; CHECK-NEXT:    // kill: def $z0 killed $z0 killed $z0_z1_z2_z3 def $z0_z1_z2_z3
; CHECK-NEXT:    fscale { z0.s - z3.s }, { z0.s - z3.s }, z4.s
; CHECK-NEXT:    ret
  %res = call { <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float> } @llvm.aarch64.sme.fp8.scale.single.x4.nxv4f32(<vscale x 4 x float> %zdn1, <vscale x 4 x float> %zdn2, <vscale x 4 x float> %zdn3, <vscale x 4 x float> %zdn4, <vscale x 4 x i32> %zm)
  ret { <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float> } %res
}

define { <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double> } @multi_vec_scale_single_x4_double( <vscale x 2 x double> %zdn1, <vscale x 2 x double> %zdn2, <vscale x 2 x double> %zdn3, <vscale x 2 x double> %zdn4, <vscale x 2 x i64> %zm) {
; CHECK-LABEL: multi_vec_scale_single_x4_double:
; CHECK:       // %bb.0:
; CHECK-NEXT:    // kill: def $z3 killed $z3 killed $z0_z1_z2_z3 def $z0_z1_z2_z3
; CHECK-NEXT:    // kill: def $z2 killed $z2 killed $z0_z1_z2_z3 def $z0_z1_z2_z3
; CHECK-NEXT:    // kill: def $z1 killed $z1 killed $z0_z1_z2_z3 def $z0_z1_z2_z3
; CHECK-NEXT:    // kill: def $z0 killed $z0 killed $z0_z1_z2_z3 def $z0_z1_z2_z3
; CHECK-NEXT:    fscale { z0.d - z3.d }, { z0.d - z3.d }, z4.d
; CHECK-NEXT:    ret
  %res = call { <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double> } @llvm.aarch64.sme.fp8.scale.single.x4.nxv2f64(<vscale x 2 x double> %zdn1, <vscale x 2 x double> %zdn2, <vscale x 2 x double> %zdn3, <vscale x 2 x double> %zdn4, <vscale x 2 x i64> %zm)
  ret { <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double> } %res
}

; FSCALE (Multi, x2)
define { <vscale x 8 x half>, <vscale x 8 x half> } @multi_vec_scale_x2_half( <vscale x 8 x half> %zdn1, <vscale x 8 x half> %zdn2, <vscale x 8 x i16> %zm1, <vscale x 8 x i16> %zm2) {
; CHECK-LABEL: multi_vec_scale_x2_half:
; CHECK:       // %bb.0:
; CHECK-NEXT:    // kill: def $z1 killed $z1 killed $z0_z1 def $z0_z1
; CHECK-NEXT:    // kill: def $z3 killed $z3 killed $z2_z3 def $z2_z3
; CHECK-NEXT:    // kill: def $z0 killed $z0 killed $z0_z1 def $z0_z1
; CHECK-NEXT:    // kill: def $z2 killed $z2 killed $z2_z3 def $z2_z3
; CHECK-NEXT:    fscale { z0.h, z1.h }, { z0.h, z1.h }, { z2.h, z3.h }
; CHECK-NEXT:    ret
  %res = call { <vscale x 8 x half>, <vscale x 8 x half> } @llvm.aarch64.sme.fp8.scale.x2.nxv8f16(<vscale x 8 x half> %zdn1, <vscale x 8 x half> %zdn2, <vscale x 8 x i16> %zm1, <vscale x 8 x i16> %zm2)
  ret { <vscale x 8 x half>, <vscale x 8 x half> } %res
}

