; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -global-isel -mtriple=amdgcn-amd-amdpal -mcpu=hawaii < %s | FileCheck -check-prefix=GFX7 %s
; RUN: llc -global-isel -mtriple=amdgcn-amd-amdpal -mcpu=gfx900 < %s | FileCheck -check-prefix=GFX9 %s
; RUN: llc -global-isel -mtriple=amdgcn-amd-amdpal -mcpu=fiji < %s | FileCheck -check-prefix=GFX8 %s
; RUN: llc -global-isel -mtriple=amdgcn-amd-amdpal -mcpu=gfx1010 < %s | FileCheck -check-prefix=GFX10 %s
; RUN: llc -global-isel -mtriple=amdgcn-amd-amdpal -mcpu=gfx1100 -amdgpu-enable-delay-alu=0 < %s | FileCheck -check-prefix=GFX10 %s
define <2 x i16> @v_add_v2i16(<2 x i16> %a, <2 x i16> %b) {
; GFX7-LABEL: v_add_v2i16:
; GFX7: ; %bb.0:
; GFX7-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX7-NEXT: v_add_i32_e32 v0, vcc, v0, v2
; GFX7-NEXT: v_add_i32_e32 v1, vcc, v1, v3
; GFX7-NEXT: s_setpc_b64 s[30:31]
;
; GFX9-LABEL: v_add_v2i16:
; GFX9: ; %bb.0:
; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX9-NEXT: v_pk_add_u16 v0, v0, v1
; GFX9-NEXT: s_setpc_b64 s[30:31]
;
; GFX8-LABEL: v_add_v2i16:
; GFX8: ; %bb.0:
; GFX8-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX8-NEXT: v_add_u16_e32 v2, v0, v1
; GFX8-NEXT: v_add_u16_sdwa v0, v0, v1 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_or_b32_e32 v0, v2, v0
; GFX8-NEXT: s_setpc_b64 s[30:31]
;
; GFX10-LABEL: v_add_v2i16:
; GFX10: ; %bb.0:
; GFX10-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX10-NEXT: v_pk_add_u16 v0, v0, v1
; GFX10-NEXT: s_setpc_b64 s[30:31]
%add = add <2 x i16> %a, %b
ret <2 x i16> %add
}
define <2 x i16> @v_add_v2i16_fneg_lhs(<2 x half> %a, <2 x i16> %b) {
; GFX7-LABEL: v_add_v2i16_fneg_lhs:
; GFX7: ; %bb.0:
; GFX7-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX7-NEXT: v_lshlrev_b32_e32 v1, 16, v1
; GFX7-NEXT: v_and_b32_e32 v0, 0xffff, v0
; GFX7-NEXT: v_or_b32_e32 v0, v1, v0
; GFX7-NEXT: v_xor_b32_e32 v0, 0x80008000, v0
; GFX7-NEXT: v_lshrrev_b32_e32 v1, 16, v0
; GFX7-NEXT: v_add_i32_e32 v0, vcc, v0, v2
; GFX7-NEXT: v_add_i32_e32 v1, vcc, v1, v3
; GFX7-NEXT: s_setpc_b64 s[30:31]
;
; GFX9-LABEL: v_add_v2i16_fneg_lhs:
; GFX9: ; %bb.0:
; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX9-NEXT: v_pk_add_u16 v0, v0, v1 neg_lo:[1,0] neg_hi:[1,0]
; GFX9-NEXT: s_setpc_b64 s[30:31]
;
; GFX8-LABEL: v_add_v2i16_fneg_lhs:
; GFX8: ; %bb.0:
; GFX8-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX8-NEXT: v_xor_b32_e32 v0, 0x80008000, v0
; GFX8-NEXT: v_add_u16_e32 v2, v0, v1
; GFX8-NEXT: v_add_u16_sdwa v0, v0, v1 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_or_b32_e32 v0, v2, v0
; GFX8-NEXT: s_setpc_b64 s[30:31]
;
; GFX10-LABEL: v_add_v2i16_fneg_lhs:
