; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -mtriple=amdgpu10.10-- < %s | FileCheck -check-prefix=GFX10 %s
; RUN: llc -mtriple=amdgpu9.00-- < %s | FileCheck -check-prefix=GFX9 %s
; RUN: llc -mtriple=amdgpu8.10-- < %s | FileCheck -check-prefix=GFX8 %s
; RUN: llc -mattr=+real-true16 -mtriple=amdgpu11.00-- < %s | FileCheck -check-prefixes=GFX11,GFX11-TRUE16 %s
; RUN: llc -mattr=-real-true16 -mtriple=amdgpu11.00-- < %s | FileCheck -check-prefixes=GFX11,GFX11-FAKE16 %s
; RUN: llc -mattr=+real-true16 -mtriple=amdgpu12.00-- < %s | FileCheck -check-prefixes=GFX12,GFX12-TRUE16 %s
; RUN: llc -mattr=-real-true16 -mtriple=amdgpu12.00-- < %s | FileCheck -check-prefixes=GFX12,GFX12-FAKE16 %s

@esgs_ring = external addrspace(3) global [0 x i32], align 65536

define amdgpu_gs void @main(<4 x i32> %arg, i32 %arg1) {
; GFX10-LABEL: main:
; GFX10:       ; %bb.0: ; %bb
; GFX10-NEXT:    s_mov_b32 s4, exec_lo
; GFX10-NEXT:    s_mov_b32 s5, s4
; GFX10-NEXT:  .LBB0_1: ; =>This Inner Loop Header: Depth=1
; GFX10-NEXT:    v_readfirstlane_b32 s0, v0
; GFX10-NEXT:    v_readfirstlane_b32 s1, v1
; GFX10-NEXT:    v_readfirstlane_b32 s2, v2
; GFX10-NEXT:    v_readfirstlane_b32 s3, v3
; GFX10-NEXT:    s_waitcnt_depctr depctr_sa_sdst(0)
; GFX10-NEXT:    v_cmpx_eq_u64_e32 s[0:1], v[0:1]
; GFX10-NEXT:    v_cmpx_eq_u64_e32 s[2:3], v[2:3]
; GFX10-NEXT:    buffer_load_format_d16_xyz v[5:6], v4, s[0:3], 0 idxen
; GFX10-NEXT:    s_andn2_wrexec_b32 s5, s5
; GFX10-NEXT:    ; implicit-def: $vgpr0_vgpr1_vgpr2_vgpr3
; GFX10-NEXT:    ; implicit-def: $vgpr4
; GFX10-NEXT:    s_cbranch_execnz .LBB0_1
; GFX10-NEXT:  ; %bb.2:
; GFX10-NEXT:    s_waitcnt_depctr depctr_vm_vsrc(0)
; GFX10-NEXT:    s_mov_b32 exec_lo, s4
; GFX10-NEXT:    s_waitcnt vmcnt(0)
; GFX10-NEXT:    v_lshrrev_b32_e32 v0, 16, v5
; GFX10-NEXT:    v_and_b32_e32 v1, 0xffff, v6
; GFX10-NEXT:    v_mov_b32_e32 v2, 0
; GFX10-NEXT:    s_waitcnt_vscnt null, 0x0
; GFX10-NEXT:    ds_write2_b32 v2, v0, v1 offset0:7 offset1:8
;
; GFX9-LABEL: main:
; GFX9:       ; %bb.0: ; %bb
; GFX9-NEXT:    s_mov_b64 s[2:3], exec
; GFX9-NEXT:  .LBB0_1: ; =>This Inner Loop Header: Depth=1
; GFX9-NEXT:    v_readfirstlane_b32 s4, v0
; GFX9-NEXT:    v_readfirstlane_b32 s5, v1
; GFX9-NEXT:    v_readfirstlane_b32 s6, v2
; GFX9-NEXT:    v_readfirstlane_b32 s7, v3
; GFX9-NEXT:    v_cmp_eq_u64_e32 vcc, s[4:5], v[0:1]
; GFX9-NEXT:    v_cmp_eq_u64_e64 s[0:1], s[6:7], v[2:3]
; GFX9-NEXT:    s_and_b64 s[0:1], vcc, s[0:1]
; GFX9-NEXT:    s_and_saveexec_b64 s[0:1], s[0:1]
; GFX9-NEXT:    s_nop 0
; GFX9-NEXT:    buffer_load_format_d16_xyz v[5:6], v4, s[4:7], 0 idxen