define { <vscale x 4 x float>, <vscale x 4 x float> } @multi_vec_scale_x2_float( <vscale x 4 x float> %zdn1, <vscale x 4 x float> %zdn2, <vscale x 4 x i32> %zm1, <vscale x 4 x i32> %zm2 ) {
; CHECK-LABEL: multi_vec_scale_x2_float:
; CHECK:       // %bb.0:
; CHECK-NEXT:    // kill: def $z1 killed $z1 killed $z0_z1 def $z0_z1
; CHECK-NEXT:    // kill: def $z3 killed $z3 killed $z2_z3 def $z2_z3
; CHECK-NEXT:    // kill: def $z0 killed $z0 killed $z0_z1 def $z0_z1
; CHECK-NEXT:    // kill: def $z2 killed $z2 killed $z2_z3 def $z2_z3
; CHECK-NEXT:    fscale { z0.s, z1.s }, { z0.s, z1.s }, { z2.s, z3.s }
; CHECK-NEXT:    ret
  %res = call { <vscale x 4 x float>, <vscale x 4 x float> } @llvm.aarch64.sme.fp8.scale.x2.nxv4f32(<vscale x 4 x float> %zdn1, <vscale x 4 x float> %zdn2, <vscale x 4 x i32> %zm1, <vscale x 4 x i32> %zm2)
  ret { <vscale x 4 x float>, <vscale x 4 x float> } %res
}

define { <vscale x 2 x double>, <vscale x 2 x double> } @multi_vec_scale_x2_double( <vscale x 2 x double> %zdn1, <vscale x 2 x double> %zdn2, <vscale x 2 x i64> %zm1, <vscale x 2 x i64> %zm2) {
; CHECK-LABEL: multi_vec_scale_x2_double:
; CHECK:       // %bb.0:
; CHECK-NEXT:    // kill: def $z1 killed $z1 killed $z0_z1 def $z0_z1
; CHECK-NEXT:    // kill: def $z3 killed $z3 killed $z2_z3 def $z2_z3
; CHECK-NEXT:    // kill: def $z0 killed $z0 killed $z0_z1 def $z0_z1
; CHECK-NEXT:    // kill: def $z2 killed $z2 killed $z2_z3 def $z2_z3
; CHECK-NEXT:    fscale { z0.d, z1.d }, { z0.d, z1.d }, { z2.d, z3.d }
; CHECK-NEXT:    ret
  %res = call { <vscale x 2 x double>, <vscale x 2 x double> } @llvm.aarch64.sme.fp8.scale.x2.nxv2f64(<vscale x 2 x double> %zdn1, <vscale x 2 x double> %zdn2, <vscale x 2 x i64> %zm1, <vscale x 2 x i64> %zm2)
  ret { <vscale x 2 x double>, <vscale x 2 x double> } %res
}

; FSCALE (Multi, x4)
define { <vscale x 8 x half>, <vscale x 8 x half>, <vscale x 8 x half>, <vscale x 8 x half> } @multi_vec_scale_x4_half( <vscale x 8 x half> %zdn1, <vscale x 8 x half> %zdn2, <vscale x 8 x half> %zdn3, <vscale x 8 x half> %zdn4, <vscale x 8 x i16> %zm1, <vscale x 8 x i16> %zm2, <vscale x 8 x i16> %zm3, <vscale x 8 x i16> %zm4) {
; CHECK-LABEL: multi_vec_scale_x4_half:
; CHECK:       // %bb.0:
; CHECK-NEXT:    // kill: def $z3 killed $z3 killed $z0_z1_z2_z3 def $z0_z1_z2_z3
; CHECK-NEXT:    // kill: def $z7 killed $z7 killed $z4_z5_z6_z7 def $z4_z5_z6_z7
; CHECK-NEXT:    // kill: def $z2 killed $z2 killed $z0_z1_z2_z3 def $z0_z1_z2_z3
; CHECK-NEXT:    // kill: def $z6 killed $z6 killed $z4_z5_z6_z7 def $z4_z5_z6_z7
; CHECK-NEXT:    // kill: def $z1 killed $z1 killed $z0_z1_z2_z3 def $z0_z1_z2_z3
; CHECK-NEXT:    // kill: def $z5 killed $z5 killed $z4_z5_z6_z7 def $z4_z5_z6_z7
; CHECK-NEXT:    // kill: def $z0 killed $z0 killed $z0_z1_z2_z3 def $z0_z1_z2_z3
; CHECK-NEXT:    // kill: def $z4 killed $z4 killed $z4_z5_z6_z7 def $z4_z5_z6_z7
; CHECK-NEXT:    fscale { z0.h - z3.h }, { z0.h - z3.h }, { z4.h - z7.h }
; CHECK-NEXT:    ret
  %res = call { <vscale x 8 x half>, <vscale x 8 x half>, <vscale x 8 x half>, <vscale x 8 x half> } @llvm.aarch64.sme.fp8.scale.x4.nxv8f16(<vscale x 8 x half> %zdn1, <vscale x 8 x half> %zdn2, <vscale x 8 x half> %zdn3, <vscale x 8 x half> %zdn4, <vscale x 8 x i16> %zm1, <vscale x 8 x i16> %zm2, <vscale x 8 x i16> %zm3, <vscale x 8 x i16> %zm4)
  ret { <vscale x 8 x half>, <vscale x 8 x half>, <vscale x 8 x half>, <vscale x 8 x half> } %res
}