; GFX10: ; %bb.0:
; GFX10-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX10-NEXT: v_pk_add_u16 v0, v0, v1 neg_lo:[1,0] neg_hi:[1,0]
; GFX10-NEXT: s_setpc_b64 s[30:31]
%neg.a = fneg <2 x half> %a
%cast.neg.a = bitcast <2 x half> %neg.a to <2 x i16>
%add = add <2 x i16> %cast.neg.a, %b
ret <2 x i16> %add
}
define <2 x i16> @v_add_v2i16_fneg_rhs(<2 x i16> %a, <2 x half> %b) {
; GFX7-LABEL: v_add_v2i16_fneg_rhs:
; GFX7: ; %bb.0:
; GFX7-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX7-NEXT: v_lshlrev_b32_e32 v3, 16, v3
; GFX7-NEXT: v_and_b32_e32 v2, 0xffff, v2
; GFX7-NEXT: v_or_b32_e32 v2, v3, v2
; GFX7-NEXT: v_xor_b32_e32 v2, 0x80008000, v2
; GFX7-NEXT: v_lshrrev_b32_e32 v3, 16, v2
; GFX7-NEXT: v_add_i32_e32 v0, vcc, v0, v2
; GFX7-NEXT: v_add_i32_e32 v1, vcc, v1, v3
; GFX7-NEXT: s_setpc_b64 s[30:31]
;
; GFX9-LABEL: v_add_v2i16_fneg_rhs:
; GFX9: ; %bb.0:
; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX9-NEXT: v_pk_add_u16 v0, v0, v1 neg_lo:[0,1] neg_hi:[0,1]
; GFX9-NEXT: s_setpc_b64 s[30:31]
;
; GFX8-LABEL: v_add_v2i16_fneg_rhs:
; GFX8: ; %bb.0:
; GFX8-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX8-NEXT: v_xor_b32_e32 v1, 0x80008000, v1
; GFX8-NEXT: v_add_u16_e32 v2, v0, v1
; GFX8-NEXT: v_add_u16_sdwa v0, v0, v1 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_or_b32_e32 v0, v2, v0
; GFX8-NEXT: s_setpc_b64 s[30:31]
;
; GFX10-LABEL: v_add_v2i16_fneg_rhs:
; GFX10: ; %bb.0:
; GFX10-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX10-NEXT: v_pk_add_u16 v0, v0, v1 neg_lo:[0,1] neg_hi:[0,1]
; GFX10-NEXT: s_setpc_b64 s[30:31]
%neg.b = fneg <2 x half> %b
%cast.neg.b = bitcast <2 x half> %neg.b to <2 x i16>
%add = add <2 x i16> %a, %cast.neg.b
ret <2 x i16> %add
}
define <2 x i16> @v_add_v2i16_fneg_lhs_fneg_rhs(<2 x half> %a, <2 x half> %b) {
; GFX7-LABEL: v_add_v2i16_fneg_lhs_fneg_rhs:
; GFX7: ; %bb.0:
; GFX7-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX7-NEXT: v_lshlrev_b32_e32 v1, 16, v1
; GFX7-NEXT: v_and_b32_e32 v0, 0xffff, v0
; GFX7-NEXT: v_or_b32_e32 v0, v1, v0
; GFX7-NEXT: v_lshlrev_b32_e32 v1, 16, v3
; GFX7-NEXT: v_and_b32_e32 v2, 0xffff, v2
; GFX7-NEXT: v_or_b32_e32 v1, v1, v2
; GFX7-NEXT: v_xor_b32_e32 v0, 0x80008000, v0
; GFX7-NEXT: v_xor_b32_e32 v1, 0x80008000, v1
; GFX7-NEXT: v_lshrrev_b32_e32 v2, 16, v0
; GFX7-NEXT: v_lshrrev_b32_e32 v3, 16, v1
; GFX7-NEXT: v_add_i32_e32 v0, vcc, v0, v1
; GFX7-NEXT: v_add_i32_e32 v1, vcc, v2, v3
; GFX7-NEXT: s_setpc_b64 s[30:31]
;
; GFX9-LABEL: v_add_v2i16_fneg_lhs_fneg_rhs:
; GFX9: ; %bb.0:
; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX9-NEXT: v_pk_add_u16 v0, v0, v1 neg_lo:[1,1] neg_hi:[1,1]
