llvm/test/CodeGen/AMDGPU/v_pack.ll

; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -mtriple=amdgpu10.10 -mattr=-flat-for-global < %s | FileCheck -enable-var-scope -check-prefix=GCN %s
; RUN: llc -global-isel -mtriple=amdgpu10.10 -mattr=-flat-for-global < %s | FileCheck -enable-var-scope -check-prefix=GISEL %s
; RUN: llc -mtriple=amdgpu11.00 -mattr=-flat-for-global,-real-true16 < %s | FileCheck -enable-var-scope -check-prefixes=GFX11-GCN-FAKE16 %s
; RUN: llc -global-isel -mtriple=amdgpu11.00 -mattr=-flat-for-global,-real-true16 < %s | FileCheck -enable-var-scope -check-prefixes=GFX11-GISEL-FAKE16 %s
; RUN: llc -mtriple=amdgpu11.00 -mattr=-flat-for-global,+real-true16 < %s | FileCheck -enable-var-scope -check-prefixes=GFX11-GCN-REAL16 %s
; RUN: llc -global-isel -mtriple=amdgpu11.00 -mattr=-flat-for-global,+real-true16 < %s | FileCheck -enable-var-scope -check-prefixes=GFX11-GISEL-REAL16 %s

declare i32 @llvm.amdgcn.workitem.id.x() #1

define amdgpu_kernel void @v_pack_b32_v2f16(ptr addrspace(1) %in0, ptr addrspace(1) %in1) #0 {
; GCN-LABEL: v_pack_b32_v2f16:
; GCN:       ; %bb.0:
; GCN-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; GCN-NEXT:    v_lshlrev_b32_e32 v0, 1, v0
; GCN-NEXT:    s_waitcnt lgkmcnt(0)
; GCN-NEXT:    global_load_ushort v1, v0, s[0:1] glc dlc
; GCN-NEXT:    s_waitcnt vmcnt(0)
; GCN-NEXT:    global_load_ushort v2, v0, s[2:3] glc dlc
; GCN-NEXT:    s_waitcnt vmcnt(0)
; GCN-NEXT:    v_add_f16_e32 v0, 2.0, v1
; GCN-NEXT:    v_add_f16_e32 v1, 2.0, v2
; GCN-NEXT:    v_pack_b32_f16 v0, v0, v1
; GCN-NEXT:    ;;#ASMSTART
; GCN-NEXT:    ; use v0
; GCN-NEXT:    ;;#ASMEND
; GCN-NEXT:    s_endpgm
;
; GISEL-LABEL: v_pack_b32_v2f16:
; GISEL:       ; %bb.0:
; GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; GISEL-NEXT:    v_lshlrev_b32_e32 v0, 1, v0
; GISEL-NEXT:    s_waitcnt lgkmcnt(0)
; GISEL-NEXT:    global_load_ushort v1, v0, s[0:1] glc dlc
; GISEL-NEXT:    s_waitcnt vmcnt(0)
; GISEL-NEXT:    global_load_ushort v2, v0, s[2:3] glc dlc
; GISEL-NEXT:    s_waitcnt vmcnt(0)
; GISEL-NEXT:    v_add_f16_e32 v0, 2.0, v1
; GISEL-NEXT:    v_add_f16_e32 v1, 2.0, v2
; GISEL-NEXT:    v_pack_b32_f16 v0, v0, v1
; GISEL-NEXT:    ;;#ASMSTART
; GISEL-NEXT:    ; use v0
; GISEL-NEXT:    ;;#ASMEND
; GISEL-NEXT:    s_endpgm
;
; GFX11-GCN-FAKE16-LABEL: v_pack_b32_v2f16:
; GFX11-GCN-FAKE16:       ; %bb.0:
; GFX11-GCN-FAKE16-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
; GFX11-GCN-FAKE16-NEXT:    v_and_b32_e32 v0, 0x3ff, v0
; GFX11-GCN-FAKE16-NEXT:    s_delay_alu instid0(VALU_DEP_1)
; GFX11-GCN-FAKE16-NEXT:    v_lshlrev_b32_e32 v0, 1, v0
; GFX11-GCN-FAKE16-NEXT:    s_waitcnt lgkmcnt(0)
; GFX11-GCN-FAKE16-NEXT:    global_load_u16 v1, v0, s[0:1] glc dlc
; GFX11-GCN-FAKE16-NEXT:    s_waitcnt vmcnt(0)
; GFX11-GCN-FAKE16-NEXT:    global_load_u16 v0, v0, s[2:3] glc dlc
; GFX11-GCN-FAKE16-NEXT:    s_waitcnt vmcnt(0)
; GFX11-GCN-FAKE16-NEXT:    v_add_f16_e32 v1, 2.0, v1
; GFX11-GCN-FAKE16-NEXT:    v_add_f16_e32 v0, 2.0, v0
; GFX11-GCN-FAKE16-NEXT:    s_delay_alu instid0(VALU_DEP_1)
; GFX11-GCN-FAKE16-NEXT:    v_pack_b32_f16 v0, v1, v0
; GFX11-GCN-FAKE16-NEXT:    ;;#ASMSTART
; GFX11-GCN-FAKE16-NEXT:    ; use v0
; GFX11-GCN-FAKE16-NEXT:    ;;#ASMEND
; GFX11-GCN-FAKE16-NEXT:    s_endpgm
;
; GFX11-GISEL-FAKE16-LABEL: v_pack_b32_v2f16:
; GFX11-GISEL-FAKE16:       ; %bb.0:
; GFX11-GISEL-FAKE16-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