define { <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float> } @multi_vec_scale_x4_float( <vscale x 4 x float> %zdn1, <vscale x 4 x float> %zdn2,  <vscale x 4 x float> %zdn3, <vscale x 4 x float> %zdn4, <vscale x 4 x i32> %zm1, <vscale x 4 x i32> %zm2, <vscale x 4 x i32> %zm3, <vscale x 4 x i32> %zm4) {
; CHECK-LABEL: multi_vec_scale_x4_float:
; CHECK:       // %bb.0:
; CHECK-NEXT:    // kill: def $z3 killed $z3 killed $z0_z1_z2_z3 def $z0_z1_z2_z3
; CHECK-NEXT:    // kill: def $z7 killed $z7 killed $z4_z5_z6_z7 def $z4_z5_z6_z7
; CHECK-NEXT:    // kill: def $z2 killed $z2 killed $z0_z1_z2_z3 def $z0_z1_z2_z3
; CHECK-NEXT:    // kill: def $z6 killed $z6 killed $z4_z5_z6_z7 def $z4_z5_z6_z7
; CHECK-NEXT:    // kill: def $z1 killed $z1 killed $z0_z1_z2_z3 def $z0_z1_z2_z3
; CHECK-NEXT:    // kill: def $z5 killed $z5 killed $z4_z5_z6_z7 def $z4_z5_z6_z7
; CHECK-NEXT:    // kill: def $z0 killed $z0 killed $z0_z1_z2_z3 def $z0_z1_z2_z3
; CHECK-NEXT:    // kill: def $z4 killed $z4 killed $z4_z5_z6_z7 def $z4_z5_z6_z7
; CHECK-NEXT:    fscale { z0.s - z3.s }, { z0.s - z3.s }, { z4.s - z7.s }
; CHECK-NEXT:    ret
  %res = call { <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float> } @llvm.aarch64.sme.fp8.scale.x4.nxv4f32(<vscale x 4 x float> %zdn1, <vscale x 4 x float> %zdn2, <vscale x 4 x float> %zdn3, <vscale x 4 x float> %zdn4, <vscale x 4 x i32> %zm1,  <vscale x 4 x i32> %zm2, <vscale x 4 x i32> %zm3, <vscale x 4 x i32> %zm4)
  ret { <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float> } %res
}