; GFX9-NEXT: s_setpc_b64 s[30:31]
;
; GFX8-LABEL: v_add_v2i16_fneg_lhs_fneg_rhs:
; GFX8: ; %bb.0:
; GFX8-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX8-NEXT: v_xor_b32_e32 v0, 0x80008000, v0
; GFX8-NEXT: v_xor_b32_e32 v1, 0x80008000, v1
; GFX8-NEXT: v_add_u16_e32 v2, v0, v1
; GFX8-NEXT: v_add_u16_sdwa v0, v0, v1 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1
; GFX8-NEXT: v_or_b32_e32 v0, v2, v0
; GFX8-NEXT: s_setpc_b64 s[30:31]
;
; GFX10-LABEL: v_add_v2i16_fneg_lhs_fneg_rhs:
; GFX10: ; %bb.0:
; GFX10-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX10-NEXT: v_pk_add_u16 v0, v0, v1 neg_lo:[1,1] neg_hi:[1,1]
; GFX10-NEXT: s_setpc_b64 s[30:31]
%neg.a = fneg <2 x half> %a
%neg.b = fneg <2 x half> %b
%cast.neg.a = bitcast <2 x half> %neg.a to <2 x i16>
%cast.neg.b = bitcast <2 x half> %neg.b to <2 x i16>
%add = add <2 x i16> %cast.neg.a, %cast.neg.b
ret <2 x i16> %add
}
define <2 x i16> @v_add_v2i16_neg_inline_imm_splat(<2 x i16> %a) {
; GFX7-LABEL: v_add_v2i16_neg_inline_imm_splat:
; GFX7: ; %bb.0:
; GFX7-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX7-NEXT: v_add_i32_e32 v0, vcc, 0xffffffc0, v0
; GFX7-NEXT: v_add_i32_e32 v1, vcc, 0xffffffc0, v1
; GFX7-NEXT: s_setpc_b64 s[30:31]
;
; GFX9-LABEL: v_add_v2i16_neg_inline_imm_splat:
; GFX9: ; %bb.0:
; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX9-NEXT: v_pk_sub_u16 v0, v0, 64 op_sel_hi:[1,0]
; GFX9-NEXT: s_setpc_b64 s[30:31]
;
; GFX8-LABEL: v_add_v2i16_neg_inline_imm_splat:
; GFX8: ; %bb.0:
; GFX8-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX8-NEXT: v_not_b32_e32 v1, 63
; GFX8-NEXT: v_add_u16_e32 v2, 0xffc0, v0
; GFX8-NEXT: v_add_u16_sdwa v0, v0, v1 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:DWORD
; GFX8-NEXT: v_or_b32_e32 v0, v2, v0
; GFX8-NEXT: s_setpc_b64 s[30:31]
;
; GFX10-LABEL: v_add_v2i16_neg_inline_imm_splat:
; GFX10: ; %bb.0:
; GFX10-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX10-NEXT: v_pk_sub_u16 v0, v0, 64 op_sel_hi:[1,0]
; GFX10-NEXT: s_setpc_b64 s[30:31]
%add = add <2 x i16> %a, <i16 -64, i16 -64>
ret <2 x i16> %add
}
define <2 x i16> @v_add_v2i16_neg_inline_imm_lo(<2 x i16> %a) {
; GFX7-LABEL: v_add_v2i16_neg_inline_imm_lo:
; GFX7: ; %bb.0:
; GFX7-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX7-NEXT: v_add_i32_e32 v0, vcc, 0xffffffc0, v0
; GFX7-NEXT: v_add_i32_e32 v1, vcc, 4, v1
; GFX7-NEXT: s_setpc_b64 s[30:31]
;
; GFX9-LABEL: v_add_v2i16_neg_inline_imm_lo:
; GFX9: ; %bb.0:
; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX9-NEXT: v_mov_b32_e32 v1, 0x4ffc0
; GFX9-NEXT: v_pk_add_u16 v0, v0, v1
; GFX9-NEXT: s_setpc_b64 s[30:31]