; GFX11-GISEL-FAKE16-NEXT:    v_and_b32_e32 v0, 0x3ff, v0
; GFX11-GISEL-FAKE16-NEXT:    s_delay_alu instid0(VALU_DEP_1)
; GFX11-GISEL-FAKE16-NEXT:    v_lshlrev_b32_e32 v0, 1, v0
; GFX11-GISEL-FAKE16-NEXT:    s_waitcnt lgkmcnt(0)
; GFX11-GISEL-FAKE16-NEXT:    global_load_u16 v1, v0, s[0:1] glc dlc
; GFX11-GISEL-FAKE16-NEXT:    s_waitcnt vmcnt(0)
; GFX11-GISEL-FAKE16-NEXT:    global_load_u16 v0, v0, s[2:3] glc dlc
; GFX11-GISEL-FAKE16-NEXT:    s_waitcnt vmcnt(0)
; GFX11-GISEL-FAKE16-NEXT:    v_add_f16_e32 v1, 2.0, v1
; GFX11-GISEL-FAKE16-NEXT:    v_add_f16_e32 v0, 2.0, v0
; GFX11-GISEL-FAKE16-NEXT:    s_delay_alu instid0(VALU_DEP_1)
; GFX11-GISEL-FAKE16-NEXT:    v_pack_b32_f16 v0, v1, v0
; GFX11-GISEL-FAKE16-NEXT:    ;;#ASMSTART
; GFX11-GISEL-FAKE16-NEXT:    ; use v0
; GFX11-GISEL-FAKE16-NEXT:    ;;#ASMEND
; GFX11-GISEL-FAKE16-NEXT:    s_endpgm
;
; GFX11-GCN-REAL16-LABEL: v_pack_b32_v2f16:
; GFX11-GCN-REAL16:       ; %bb.0:
; GFX11-GCN-REAL16-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
; GFX11-GCN-REAL16-NEXT:    v_and_b32_e32 v0, 0x3ff, v0
; GFX11-GCN-REAL16-NEXT:    s_delay_alu instid0(VALU_DEP_1)
; GFX11-GCN-REAL16-NEXT:    v_lshlrev_b32_e32 v1, 1, v0
; GFX11-GCN-REAL16-NEXT:    s_waitcnt lgkmcnt(0)
; GFX11-GCN-REAL16-NEXT:    global_load_d16_b16 v0, v1, s[0:1] glc dlc
; GFX11-GCN-REAL16-NEXT:    s_waitcnt vmcnt(0)
; GFX11-GCN-REAL16-NEXT:    global_load_d16_hi_b16 v0, v1, s[2:3] glc dlc
; GFX11-GCN-REAL16-NEXT:    s_waitcnt vmcnt(0)
; GFX11-GCN-REAL16-NEXT:    v_add_f16_e32 v0.l, 2.0, v0.l
; GFX11-GCN-REAL16-NEXT:    v_add_f16_e32 v0.h, 2.0, v0.h
; GFX11-GCN-REAL16-NEXT:    ;;#ASMSTART
; GFX11-GCN-REAL16-NEXT:    ; use v0
; GFX11-GCN-REAL16-NEXT:    ;;#ASMEND
; GFX11-GCN-REAL16-NEXT:    s_endpgm
;
; GFX11-GISEL-REAL16-LABEL: v_pack_b32_v2f16:
; GFX11-GISEL-REAL16:       ; %bb.0:
; GFX11-GISEL-REAL16-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
; GFX11-GISEL-REAL16-NEXT:    v_and_b32_e32 v0, 0x3ff, v0
; GFX11-GISEL-REAL16-NEXT:    s_delay_alu instid0(VALU_DEP_1)
; GFX11-GISEL-REAL16-NEXT:    v_lshlrev_b32_e32 v1, 1, v0
; GFX11-GISEL-REAL16-NEXT:    s_waitcnt lgkmcnt(0)
; GFX11-GISEL-REAL16-NEXT:    global_load_d16_b16 v0, v1, s[0:1] glc dlc
; GFX11-GISEL-REAL16-NEXT:    s_waitcnt vmcnt(0)
; GFX11-GISEL-REAL16-NEXT:    global_load_d16_hi_b16 v0, v1, s[2:3] glc dlc
; GFX11-GISEL-REAL16-NEXT:    s_waitcnt vmcnt(0)
; GFX11-GISEL-REAL16-NEXT:    v_add_f16_e32 v0.l, 2.0, v0.l
; GFX11-GISEL-REAL16-NEXT:    v_add_f16_e32 v0.h, 2.0, v0.h
; GFX11-GISEL-REAL16-NEXT:    ;;#ASMSTART
; GFX11-GISEL-REAL16-NEXT:    ; use v0
; GFX11-GISEL-REAL16-NEXT:    ;;#ASMEND
; GFX11-GISEL-REAL16-NEXT:    s_endpgm
  %tid = call i32 @llvm.amdgcn.workitem.id.x()
  %tid.ext = sext i32 %tid to i64
  %in0.gep = getelementptr inbounds half, ptr addrspace(1) %in0, i64 %tid.ext
  %in1.gep = getelementptr inbounds half, ptr addrspace(1) %in1, i64 %tid.ext
  %v0 = load volatile half, ptr addrspace(1) %in0.gep
  %v1 = load volatile half, ptr addrspace(1) %in1.gep
  %v0.add = fadd half %v0, 2.0
  %v1.add = fadd half %v1, 2.0
  %vec.0 = insertelement <2 x half> poison, half %v0.add, i32 0
  %vec.1 = insertelement <2 x half> %vec.0, half %v1.add, i32 1
  %vec.i32 = bitcast <2 x half> %vec.1 to i32
  call void asm sideeffect "; use $0", "v"(i32 %vec.i32) #0
  ret void
}

define amdgpu_kernel void @v_pack_b32_v2f16_sub(ptr addrspace(1) %in0, ptr addrspace(1) %in1) #0 {
; GCN-LABEL: v_pack_b32_v2f16_sub:
; GCN:       ; %bb.0:
; GCN-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; GCN-NEXT:    v_lshlrev_b32_e32 v0, 1, v0
; GCN-NEXT:    s_waitcnt lgkmcnt(0)
; GCN-NEXT:    global_load_ushort v1, v0, s[0:1] glc dlc
; GCN-NEXT:    s_waitcnt vmcnt(0)
; GCN-NEXT:    global_load_ushort v2, v0, s[2:3] glc dlc
; GCN-NEXT:    s_waitcnt vmcnt(0)
; GCN-NEXT:    v_subrev_f16_e32 v0, 2.0, v1
; GCN-NEXT:    v_add_f16_e32 v1, 2.0, v2
; GCN-NEXT:    v_pack_b32_f16 v0, v0, v1
; GCN-NEXT:    ;;#ASMSTART
; GCN-NEXT:    ; use v0
; GCN-NEXT:    ;;#ASMEND
; GCN-NEXT:    s_endpgm
;
; GISEL-LABEL: v_pack_b32_v2f16_sub:
; GISEL:       ; %bb.0:
; GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; GISEL-NEXT:    v_lshlrev_b32_e32 v0, 1, v0
; GISEL-NEXT:    s_waitcnt lgkmcnt(0)
; GISEL-NEXT:    global_load_ushort v1, v0, s[0:1] glc dlc
; GISEL-NEXT:    s_waitcnt vmcnt(0)
; GISEL-NEXT:    global_load_ushort v2, v0, s[2:3] glc dlc
; GISEL-NEXT:    s_waitcnt vmcnt(0)
; GISEL-NEXT:    v_subrev_f16_e32 v0, 2.0, v1
; GISEL-NEXT:    v_add_f16_e32 v1, 2.0, v2
; GISEL-NEXT:    v_pack_b32_f16 v0, v0, v1
; GISEL-NEXT:    ;;#ASMSTART
; GISEL-NEXT:    ; use v0
; GISEL-NEXT:    ;;#ASMEND
; GISEL-NEXT:    s_endpgm
;
; GFX11-GCN-FAKE16-LABEL: v_pack_b32_v2f16_sub:
; GFX11-GCN-FAKE16:       ; %bb.0:
; GFX11-GCN-FAKE16-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
; GFX11-GCN-FAKE16-NEXT:    v_and_b32_e32 v0, 0x3ff, v0
; GFX11-GCN-FAKE16-NEXT:    s_delay_alu instid0(VALU_DEP_1)
; GFX11-GCN-FAKE16-NEXT:    v_lshlrev_b32_e32 v0, 1, v0
; GFX11-GCN-FAKE16-NEXT:    s_waitcnt lgkmcnt(0)
; GFX11-GCN-FAKE16-NEXT:    global_load_u16 v1, v0, s[0:1] glc dlc
; GFX11-GCN-FAKE16-NEXT:    s_waitcnt vmcnt(0)
; GFX11-GCN-FAKE16-NEXT:    global_load_u16 v0, v0, s[2:3] glc dlc
; GFX11-GCN-FAKE16-NEXT:    s_waitcnt vmcnt(0)
; GFX11-GCN-FAKE16-NEXT:    v_subrev_f16_e32 v1, 2.0, v1
; GFX11-GCN-FAKE16-NEXT:    v_add_f16_e32 v0, 2.0, v0
; GFX11-GCN-FAKE16-NEXT:    s_delay_alu instid0(VALU_DEP_1)
; GFX11-GCN-FAKE16-NEXT:    v_pack_b32_f16 v0, v1, v0
; GFX11-GCN-FAKE16-NEXT:    ;;#ASMSTART
; GFX11-GCN-FAKE16-NEXT:    ; use v0
; GFX11-GCN-FAKE16-NEXT:    ;;#ASMEND
; GFX11-GCN-FAKE16-NEXT:    s_endpgm
;
; GFX11-GISEL-FAKE16-LABEL: v_pack_b32_v2f16_sub:
; GFX11-GISEL-FAKE16:       ; %bb.0:
; GFX11-GISEL-FAKE16-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
; GFX11-GISEL-FAKE16-NEXT:    v_and_b32_e32 v0, 0x3ff, v0
; GFX11-GISEL-FAKE16-NEXT:    s_delay_alu instid0(VALU_DEP_1)
; GFX11-GISEL-FAKE16-NEXT:    v_lshlrev_b32_e32 v0, 1, v0
; GFX11-GISEL-FAKE16-NEXT:    s_waitcnt lgkmcnt(0)
; GFX11-GISEL-FAKE16-NEXT:    global_load_u16 v1, v0, s[0:1] glc dlc
; GFX11-GISEL-FAKE16-NEXT:    s_waitcnt vmcnt(0)
; GFX11-GISEL-FAKE16-NEXT:    global_load_u16 v0, v0, s[2:3] glc dlc
; GFX11-GISEL-FAKE16-NEXT:    s_waitcnt vmcnt(0)
; GFX11-GISEL-FAKE16-NEXT:    v_subrev_f16_e32 v1, 2.0, v1
; GFX11-GISEL-FAKE16-NEXT:    v_add_f16_e32 v0, 2.0, v0
; GFX11-GISEL-FAKE16-NEXT:    s_delay_alu instid0(VALU_DEP_1)
; GFX11-GISEL-FAKE16-NEXT:    v_pack_b32_f16 v0, v1, v0
; GFX11-GISEL-FAKE16-NEXT:    ;;#ASMSTART
; GFX11-GISEL-FAKE16-NEXT:    ; use v0
; GFX11-GISEL-FAKE16-NEXT:    ;;#ASMEND
; GFX11-GISEL-FAKE16-NEXT:    s_endpgm
;
; GFX11-GCN-REAL16-LABEL: v_pack_b32_v2f16_sub:
; GFX11-GCN-REAL16:       ; %bb.0:
; GFX11-GCN-REAL16-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
; GFX11-GCN-REAL16-NEXT:    v_and_b32_e32 v0, 0x3ff, v0