define { <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double> } @multi_vec_scale_x4_double( <vscale x 2 x double> %zdn1, <vscale x 2 x double> %zdn2, <vscale x 2 x double> %zdn3, <vscale x 2 x double> %zdn4, <vscale x 2 x i64> %zm1, <vscale x 2 x i64> %zm2, <vscale x 2 x i64> %zm3, <vscale x 2 x i64> %zm4) {
; CHECK-LABEL: multi_vec_scale_x4_double:
; CHECK:       // %bb.0:
; CHECK-NEXT:    // kill: def $z3 killed $z3 killed $z0_z1_z2_z3 def $z0_z1_z2_z3
; CHECK-NEXT:    // kill: def $z7 killed $z7 killed $z4_z5_z6_z7 def $z4_z5_z6_z7
; CHECK-NEXT:    // kill: def $z2 killed $z2 killed $z0_z1_z2_z3 def $z0_z1_z2_z3
; CHECK-NEXT:    // kill: def $z6 killed $z6 killed $z4_z5_z6_z7 def $z4_z5_z6_z7
; CHECK-NEXT:    // kill: def $z1 killed $z1 killed $z0_z1_z2_z3 def $z0_z1_z2_z3
; CHECK-NEXT:    // kill: def $z5 killed $z5 killed $z4_z5_z6_z7 def $z4_z5_z6_z7
; CHECK-NEXT:    // kill: def $z0 killed $z0 killed $z0_z1_z2_z3 def $z0_z1_z2_z3
; CHECK-NEXT:    // kill: def $z4 killed $z4 killed $z4_z5_z6_z7 def $z4_z5_z6_z7
; CHECK-NEXT:    fscale { z0.d - z3.d }, { z0.d - z3.d }, { z4.d - z7.d }
; CHECK-NEXT:    ret
  %res = call { <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double> } @llvm.aarch64.sme.fp8.scale.x4.nxv2f64(<vscale x 2 x double> %zdn1, <vscale x 2 x double> %zdn2, <vscale x 2 x double> %zdn3, <vscale x 2 x double> %zdn4, <vscale x 2 x i64> %zm1, <vscale x 2 x i64> %zm2, <vscale x 2 x i64> %zm3, <vscale x 2 x i64> %zm4)
  ret { <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double> } %res
}

declare { <vscale x 8 x half>, <vscale x 8 x half> } @llvm.aarch64.sme.fp8.scale.single.x2.nxv8f16(<vscale x 8 x half>, <vscale x 8 x half>, <vscale x 8 x i16>)
declare { <vscale x 4 x float>, <vscale x 4 x float> } @llvm.aarch64.sme.fp8.scale.single.x2.nxv4f32(<vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x i32>)
declare { <vscale x 2 x double>, <vscale x 2 x double> } @llvm.aarch64.sme.fp8.scale.single.x2.nxv2f64(<vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x i64>)

declare { <vscale x 8 x half>, <vscale x 8 x half>, <vscale x 8 x half>, <vscale x 8 x half> } @llvm.aarch64.sme.fp8.scale.single.x4.nxv8f16(<vscale x 8 x half>, <vscale x 8 x half>,<vscale x 8 x half>, <vscale x 8 x half>, <vscale x 8 x i16>)
declare { <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float> } @llvm.aarch64.sme.fp8.scale.single.x4.nxv4f32(<vscale x 4 x float>, <vscale x 4 x float>,<vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x i32>)
declare { <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double> } @llvm.aarch64.sme.fp8.scale.single.x4.nxv2f64(<vscale x 2 x double>, <vscale x 2 x double>,<vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x i64>)

declare { <vscale x 8 x half>, <vscale x 8 x half> } @llvm.aarch64.sme.fp8.scale.x2.nxv8f16(<vscale x 8 x half>, <vscale x 8 x half>, <vscale x 8 x i16>, <vscale x 8 x i16>)
declare { <vscale x 4 x float>, <vscale x 4 x float> } @llvm.aarch64.sme.fp8.scale.x2.nxv4f32(<vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x i32>, <vscale x 4 x i32>)
declare { <vscale x 2 x double>, <vscale x 2 x double> } @llvm.aarch64.sme.fp8.scale.x2.nxv2f64(<vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x i64>, <vscale x 2 x i64>)

declare { <vscale x 8 x half>, <vscale x 8 x half>, <vscale x 8 x half>, <vscale x 8 x half> } @llvm.aarch64.sme.fp8.scale.x4.nxv8f16(<vscale x 8 x half>, <vscale x 8 x half>,<vscale x 8 x half>, <vscale x 8 x half>, <vscale x 8 x i16>, <vscale x 8 x i16>, <vscale x 8 x i16>, <vscale x 8 x i16>)
declare { <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float> } @llvm.aarch64.sme.fp8.scale.x4.nxv4f32(<vscale x 4 x float>, <vscale x 4 x float>,<vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32>)
declare { <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double> } @llvm.aarch64.sme.fp8.scale.x4.nxv2f64(<vscale x 2 x double>, <vscale x 2 x double>,<vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x i64>, <vscale x 2 x i64>, <vscale x 2 x i64>, <vscale x 2 x i64>)