;
; GFX8-LABEL: v_add_v2i16_neg_inline_imm_lo:
; GFX8: ; %bb.0:
; GFX8-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX8-NEXT: v_mov_b32_e32 v2, 4
; GFX8-NEXT: v_add_u16_e32 v1, 0xffc0, v0
; GFX8-NEXT: v_add_u16_sdwa v0, v0, v2 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:DWORD
; GFX8-NEXT: v_or_b32_e32 v0, v1, v0
; GFX8-NEXT: s_setpc_b64 s[30:31]
;
; GFX10-LABEL: v_add_v2i16_neg_inline_imm_lo:
; GFX10: ; %bb.0:
; GFX10-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX10-NEXT: v_pk_add_u16 v0, 0x4ffc0, v0
; GFX10-NEXT: s_setpc_b64 s[30:31]
%add = add <2 x i16> %a, <i16 -64, i16 4>
ret <2 x i16> %add
}
define <2 x i16> @v_add_v2i16_neg_inline_imm_hi(<2 x i16> %a) {
; GFX7-LABEL: v_add_v2i16_neg_inline_imm_hi:
; GFX7: ; %bb.0:
; GFX7-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX7-NEXT: v_add_i32_e32 v0, vcc, 4, v0
; GFX7-NEXT: v_add_i32_e32 v1, vcc, 0xffffffc0, v1
; GFX7-NEXT: s_setpc_b64 s[30:31]
;
; GFX9-LABEL: v_add_v2i16_neg_inline_imm_hi:
; GFX9: ; %bb.0:
; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX9-NEXT: v_mov_b32_e32 v1, 0xffc00004
; GFX9-NEXT: v_pk_add_u16 v0, v0, v1
; GFX9-NEXT: s_setpc_b64 s[30:31]
;
; GFX8-LABEL: v_add_v2i16_neg_inline_imm_hi:
; GFX8: ; %bb.0:
; GFX8-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX8-NEXT: v_not_b32_e32 v1, 63
; GFX8-NEXT: v_add_u16_e32 v2, 4, v0
; GFX8-NEXT: v_add_u16_sdwa v0, v0, v1 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:DWORD
; GFX8-NEXT: v_or_b32_e32 v0, v2, v0
; GFX8-NEXT: s_setpc_b64 s[30:31]
;
; GFX10-LABEL: v_add_v2i16_neg_inline_imm_hi:
; GFX10: ; %bb.0:
; GFX10-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX10-NEXT: v_pk_add_u16 v0, 0xffc00004, v0
; GFX10-NEXT: s_setpc_b64 s[30:31]
%add = add <2 x i16> %a, <i16 4, i16 -64>
ret <2 x i16> %add
}
define amdgpu_ps i32 @s_add_v2i16_neg_inline_imm_splat(<2 x i16> inreg %a) {
; GFX7-LABEL: s_add_v2i16_neg_inline_imm_splat:
; GFX7: ; %bb.0:
; GFX7-NEXT: s_sub_i32 s1, s1, 64
; GFX7-NEXT: s_sub_i32 s0, s0, 64
; GFX7-NEXT: s_and_b32 s1, s1, 0xffff
; GFX7-NEXT: s_and_b32 s0, s0, 0xffff
; GFX7-NEXT: s_lshl_b32 s1, s1, 16
; GFX7-NEXT: s_or_b32 s0, s0, s1
; GFX7-NEXT: ; return to shader part epilog
;
; GFX9-LABEL: s_add_v2i16_neg_inline_imm_splat:
; GFX9: ; %bb.0:
; GFX9-NEXT: s_lshr_b32 s1, s0, 16
; GFX9-NEXT: s_add_i32 s0, s0, 0xffc0ffc0
; GFX9-NEXT: s_add_i32 s1, s1, 0xffc0
; GFX9-NEXT: s_pack_ll_b32_b16 s0, s0, s1
; GFX9-NEXT: ; return to shader part epilog
;
; GFX8-LABEL: s_add_v2i16_neg_inline_imm_splat:
; GFX8: ; %bb.0:
; GFX8-NEXT: s_lshr_b32 s1, s0, 16
; GFX8-NEXT: s_and_b32 s0, s0, 0xffff
; GFX8-NEXT: s_add_i32 s0, s0, 0xffc0