; GFX11-GCN-REAL16-NEXT:    s_delay_alu instid0(VALU_DEP_1)
; GFX11-GCN-REAL16-NEXT:    v_lshlrev_b32_e32 v1, 1, v0
; GFX11-GCN-REAL16-NEXT:    s_waitcnt lgkmcnt(0)
; GFX11-GCN-REAL16-NEXT:    global_load_d16_b16 v0, v1, s[0:1] glc dlc
; GFX11-GCN-REAL16-NEXT:    s_waitcnt vmcnt(0)
; GFX11-GCN-REAL16-NEXT:    global_load_d16_hi_b16 v0, v1, s[2:3] glc dlc
; GFX11-GCN-REAL16-NEXT:    s_waitcnt vmcnt(0)
; GFX11-GCN-REAL16-NEXT:    v_subrev_f16_e32 v0.l, 2.0, v0.l
; GFX11-GCN-REAL16-NEXT:    v_add_f16_e32 v0.h, 2.0, v0.h
; GFX11-GCN-REAL16-NEXT:    ;;#ASMSTART
; GFX11-GCN-REAL16-NEXT:    ; use v0
; GFX11-GCN-REAL16-NEXT:    ;;#ASMEND
; GFX11-GCN-REAL16-NEXT:    s_endpgm
;
; GFX11-GISEL-REAL16-LABEL: v_pack_b32_v2f16_sub:
; GFX11-GISEL-REAL16:       ; %bb.0:
; GFX11-GISEL-REAL16-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
; GFX11-GISEL-REAL16-NEXT:    v_and_b32_e32 v0, 0x3ff, v0
; GFX11-GISEL-REAL16-NEXT:    s_delay_alu instid0(VALU_DEP_1)
; GFX11-GISEL-REAL16-NEXT:    v_lshlrev_b32_e32 v1, 1, v0
; GFX11-GISEL-REAL16-NEXT:    s_waitcnt lgkmcnt(0)
; GFX11-GISEL-REAL16-NEXT:    global_load_d16_b16 v0, v1, s[0:1] glc dlc
; GFX11-GISEL-REAL16-NEXT:    s_waitcnt vmcnt(0)
; GFX11-GISEL-REAL16-NEXT:    global_load_d16_hi_b16 v0, v1, s[2:3] glc dlc
; GFX11-GISEL-REAL16-NEXT:    s_waitcnt vmcnt(0)
; GFX11-GISEL-REAL16-NEXT:    v_subrev_f16_e32 v0.l, 2.0, v0.l
; GFX11-GISEL-REAL16-NEXT:    v_add_f16_e32 v0.h, 2.0, v0.h
; GFX11-GISEL-REAL16-NEXT:    ;;#ASMSTART
; GFX11-GISEL-REAL16-NEXT:    ; use v0
; GFX11-GISEL-REAL16-NEXT:    ;;#ASMEND
; GFX11-GISEL-REAL16-NEXT:    s_endpgm
  %tid = call i32 @llvm.amdgcn.workitem.id.x()
  %tid.ext = sext i32 %tid to i64
  %in0.gep = getelementptr inbounds half, ptr addrspace(1) %in0, i64 %tid.ext
  %in1.gep = getelementptr inbounds half, ptr addrspace(1) %in1, i64 %tid.ext
  %v0 = load volatile half, ptr addrspace(1) %in0.gep
  %v1 = load volatile half, ptr addrspace(1) %in1.gep
  %v0.add = fsub half %v0, 2.0
  %v1.add = fadd half %v1, 2.0
  %vec.0 = insertelement <2 x half> poison, half %v0.add, i32 0
  %vec.1 = insertelement <2 x half> %vec.0, half %v1.add, i32 1
  %vec.i32 = bitcast <2 x half> %vec.1 to i32
  call void asm sideeffect "; use $0", "v"(i32 %vec.i32) #0
  ret void
}

define <2 x half> @fptrunc(<2 x float> %a) {
; GCN-LABEL: fptrunc:
; GCN:       ; %bb.0:
; GCN-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GCN-NEXT:    v_cvt_f16_f32_e32 v1, v1
; GCN-NEXT:    v_cvt_f16_f32_e32 v0, v0
; GCN-NEXT:    v_pack_b32_f16 v0, v0, v1
; GCN-NEXT:    s_setpc_b64 s[30:31]
;
; GISEL-LABEL: fptrunc:
; GISEL:       ; %bb.0:
; GISEL-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GISEL-NEXT:    v_cvt_f16_f32_e32 v0, v0
; GISEL-NEXT:    v_cvt_f16_f32_e32 v1, v1
; GISEL-NEXT:    v_pack_b32_f16 v0, v0, v1
; GISEL-NEXT:    s_setpc_b64 s[30:31]
;
; GFX11-GCN-FAKE16-LABEL: fptrunc:
; GFX11-GCN-FAKE16:       ; %bb.0:
; GFX11-GCN-FAKE16-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX11-GCN-FAKE16-NEXT:    v_cvt_f16_f32_e32 v1, v1
; GFX11-GCN-FAKE16-NEXT:    v_cvt_f16_f32_e32 v0, v0
; GFX11-GCN-FAKE16-NEXT:    s_delay_alu instid0(VALU_DEP_1)
; GFX11-GCN-FAKE16-NEXT:    v_pack_b32_f16 v0, v0, v1
; GFX11-GCN-FAKE16-NEXT:    s_setpc_b64 s[30:31]
;
; GFX11-GISEL-FAKE16-LABEL: fptrunc:
; GFX11-GISEL-FAKE16:       ; %bb.0:
; GFX11-GISEL-FAKE16-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX11-GISEL-FAKE16-NEXT:    v_cvt_f16_f32_e32 v0, v0