; GFX9-NEXT:    ; implicit-def: $vgpr0_vgpr1_vgpr2_vgpr3
; GFX9-NEXT:    ; implicit-def: $vgpr4
; GFX9-NEXT:    s_xor_b64 exec, exec, s[0:1]
; GFX9-NEXT:    s_cbranch_execnz .LBB0_1
; GFX9-NEXT:  ; %bb.2:
; GFX9-NEXT:    s_mov_b64 exec, s[2:3]
; GFX9-NEXT:    s_waitcnt vmcnt(0)
; GFX9-NEXT:    v_lshrrev_b32_e32 v0, 16, v5
; GFX9-NEXT:    v_and_b32_e32 v1, 0xffff, v6
; GFX9-NEXT:    v_mov_b32_e32 v2, 0
; GFX9-NEXT:    ds_write2_b32 v2, v0, v1 offset0:7 offset1:8
;
; GFX8-LABEL: main:
; GFX8:       ; %bb.0: ; %bb
; GFX8-NEXT:    s_mov_b64 s[2:3], exec
; GFX8-NEXT:  .LBB0_1: ; =>This Inner Loop Header: Depth=1
; GFX8-NEXT:    v_readfirstlane_b32 s4, v0
; GFX8-NEXT:    v_readfirstlane_b32 s5, v1
; GFX8-NEXT:    v_readfirstlane_b32 s6, v2
; GFX8-NEXT:    v_readfirstlane_b32 s7, v3
; GFX8-NEXT:    v_cmp_eq_u64_e32 vcc, s[4:5], v[0:1]
; GFX8-NEXT:    v_cmp_eq_u64_e64 s[0:1], s[6:7], v[2:3]
; GFX8-NEXT:    s_and_b64 s[0:1], vcc, s[0:1]
; GFX8-NEXT:    s_and_saveexec_b64 s[0:1], s[0:1]
; GFX8-NEXT:    s_nop 0
; GFX8-NEXT:    buffer_load_format_d16_xyz v[5:6], v4, s[4:7], 0 idxen
; GFX8-NEXT:    ; implicit-def: $vgpr0_vgpr1_vgpr2_vgpr3
; GFX8-NEXT:    ; implicit-def: $vgpr4
; GFX8-NEXT:    s_xor_b64 exec, exec, s[0:1]
; GFX8-NEXT:    s_cbranch_execnz .LBB0_1
; GFX8-NEXT:  ; %bb.2:
; GFX8-NEXT:    s_mov_b64 exec, s[2:3]
; GFX8-NEXT:    s_waitcnt vmcnt(0)
; GFX8-NEXT:    v_lshrrev_b32_e32 v0, 16, v5
; GFX8-NEXT:    v_and_b32_e32 v1, 0xffff, v6
; GFX8-NEXT:    v_mov_b32_e32 v2, 0
; GFX8-NEXT:    s_mov_b32 m0, -1
; GFX8-NEXT:    ds_write2_b32 v2, v0, v1 offset0:7 offset1:8
;
; GFX11-LABEL: main:
; GFX11:       ; %bb.0: ; %bb
; GFX11-NEXT:    s_mov_b32 s4, exec_lo
; GFX11-NEXT:    s_delay_alu instid0(SALU_CYCLE_1)
; GFX11-NEXT:    s_mov_b32 s5, s4
; GFX11-NEXT:  .LBB0_1: ; =>This Inner Loop Header: Depth=1
; GFX11-NEXT:    v_readfirstlane_b32 s0, v0
; GFX11-NEXT:    v_readfirstlane_b32 s1, v1
; GFX11-NEXT:    v_readfirstlane_b32 s2, v2
; GFX11-NEXT:    v_readfirstlane_b32 s3, v3
; GFX11-NEXT:    s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_2)
; GFX11-NEXT:    v_cmpx_eq_u64_e32 s[0:1], v[0:1]
; GFX11-NEXT:    v_cmpx_eq_u64_e32 s[2:3], v[2:3]
; GFX11-NEXT:    buffer_load_d16_format_xyz v[5:6], v4, s[0:3], 0 idxen
; GFX11-NEXT:    s_and_not1_wrexec_b32 s5, s5
; GFX11-NEXT:    ; implicit-def: $vgpr0_vgpr1_vgpr2_vgpr3
; GFX11-NEXT:    ; implicit-def: $vgpr4
; GFX11-NEXT:    s_cbranch_execnz .LBB0_1
; GFX11-NEXT:  ; %bb.2:
; GFX11-NEXT:    s_mov_b32 exec_lo, s4
; GFX11-NEXT:    s_waitcnt vmcnt(0)
; GFX11-NEXT:    v_lshrrev_b32_e32 v0, 16, v5
; GFX11-NEXT:    v_dual_mov_b32 v2, 0 :: v_dual_and_b32 v1, 0xffff, v6