; GFX8-NEXT: s_add_i32 s1, s1, 0xffc0
; GFX8-NEXT: s_lshl_b32 s1, s1, 16
; GFX8-NEXT: s_and_b32 s0, s0, 0xffff
; GFX8-NEXT: s_or_b32 s0, s1, s0
; GFX8-NEXT: ; return to shader part epilog
;
; GFX10-LABEL: s_add_v2i16_neg_inline_imm_splat:
; GFX10: ; %bb.0:
; GFX10-NEXT: s_lshr_b32 s1, s0, 16
; GFX10-NEXT: s_add_i32 s0, s0, 0xffc0ffc0
; GFX10-NEXT: s_add_i32 s1, s1, 0xffc0
; GFX10-NEXT: s_pack_ll_b32_b16 s0, s0, s1
; GFX10-NEXT: ; return to shader part epilog
%add = add <2 x i16> %a, <i16 -64, i16 -64>
%cast = bitcast <2 x i16> %add to i32
ret i32 %cast
}
define amdgpu_ps i32 @s_add_v2i16_neg_inline_imm_lo(<2 x i16> inreg %a) {
; GFX7-LABEL: s_add_v2i16_neg_inline_imm_lo:
; GFX7: ; %bb.0:
; GFX7-NEXT: s_add_i32 s1, s1, 4
; GFX7-NEXT: s_sub_i32 s0, s0, 64
; GFX7-NEXT: s_and_b32 s1, s1, 0xffff
; GFX7-NEXT: s_and_b32 s0, s0, 0xffff
; GFX7-NEXT: s_lshl_b32 s1, s1, 16
; GFX7-NEXT: s_or_b32 s0, s0, s1
; GFX7-NEXT: ; return to shader part epilog
;
; GFX9-LABEL: s_add_v2i16_neg_inline_imm_lo:
; GFX9: ; %bb.0:
; GFX9-NEXT: s_lshr_b32 s1, s0, 16
; GFX9-NEXT: s_add_i32 s0, s0, 0x4ffc0
; GFX9-NEXT: s_add_i32 s1, s1, 4
; GFX9-NEXT: s_pack_ll_b32_b16 s0, s0, s1
; GFX9-NEXT: ; return to shader part epilog
;
; GFX8-LABEL: s_add_v2i16_neg_inline_imm_lo:
; GFX8: ; %bb.0:
; GFX8-NEXT: s_lshr_b32 s1, s0, 16
; GFX8-NEXT: s_and_b32 s0, s0, 0xffff
; GFX8-NEXT: s_add_i32 s0, s0, 0xffc0
; GFX8-NEXT: s_add_i32 s1, s1, 4
; GFX8-NEXT: s_lshl_b32 s1, s1, 16
; GFX8-NEXT: s_and_b32 s0, s0, 0xffff
; GFX8-NEXT: s_or_b32 s0, s1, s0
; GFX8-NEXT: ; return to shader part epilog
;
; GFX10-LABEL: s_add_v2i16_neg_inline_imm_lo:
; GFX10: ; %bb.0:
; GFX10-NEXT: s_lshr_b32 s1, s0, 16
; GFX10-NEXT: s_add_i32 s0, s0, 0x4ffc0
; GFX10-NEXT: s_add_i32 s1, s1, 4
; GFX10-NEXT: s_pack_ll_b32_b16 s0, s0, s1
; GFX10-NEXT: ; return to shader part epilog
%add = add <2 x i16> %a, <i16 -64, i16 4>
%cast = bitcast <2 x i16> %add to i32
ret i32 %cast
}
define amdgpu_ps i32 @s_add_v2i16_neg_inline_imm_hi(<2 x i16> inreg %a) {
; GFX7-LABEL: s_add_v2i16_neg_inline_imm_hi:
; GFX7: ; %bb.0:
; GFX7-NEXT: s_sub_i32 s1, s1, 64
; GFX7-NEXT: s_add_i32 s0, s0, 4
; GFX7-NEXT: s_and_b32 s1, s1, 0xffff
; GFX7-NEXT: s_and_b32 s0, s0, 0xffff
; GFX7-NEXT: s_lshl_b32 s1, s1, 16
; GFX7-NEXT: s_or_b32 s0, s0, s1
; GFX7-NEXT: ; return to shader part epilog
;
; GFX9-LABEL: s_add_v2i16_neg_inline_imm_hi:
; GFX9: ; %bb.0:
; GFX9-NEXT: s_lshr_b32 s1, s0, 16
; GFX9-NEXT: s_add_i32 s0, s0, 0xffc00004
; GFX9-NEXT: s_add_i32 s1, s1, 0xffc0
; GFX9-NEXT: s_pack_ll_b32_b16 s0, s0, s1
; GFX9-NEXT: ; return to shader part epilog
;
; GFX8-LABEL: s_add_v2i16_neg_inline_imm_hi:
; GFX8: ; %bb.0:
; GFX8-NEXT: s_lshr_b32 s1, s0, 16
; GFX8-NEXT: s_and_b32 s0, s0, 0xffff
; GFX8-NEXT: s_add_i32 s0, s0, 4
; GFX8-NEXT: s_add_i32 s1, s1, 0xffc0
; GFX8-NEXT: s_lshl_b32 s1, s1, 16
; GFX8-NEXT: s_and_b32 s0, s0, 0xffff
; GFX8-NEXT: s_or_b32 s0, s1, s0
; GFX8-NEXT: ; return to shader part epilog
;
; GFX10-LABEL: s_add_v2i16_neg_inline_imm_hi:
; GFX10: ; %bb.0:
; GFX10-NEXT: s_lshr_b32 s1, s0, 16
; GFX10-NEXT: s_add_i32 s0, s0, 0xffc00004
; GFX10-NEXT: s_add_i32 s1, s1, 0xffc0
; GFX10-NEXT: s_pack_ll_b32_b16 s0, s0, s1
; GFX10-NEXT: ; return to shader part epilog
%add = add <2 x i16> %a, <i16 4, i16 -64>
%cast = bitcast <2 x i16> %add to i32
ret i32 %cast
}
define amdgpu_ps i32 @s_add_v2i16(<2 x i16> inreg %a, <2 x i16> inreg %b) {
; GFX7-LABEL: s_add_v2i16:
; GFX7: ; %bb.0:
; GFX7-NEXT: s_add_i32 s1, s1, s3
; GFX7-NEXT: s_add_i32 s0, s0, s2
; GFX7-NEXT: s_and_b32 s1, s1, 0xffff
; GFX7-NEXT: s_and_b32 s0, s0, 0xffff
; GFX7-NEXT: s_lshl_b32 s1, s1, 16
; GFX7-NEXT: s_or_b32 s0, s0, s1
; GFX7-NEXT: ; return to shader part epilog
;
; GFX9-LABEL: s_add_v2i16:
; GFX9: ; %bb.0:
; GFX9-NEXT: s_lshr_b32 s2, s0, 16
; GFX9-NEXT: s_lshr_b32 s3, s1, 16
; GFX9-NEXT: s_add_i32 s0, s0, s1
; GFX9-NEXT: s_add_i32 s2, s2, s3
; GFX9-NEXT: s_pack_ll_b32_b16 s0, s0, s2
; GFX9-NEXT: ; return to shader part epilog
;
; GFX8-LABEL: s_add_v2i16:
; GFX8: ; %bb.0:
; GFX8-NEXT: s_lshr_b32 s2, s0, 16
; GFX8-NEXT: s_and_b32 s0, s0, 0xffff
; GFX8-NEXT: s_lshr_b32 s3, s1, 16
; GFX8-NEXT: s_and_b32 s1, s1, 0xffff
; GFX8-NEXT: s_add_i32 s0, s0, s1
; GFX8-NEXT: s_add_i32 s2, s2, s3
; GFX8-NEXT: s_lshl_b32 s1, s2, 16
; GFX8-NEXT: s_and_b32 s0, s0, 0xffff
; GFX8-NEXT: s_or_b32 s0, s1, s0
; GFX8-NEXT: ; return to shader part epilog
;
; GFX10-LABEL: s_add_v2i16:
; GFX10: ; %bb.0:
; GFX10-NEXT: s_lshr_b32 s2, s0, 16
; GFX10-NEXT: s_lshr_b32 s3, s1, 16
; GFX10-NEXT: s_add_i32 s0, s0, s1
; GFX10-NEXT: s_add_i32 s2, s2, s3
; GFX10-NEXT: s_pack_ll_b32_b16 s0, s0, s2
; GFX10-NEXT: ; return to shader part epilog
%add = add <2 x i16> %a, %b
%cast = bitcast <2 x i16> %add to i32
ret i32 %cast
}
define amdgpu_ps i32 @s_add_v2i16_fneg_lhs(<2 x half> inreg %a, <2 x i16> inreg %b) {
; GFX7-LABEL: s_add_v2i16_fneg_lhs:
; GFX7: ; %bb.0:
; GFX7-NEXT: s_lshl_b32 s1, s1, 16
; GFX7-NEXT: s_and_b32 s0, s0, 0xffff
; GFX7-NEXT: s_or_b32 s0, s1, s0
; GFX7-NEXT: s_xor_b32 s0, s0, 0x80008000
; GFX7-NEXT: s_lshr_b32 s1, s0, 16
; GFX7-NEXT: s_add_i32 s1, s1, s3
; GFX7-NEXT: s_add_i32 s0, s0, s2
; GFX7-NEXT: s_and_b32 s1, s1, 0xffff
; GFX7-NEXT: s_and_b32 s0, s0, 0xffff
; GFX7-NEXT: s_lshl_b32 s1, s1, 16
; GFX7-NEXT: s_or_b32 s0, s0, s1