; GFX11-GISEL-FAKE16-NEXT:    v_cvt_f16_f32_e32 v1, v1
; GFX11-GISEL-FAKE16-NEXT:    s_delay_alu instid0(VALU_DEP_1)
; GFX11-GISEL-FAKE16-NEXT:    v_pack_b32_f16 v0, v0, v1
; GFX11-GISEL-FAKE16-NEXT:    s_setpc_b64 s[30:31]
;
; GFX11-GCN-REAL16-LABEL: fptrunc:
; GFX11-GCN-REAL16:       ; %bb.0:
; GFX11-GCN-REAL16-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX11-GCN-REAL16-NEXT:    v_cvt_f16_f32_e32 v1.h, v1
; GFX11-GCN-REAL16-NEXT:    v_cvt_f16_f32_e32 v1.l, v0
; GFX11-GCN-REAL16-NEXT:    s_delay_alu instid0(VALU_DEP_1)
; GFX11-GCN-REAL16-NEXT:    v_mov_b32_e32 v0, v1
; GFX11-GCN-REAL16-NEXT:    s_setpc_b64 s[30:31]
;
; GFX11-GISEL-REAL16-LABEL: fptrunc:
; GFX11-GISEL-REAL16:       ; %bb.0:
; GFX11-GISEL-REAL16-NEXT:    s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX11-GISEL-REAL16-NEXT:    v_cvt_f16_f32_e32 v0.l, v0
; GFX11-GISEL-REAL16-NEXT:    v_cvt_f16_f32_e32 v0.h, v1
; GFX11-GISEL-REAL16-NEXT:    s_setpc_b64 s[30:31]
  %r.val = fptrunc <2 x float> %a to <2 x half>
  ret <2 x half> %r.val
}

define amdgpu_kernel void @v_pack_b32.fabs(ptr addrspace(1) %in0, ptr addrspace(1) %in1) #0 {
; GCN-LABEL: v_pack_b32.fabs:
; GCN:       ; %bb.0:
; GCN-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; GCN-NEXT:    v_lshlrev_b32_e32 v0, 1, v0
; GCN-NEXT:    s_waitcnt lgkmcnt(0)
; GCN-NEXT:    global_load_ushort v1, v0, s[0:1] glc dlc
; GCN-NEXT:    s_waitcnt vmcnt(0)
; GCN-NEXT:    global_load_ushort v2, v0, s[2:3] glc dlc
; GCN-NEXT:    s_waitcnt vmcnt(0)
; GCN-NEXT:    v_add_f16_e32 v0, 2.0, v1
; GCN-NEXT:    v_add_f16_e32 v1, 2.0, v2
; GCN-NEXT:    v_pack_b32_f16 v0, |v0|, |v1|
; GCN-NEXT:    ;;#ASMSTART
; GCN-NEXT:    ; use v0
; GCN-NEXT:    ;;#ASMEND
; GCN-NEXT:    s_endpgm
;
; GISEL-LABEL: v_pack_b32.fabs:
; GISEL:       ; %bb.0:
; GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; GISEL-NEXT:    v_lshlrev_b32_e32 v0, 1, v0
; GISEL-NEXT:    s_waitcnt lgkmcnt(0)
; GISEL-NEXT:    global_load_ushort v1, v0, s[0:1] glc dlc
; GISEL-NEXT:    s_waitcnt vmcnt(0)
; GISEL-NEXT:    global_load_ushort v2, v0, s[2:3] glc dlc
; GISEL-NEXT:    s_waitcnt vmcnt(0)
; GISEL-NEXT:    v_add_f16_e32 v0, 2.0, v1
; GISEL-NEXT:    v_add_f16_e32 v1, 2.0, v2
; GISEL-NEXT:    v_pack_b32_f16 v0, |v0|, |v1|
; GISEL-NEXT:    ;;#ASMSTART
; GISEL-NEXT:    ; use v0
; GISEL-NEXT:    ;;#ASMEND
; GISEL-NEXT:    s_endpgm
;
; GFX11-GCN-FAKE16-LABEL: v_pack_b32.fabs:
; GFX11-GCN-FAKE16:       ; %bb.0:
; GFX11-GCN-FAKE16-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
; GFX11-GCN-FAKE16-NEXT:    v_and_b32_e32 v0, 0x3ff, v0
; GFX11-GCN-FAKE16-NEXT:    s_delay_alu instid0(VALU_DEP_1)
; GFX11-GCN-FAKE16-NEXT:    v_lshlrev_b32_e32 v0, 1, v0
; GFX11-GCN-FAKE16-NEXT:    s_waitcnt lgkmcnt(0)
; GFX11-GCN-FAKE16-NEXT:    global_load_u16 v1, v0, s[0:1] glc dlc
; GFX11-GCN-FAKE16-NEXT:    s_waitcnt vmcnt(0)
; GFX11-GCN-FAKE16-NEXT:    global_load_u16 v0, v0, s[2:3] glc dlc
; GFX11-GCN-FAKE16-NEXT:    s_waitcnt vmcnt(0)
; GFX11-GCN-FAKE16-NEXT:    v_add_f16_e32 v1, 2.0, v1
; GFX11-GCN-FAKE16-NEXT:    v_add_f16_e32 v0, 2.0, v0
; GFX11-GCN-FAKE16-NEXT:    s_delay_alu instid0(VALU_DEP_1)
; GFX11-GCN-FAKE16-NEXT:    v_pack_b32_f16 v0, |v1|, |v0|
; GFX11-GCN-FAKE16-NEXT:    ;;#ASMSTART
; GFX11-GCN-FAKE16-NEXT:    ; use v0
; GFX11-GCN-FAKE16-NEXT:    ;;#ASMEND
; GFX11-GCN-FAKE16-NEXT:    s_endpgm
;
; GFX11-GISEL-FAKE16-LABEL: v_pack_b32.fabs:
; GFX11-GISEL-FAKE16:       ; %bb.0:
; GFX11-GISEL-FAKE16-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
; GFX11-GISEL-FAKE16-NEXT:    v_and_b32_e32 v0, 0x3ff, v0
; GFX11-GISEL-FAKE16-NEXT:    s_delay_alu instid0(VALU_DEP_1)
; GFX11-GISEL-FAKE16-NEXT:    v_lshlrev_b32_e32 v0, 1, v0
; GFX11-GISEL-FAKE16-NEXT:    s_waitcnt lgkmcnt(0)
; GFX11-GISEL-FAKE16-NEXT:    global_load_u16 v1, v0, s[0:1] glc dlc
; GFX11-GISEL-FAKE16-NEXT:    s_waitcnt vmcnt(0)
; GFX11-GISEL-FAKE16-NEXT:    global_load_u16 v0, v0, s[2:3] glc dlc
; GFX11-GISEL-FAKE16-NEXT:    s_waitcnt vmcnt(0)
; GFX11-GISEL-FAKE16-NEXT:    v_add_f16_e32 v1, 2.0, v1
; GFX11-GISEL-FAKE16-NEXT:    v_add_f16_e32 v0, 2.0, v0
; GFX11-GISEL-FAKE16-NEXT:    s_delay_alu instid0(VALU_DEP_1)
; GFX11-GISEL-FAKE16-NEXT:    v_pack_b32_f16 v0, |v1|, |v0|
; GFX11-GISEL-FAKE16-NEXT:    ;;#ASMSTART
; GFX11-GISEL-FAKE16-NEXT:    ; use v0
; GFX11-GISEL-FAKE16-NEXT:    ;;#ASMEND
; GFX11-GISEL-FAKE16-NEXT:    s_endpgm
;
; GFX11-GCN-REAL16-LABEL: v_pack_b32.fabs:
; GFX11-GCN-REAL16:       ; %bb.0:
; GFX11-GCN-REAL16-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
; GFX11-GCN-REAL16-NEXT:    v_and_b32_e32 v0, 0x3ff, v0
; GFX11-GCN-REAL16-NEXT:    s_delay_alu instid0(VALU_DEP_1)
; GFX11-GCN-REAL16-NEXT:    v_lshlrev_b32_e32 v1, 1, v0
; GFX11-GCN-REAL16-NEXT:    s_waitcnt lgkmcnt(0)
; GFX11-GCN-REAL16-NEXT:    global_load_d16_b16 v0, v1, s[0:1] glc dlc
; GFX11-GCN-REAL16-NEXT:    s_waitcnt vmcnt(0)
; GFX11-GCN-REAL16-NEXT:    global_load_d16_hi_b16 v0, v1, s[2:3] glc dlc
; GFX11-GCN-REAL16-NEXT:    s_waitcnt vmcnt(0)
; GFX11-GCN-REAL16-NEXT:    v_add_f16_e32 v0.l, 2.0, v0.l
; GFX11-GCN-REAL16-NEXT:    v_add_f16_e32 v0.h, 2.0, v0.h
; GFX11-GCN-REAL16-NEXT:    s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2)
; GFX11-GCN-REAL16-NEXT:    v_and_b16 v0.l, 0x7fff, v0.l
; GFX11-GCN-REAL16-NEXT:    v_and_b16 v0.h, 0x7fff, v0.h
; GFX11-GCN-REAL16-NEXT:    ;;#ASMSTART
; GFX11-GCN-REAL16-NEXT:    ; use v0
; GFX11-GCN-REAL16-NEXT:    ;;#ASMEND
; GFX11-GCN-REAL16-NEXT:    s_endpgm
;
; GFX11-GISEL-REAL16-LABEL: v_pack_b32.fabs:
; GFX11-GISEL-REAL16:       ; %bb.0:
; GFX11-GISEL-REAL16-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
; GFX11-GISEL-REAL16-NEXT:    v_and_b32_e32 v0, 0x3ff, v0
; GFX11-GISEL-REAL16-NEXT:    s_delay_alu instid0(VALU_DEP_1)
; GFX11-GISEL-REAL16-NEXT:    v_lshlrev_b32_e32 v1, 1, v0
; GFX11-GISEL-REAL16-NEXT:    s_waitcnt lgkmcnt(0)
; GFX11-GISEL-REAL16-NEXT:    global_load_d16_b16 v0, v1, s[0:1] glc dlc
; GFX11-GISEL-REAL16-NEXT:    s_waitcnt vmcnt(0)
; GFX11-GISEL-REAL16-NEXT:    global_load_d16_hi_b16 v0, v1, s[2:3] glc dlc
; GFX11-GISEL-REAL16-NEXT:    s_waitcnt vmcnt(0)
; GFX11-GISEL-REAL16-NEXT:    v_add_f16_e32 v0.l, 2.0, v0.l
; GFX11-GISEL-REAL16-NEXT:    v_add_f16_e32 v0.h, 2.0, v0.h
; GFX11-GISEL-REAL16-NEXT:    s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2)
; GFX11-GISEL-REAL16-NEXT:    v_and_b16 v0.l, 0x7fff, v0.l
; GFX11-GISEL-REAL16-NEXT:    v_and_b16 v0.h, 0x7fff, v0.h
; GFX11-GISEL-REAL16-NEXT:    ;;#ASMSTART
; GFX11-GISEL-REAL16-NEXT:    ; use v0
; GFX11-GISEL-REAL16-NEXT:    ;;#ASMEND
; GFX11-GISEL-REAL16-NEXT:    s_endpgm
  %tid = call i32 @llvm.amdgcn.workitem.id.x()
  %tid.ext = sext i32 %tid to i64
  %in0.gep = getelementptr inbounds half, ptr addrspace(1) %in0, i64 %tid.ext
  %in1.gep = getelementptr inbounds half, ptr addrspace(1) %in1, i64 %tid.ext