; GFX11-NEXT:    ds_store_2addr_b32 v2, v0, v1 offset0:7 offset1:8
;
; GFX12-LABEL: main:
; GFX12:       ; %bb.0: ; %bb
; GFX12-NEXT:    s_mov_b32 s4, exec_lo
; GFX12-NEXT:    s_delay_alu instid0(SALU_CYCLE_1)
; GFX12-NEXT:    s_mov_b32 s5, s4
; GFX12-NEXT:  .LBB0_1: ; =>This Inner Loop Header: Depth=1
; GFX12-NEXT:    v_readfirstlane_b32 s0, v0
; GFX12-NEXT:    v_readfirstlane_b32 s1, v1
; GFX12-NEXT:    v_readfirstlane_b32 s2, v2
; GFX12-NEXT:    v_readfirstlane_b32 s3, v3
; GFX12-NEXT:    s_wait_alu depctr_va_sdst(0)
; GFX12-NEXT:    s_delay_alu instid0(VALU_DEP_3) | instskip(NEXT) | instid1(VALU_DEP_2)
; GFX12-NEXT:    v_cmpx_eq_u64_e32 s[0:1], v[0:1]
; GFX12-NEXT:    v_cmpx_eq_u64_e32 s[2:3], v[2:3]
; GFX12-NEXT:    s_wait_loadcnt 0x0
; GFX12-NEXT:    buffer_load_d16_format_xyz v[5:6], v4, s[0:3], null idxen
; GFX12-NEXT:    s_and_not1_wrexec_b32 s5, s5
; GFX12-NEXT:    ; implicit-def: $vgpr0_vgpr1_vgpr2_vgpr3
; GFX12-NEXT:    ; implicit-def: $vgpr4
; GFX12-NEXT:    s_cbranch_execnz .LBB0_1
; GFX12-NEXT:  ; %bb.2:
; GFX12-NEXT:    s_mov_b32 exec_lo, s4
; GFX12-NEXT:    s_wait_loadcnt 0x0
; GFX12-NEXT:    v_lshrrev_b32_e32 v0, 16, v5
; GFX12-NEXT:    v_dual_mov_b32 v2, 0 :: v_dual_and_b32 v1, 0xffff, v6
; GFX12-NEXT:    ds_store_2addr_b32 v2, v0, v1 offset0:7 offset1:8
bb:
  %i = call i32 @llvm.amdgcn.mbcnt.hi(i32 -1, i32 poison)
  %i2 = call nsz arcp <3 x half> @llvm.amdgcn.struct.buffer.load.format.v3f16(<4 x i32> %arg, i32 %arg1, i32 0, i32 0, i32 0)
  %i3 = bitcast <3 x half> %i2 to <3 x i16>
  %i4 = extractelement <3 x i16> %i3, i32 1
  %i5 = bitcast <3 x half> %i2 to <3 x i16>
  %i6 = extractelement <3 x i16> %i5, i32 2
  %i7 = zext i16 %i4 to i32
  %i8 = zext i16 %i6 to i32
  %i9 = add nuw nsw i32 0, 7
  %i10 = getelementptr [0 x i32], ptr addrspace(3) @esgs_ring, i32 0, i32 %i9
  store i32 %i7, ptr addrspace(3) %i10, align 4
  %i11 = add nuw nsw i32 0, 8
  %i12 = getelementptr [0 x i32], ptr addrspace(3) @esgs_ring, i32 0, i32 %i11
  store i32 %i8, ptr addrspace(3) %i12, align 4
  unreachable
}
; Function Attrs: nounwind readnone willreturn
declare i32 @llvm.amdgcn.mbcnt.hi(i32, i32) #0
; Function Attrs: nounwind readonly willreturn
declare <3 x half> @llvm.amdgcn.struct.buffer.load.format.v3f16(<4 x i32>, i32, i32, i32, i32 immarg) #1
attributes #0 = { nounwind readnone willreturn }
attributes #1 = { nounwind readonly willreturn }
;; NOTE: These prefixes are unused and the list is autogenerated. Do not add tests below this line:
; GFX11-FAKE16: {{.*}}
; GFX11-TRUE16: {{.*}}
; GFX12-FAKE16: {{.*}}
; GFX12-TRUE16: {{.*}}