; GFX7-NEXT: ; return to shader part epilog
;
; GFX9-LABEL: s_add_v2i16_fneg_lhs:
; GFX9: ; %bb.0:
; GFX9-NEXT: s_xor_b32 s0, s0, 0x80008000
; GFX9-NEXT: s_lshr_b32 s2, s0, 16
; GFX9-NEXT: s_lshr_b32 s3, s1, 16
; GFX9-NEXT: s_add_i32 s0, s0, s1
; GFX9-NEXT: s_add_i32 s2, s2, s3
; GFX9-NEXT: s_pack_ll_b32_b16 s0, s0, s2
; GFX9-NEXT: ; return to shader part epilog
;
; GFX8-LABEL: s_add_v2i16_fneg_lhs:
; GFX8: ; %bb.0:
; GFX8-NEXT: s_xor_b32 s0, s0, 0x80008000
; GFX8-NEXT: s_lshr_b32 s2, s0, 16
; GFX8-NEXT: s_and_b32 s0, s0, 0xffff
; GFX8-NEXT: s_lshr_b32 s3, s1, 16
; GFX8-NEXT: s_and_b32 s1, s1, 0xffff
; GFX8-NEXT: s_add_i32 s0, s0, s1
; GFX8-NEXT: s_add_i32 s2, s2, s3
; GFX8-NEXT: s_lshl_b32 s1, s2, 16
; GFX8-NEXT: s_and_b32 s0, s0, 0xffff
; GFX8-NEXT: s_or_b32 s0, s1, s0
; GFX8-NEXT: ; return to shader part epilog
;
; GFX10-LABEL: s_add_v2i16_fneg_lhs:
; GFX10: ; %bb.0:
; GFX10-NEXT: s_xor_b32 s0, s0, 0x80008000
; GFX10-NEXT: s_lshr_b32 s3, s1, 16
; GFX10-NEXT: s_lshr_b32 s2, s0, 16
; GFX10-NEXT: s_add_i32 s0, s0, s1
; GFX10-NEXT: s_add_i32 s2, s2, s3
; GFX10-NEXT: s_pack_ll_b32_b16 s0, s0, s2
; GFX10-NEXT: ; return to shader part epilog
%neg.a = fneg <2 x half> %a
%cast.neg.a = bitcast <2 x half> %neg.a to <2 x i16>
%add = add <2 x i16> %cast.neg.a, %b
%cast = bitcast <2 x i16> %add to i32
ret i32 %cast
}
define amdgpu_ps i32 @s_add_v2i16_fneg_rhs(<2 x i16> inreg %a, <2 x half> inreg %b) {
; GFX7-LABEL: s_add_v2i16_fneg_rhs:
; GFX7: ; %bb.0:
; GFX7-NEXT: s_lshl_b32 s3, s3, 16
; GFX7-NEXT: s_and_b32 s2, s2, 0xffff
; GFX7-NEXT: s_or_b32 s2, s3, s2
; GFX7-NEXT: s_xor_b32 s2, s2, 0x80008000
; GFX7-NEXT: s_lshr_b32 s3, s2, 16
; GFX7-NEXT: s_add_i32 s1, s1, s3
; GFX7-NEXT: s_add_i32 s0, s0, s2
; GFX7-NEXT: s_and_b32 s1, s1, 0xffff
; GFX7-NEXT: s_and_b32 s0, s0, 0xffff
; GFX7-NEXT: s_lshl_b32 s1, s1, 16
; GFX7-NEXT: s_or_b32 s0, s0, s1
; GFX7-NEXT: ; return to shader part epilog
;
; GFX9-LABEL: s_add_v2i16_fneg_rhs:
; GFX9: ; %bb.0:
; GFX9-NEXT: s_xor_b32 s1, s1, 0x80008000
; GFX9-NEXT: s_lshr_b32 s2, s0, 16
; GFX9-NEXT: s_lshr_b32 s3, s1, 16
; GFX9-NEXT: s_add_i32 s0, s0, s1
; GFX9-NEXT: s_add_i32 s2, s2, s3
; GFX9-NEXT: s_pack_ll_b32_b16 s0, s0, s2
; GFX9-NEXT: ; return to shader part epilog
;
; GFX8-LABEL: s_add_v2i16_fneg_rhs:
; GFX8: ; %bb.0:
; GFX8-NEXT: s_xor_b32 s1, s1, 0x80008000
; GFX8-NEXT: s_lshr_b32 s2, s0, 16
; GFX8-NEXT: s_and_b32 s0, s0, 0xffff
; GFX8-NEXT: s_lshr_b32 s3, s1, 16
; GFX8-NEXT: s_and_b32 s1, s1, 0xffff
; GFX8-NEXT: s_add_i32 s0, s0, s1
; GFX8-NEXT: s_add_i32 s2, s2, s3
; GFX8-NEXT: s_lshl_b32 s1, s2, 16
; GFX8-NEXT: s_and_b32 s0, s0, 0xffff