  %v0 = load volatile half, ptr addrspace(1) %in0.gep
  %v1 = load volatile half, ptr addrspace(1) %in1.gep
  %v0.add = fadd half %v0, 2.0
  %v1.add = fadd half %v1, 2.0
  %v0.fabs = call half @llvm.fabs.f16(half %v0.add)
  %v1.fabs = call half @llvm.fabs.f16(half %v1.add)
  %vec.0 = insertelement <2 x half> poison, half %v0.fabs, i32 0
  %vec.1 = insertelement <2 x half> %vec.0, half %v1.fabs, i32 1
  %vec.i32 = bitcast <2 x half> %vec.1 to i32
  call void asm sideeffect "; use $0", "v"(i32 %vec.i32) #0
  ret void
}

define amdgpu_kernel void @v_pack_b32.fneg(ptr addrspace(1) %in0, ptr addrspace(1) %in1) #0 {
; GCN-LABEL: v_pack_b32.fneg:
; GCN:       ; %bb.0:
; GCN-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; GCN-NEXT:    v_lshlrev_b32_e32 v0, 1, v0
; GCN-NEXT:    s_waitcnt lgkmcnt(0)
; GCN-NEXT:    global_load_ushort v1, v0, s[0:1] glc dlc
; GCN-NEXT:    s_waitcnt vmcnt(0)
; GCN-NEXT:    global_load_ushort v2, v0, s[2:3] glc dlc
; GCN-NEXT:    s_waitcnt vmcnt(0)
; GCN-NEXT:    v_add_f16_e32 v0, 2.0, v1
; GCN-NEXT:    v_add_f16_e32 v1, 2.0, v2
; GCN-NEXT:    v_pack_b32_f16 v0, -v0, -v1
; GCN-NEXT:    ;;#ASMSTART
; GCN-NEXT:    ; use v0
; GCN-NEXT:    ;;#ASMEND
; GCN-NEXT:    s_endpgm
;
; GISEL-LABEL: v_pack_b32.fneg:
; GISEL:       ; %bb.0:
; GISEL-NEXT:    s_load_dwordx4 s[0:3], s[4:5], 0x24
; GISEL-NEXT:    v_lshlrev_b32_e32 v0, 1, v0
; GISEL-NEXT:    s_waitcnt lgkmcnt(0)
; GISEL-NEXT:    global_load_ushort v1, v0, s[0:1] glc dlc
; GISEL-NEXT:    s_waitcnt vmcnt(0)
; GISEL-NEXT:    global_load_ushort v2, v0, s[2:3] glc dlc
; GISEL-NEXT:    s_waitcnt vmcnt(0)
; GISEL-NEXT:    v_add_f16_e32 v0, 2.0, v1
; GISEL-NEXT:    v_add_f16_e32 v1, 2.0, v2
; GISEL-NEXT:    v_pack_b32_f16 v0, -v0, -v1
; GISEL-NEXT:    ;;#ASMSTART
; GISEL-NEXT:    ; use v0
; GISEL-NEXT:    ;;#ASMEND
; GISEL-NEXT:    s_endpgm
;
; GFX11-GCN-FAKE16-LABEL: v_pack_b32.fneg:
; GFX11-GCN-FAKE16:       ; %bb.0:
; GFX11-GCN-FAKE16-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
; GFX11-GCN-FAKE16-NEXT:    v_and_b32_e32 v0, 0x3ff, v0
; GFX11-GCN-FAKE16-NEXT:    s_delay_alu instid0(VALU_DEP_1)
; GFX11-GCN-FAKE16-NEXT:    v_lshlrev_b32_e32 v0, 1, v0
; GFX11-GCN-FAKE16-NEXT:    s_waitcnt lgkmcnt(0)
; GFX11-GCN-FAKE16-NEXT:    global_load_u16 v1, v0, s[0:1] glc dlc
; GFX11-GCN-FAKE16-NEXT:    s_waitcnt vmcnt(0)
; GFX11-GCN-FAKE16-NEXT:    global_load_u16 v0, v0, s[2:3] glc dlc
; GFX11-GCN-FAKE16-NEXT:    s_waitcnt vmcnt(0)
; GFX11-GCN-FAKE16-NEXT:    v_add_f16_e32 v1, 2.0, v1
; GFX11-GCN-FAKE16-NEXT:    v_add_f16_e32 v0, 2.0, v0
; GFX11-GCN-FAKE16-NEXT:    s_delay_alu instid0(VALU_DEP_1)
; GFX11-GCN-FAKE16-NEXT:    v_pack_b32_f16 v0, -v1, -v0
; GFX11-GCN-FAKE16-NEXT:    ;;#ASMSTART
; GFX11-GCN-FAKE16-NEXT:    ; use v0
; GFX11-GCN-FAKE16-NEXT:    ;;#ASMEND
; GFX11-GCN-FAKE16-NEXT:    s_endpgm
;
; GFX11-GISEL-FAKE16-LABEL: v_pack_b32.fneg:
; GFX11-GISEL-FAKE16:       ; %bb.0:
; GFX11-GISEL-FAKE16-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
; GFX11-GISEL-FAKE16-NEXT:    v_and_b32_e32 v0, 0x3ff, v0
; GFX11-GISEL-FAKE16-NEXT:    s_delay_alu instid0(VALU_DEP_1)
; GFX11-GISEL-FAKE16-NEXT:    v_lshlrev_b32_e32 v0, 1, v0
; GFX11-GISEL-FAKE16-NEXT:    s_waitcnt lgkmcnt(0)
; GFX11-GISEL-FAKE16-NEXT:    global_load_u16 v1, v0, s[0:1] glc dlc
; GFX11-GISEL-FAKE16-NEXT:    s_waitcnt vmcnt(0)
; GFX11-GISEL-FAKE16-NEXT:    global_load_u16 v0, v0, s[2:3] glc dlc
; GFX11-GISEL-FAKE16-NEXT:    s_waitcnt vmcnt(0)
; GFX11-GISEL-FAKE16-NEXT:    v_add_f16_e32 v1, 2.0, v1
; GFX11-GISEL-FAKE16-NEXT:    v_add_f16_e32 v0, 2.0, v0
; GFX11-GISEL-FAKE16-NEXT:    s_delay_alu instid0(VALU_DEP_1)
; GFX11-GISEL-FAKE16-NEXT:    v_pack_b32_f16 v0, -v1, -v0
; GFX11-GISEL-FAKE16-NEXT:    ;;#ASMSTART
; GFX11-GISEL-FAKE16-NEXT:    ; use v0
; GFX11-GISEL-FAKE16-NEXT:    ;;#ASMEND
; GFX11-GISEL-FAKE16-NEXT:    s_endpgm
;
; GFX11-GCN-REAL16-LABEL: v_pack_b32.fneg:
; GFX11-GCN-REAL16:       ; %bb.0:
; GFX11-GCN-REAL16-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
; GFX11-GCN-REAL16-NEXT:    v_and_b32_e32 v0, 0x3ff, v0
; GFX11-GCN-REAL16-NEXT:    s_delay_alu instid0(VALU_DEP_1)
; GFX11-GCN-REAL16-NEXT:    v_lshlrev_b32_e32 v1, 1, v0
; GFX11-GCN-REAL16-NEXT:    s_waitcnt lgkmcnt(0)
; GFX11-GCN-REAL16-NEXT:    global_load_d16_b16 v0, v1, s[0:1] glc dlc
; GFX11-GCN-REAL16-NEXT:    s_waitcnt vmcnt(0)
; GFX11-GCN-REAL16-NEXT:    global_load_d16_hi_b16 v0, v1, s[2:3] glc dlc
; GFX11-GCN-REAL16-NEXT:    s_waitcnt vmcnt(0)
; GFX11-GCN-REAL16-NEXT:    v_add_f16_e32 v0.l, 2.0, v0.l
; GFX11-GCN-REAL16-NEXT:    v_add_f16_e32 v0.h, 2.0, v0.h
; GFX11-GCN-REAL16-NEXT:    s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2)
; GFX11-GCN-REAL16-NEXT:    v_xor_b16 v0.l, 0x8000, v0.l
; GFX11-GCN-REAL16-NEXT:    v_xor_b16 v0.h, 0x8000, v0.h
; GFX11-GCN-REAL16-NEXT:    ;;#ASMSTART
; GFX11-GCN-REAL16-NEXT:    ; use v0
; GFX11-GCN-REAL16-NEXT:    ;;#ASMEND
; GFX11-GCN-REAL16-NEXT:    s_endpgm
;
; GFX11-GISEL-REAL16-LABEL: v_pack_b32.fneg:
; GFX11-GISEL-REAL16:       ; %bb.0:
; GFX11-GISEL-REAL16-NEXT:    s_load_b128 s[0:3], s[4:5], 0x24
; GFX11-GISEL-REAL16-NEXT:    v_and_b32_e32 v0, 0x3ff, v0
; GFX11-GISEL-REAL16-NEXT:    s_delay_alu instid0(VALU_DEP_1)
; GFX11-GISEL-REAL16-NEXT:    v_lshlrev_b32_e32 v1, 1, v0
; GFX11-GISEL-REAL16-NEXT:    s_waitcnt lgkmcnt(0)
; GFX11-GISEL-REAL16-NEXT:    global_load_d16_b16 v0, v1, s[0:1] glc dlc
; GFX11-GISEL-REAL16-NEXT:    s_waitcnt vmcnt(0)
; GFX11-GISEL-REAL16-NEXT:    global_load_d16_hi_b16 v0, v1, s[2:3] glc dlc
; GFX11-GISEL-REAL16-NEXT:    s_waitcnt vmcnt(0)
; GFX11-GISEL-REAL16-NEXT:    v_add_f16_e32 v0.l, 2.0, v0.l
; GFX11-GISEL-REAL16-NEXT:    v_add_f16_e32 v0.h, 2.0, v0.h
; GFX11-GISEL-REAL16-NEXT:    s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2)
; GFX11-GISEL-REAL16-NEXT:    v_xor_b16 v0.l, 0x8000, v0.l
; GFX11-GISEL-REAL16-NEXT:    v_xor_b16 v0.h, 0x8000, v0.h
; GFX11-GISEL-REAL16-NEXT:    ;;#ASMSTART
; GFX11-GISEL-REAL16-NEXT:    ; use v0
; GFX11-GISEL-REAL16-NEXT:    ;;#ASMEND
; GFX11-GISEL-REAL16-NEXT:    s_endpgm
  %tid = call i32 @llvm.amdgcn.workitem.id.x()
  %tid.ext = sext i32 %tid to i64
  %in0.gep = getelementptr inbounds half, ptr addrspace(1) %in0, i64 %tid.ext
  %in1.gep = getelementptr inbounds half, ptr addrspace(1) %in1, i64 %tid.ext
  %v0 = load volatile half, ptr addrspace(1) %in0.gep
  %v1 = load volatile half, ptr addrspace(1) %in1.gep
  %v0.add = fadd half %v0, 2.0
  %v1.add = fadd half %v1, 2.0
  %v0.fneg = fsub half -0.0, %v0.add
  %v1.fneg = fsub half -0.0, %v1.add
  %vec.0 = insertelement <2 x half> poison, half %v0.fneg, i32 0
  %vec.1 = insertelement <2 x half> %vec.0, half %v1.fneg, i32 1
  %vec.i32 = bitcast <2 x half> %vec.1 to i32
  call void asm sideeffect "; use $0", "v"(i32 %vec.i32) #0
  ret void
}

declare half @llvm.fabs.f16(half) #1

attributes #0 = { nounwind }
attributes #1 = { nounwind readnone }