; GFX8-NEXT: s_or_b32 s0, s1, s0
; GFX8-NEXT: ; return to shader part epilog
;
; GFX10-LABEL: s_add_v2i16_fneg_rhs:
; GFX10: ; %bb.0:
; GFX10-NEXT: s_xor_b32 s1, s1, 0x80008000
; GFX10-NEXT: s_lshr_b32 s2, s0, 16
; GFX10-NEXT: s_lshr_b32 s3, s1, 16
; GFX10-NEXT: s_add_i32 s0, s0, s1
; GFX10-NEXT: s_add_i32 s2, s2, s3
; GFX10-NEXT: s_pack_ll_b32_b16 s0, s0, s2
; GFX10-NEXT: ; return to shader part epilog
%neg.b = fneg <2 x half> %b
%cast.neg.b = bitcast <2 x half> %neg.b to <2 x i16>
%add = add <2 x i16> %a, %cast.neg.b
%cast = bitcast <2 x i16> %add to i32
ret i32 %cast
}
define amdgpu_ps i32 @s_add_v2i16_fneg_lhs_fneg_rhs(<2 x half> inreg %a, <2 x half> inreg %b) {
; GFX7-LABEL: s_add_v2i16_fneg_lhs_fneg_rhs:
; GFX7: ; %bb.0:
; GFX7-NEXT: s_lshl_b32 s1, s1, 16
; GFX7-NEXT: s_and_b32 s0, s0, 0xffff
; GFX7-NEXT: s_or_b32 s0, s1, s0
; GFX7-NEXT: s_lshl_b32 s1, s3, 16
; GFX7-NEXT: s_and_b32 s2, s2, 0xffff
; GFX7-NEXT: s_or_b32 s1, s1, s2
; GFX7-NEXT: s_xor_b32 s0, s0, 0x80008000
; GFX7-NEXT: s_xor_b32 s1, s1, 0x80008000
; GFX7-NEXT: s_lshr_b32 s2, s0, 16
; GFX7-NEXT: s_lshr_b32 s3, s1, 16
; GFX7-NEXT: s_add_i32 s2, s2, s3
; GFX7-NEXT: s_add_i32 s0, s0, s1
; GFX7-NEXT: s_and_b32 s1, s2, 0xffff
; GFX7-NEXT: s_and_b32 s0, s0, 0xffff
; GFX7-NEXT: s_lshl_b32 s1, s1, 16
; GFX7-NEXT: s_or_b32 s0, s0, s1
; GFX7-NEXT: ; return to shader part epilog
;
; GFX9-LABEL: s_add_v2i16_fneg_lhs_fneg_rhs:
; GFX9: ; %bb.0:
; GFX9-NEXT: s_xor_b32 s0, s0, 0x80008000
; GFX9-NEXT: s_xor_b32 s1, s1, 0x80008000
; GFX9-NEXT: s_lshr_b32 s2, s0, 16
; GFX9-NEXT: s_lshr_b32 s3, s1, 16
; GFX9-NEXT: s_add_i32 s0, s0, s1
; GFX9-NEXT: s_add_i32 s2, s2, s3
; GFX9-NEXT: s_pack_ll_b32_b16 s0, s0, s2
; GFX9-NEXT: ; return to shader part epilog
;
; GFX8-LABEL: s_add_v2i16_fneg_lhs_fneg_rhs:
; GFX8: ; %bb.0:
; GFX8-NEXT: s_xor_b32 s0, s0, 0x80008000
; GFX8-NEXT: s_xor_b32 s1, s1, 0x80008000
; GFX8-NEXT: s_lshr_b32 s2, s0, 16
; GFX8-NEXT: s_and_b32 s0, s0, 0xffff
; GFX8-NEXT: s_lshr_b32 s3, s1, 16
; GFX8-NEXT: s_and_b32 s1, s1, 0xffff
; GFX8-NEXT: s_add_i32 s0, s0, s1
; GFX8-NEXT: s_add_i32 s2, s2, s3
; GFX8-NEXT: s_lshl_b32 s1, s2, 16
; GFX8-NEXT: s_and_b32 s0, s0, 0xffff
; GFX8-NEXT: s_or_b32 s0, s1, s0
; GFX8-NEXT: ; return to shader part epilog
;
; GFX10-LABEL: s_add_v2i16_fneg_lhs_fneg_rhs:
; GFX10: ; %bb.0:
; GFX10-NEXT: s_xor_b32 s0, s0, 0x80008000
; GFX10-NEXT: s_xor_b32 s1, s1, 0x80008000
; GFX10-NEXT: s_lshr_b32 s2, s0, 16
; GFX10-NEXT: s_lshr_b32 s3, s1, 16
; GFX10-NEXT: s_add_i32 s0, s0, s1
; GFX10-NEXT: s_add_i32 s2, s2, s3
; GFX10-NEXT: s_pack_ll_b32_b16 s0, s0, s2
; GFX10-NEXT: ; return to shader part epilog
%neg.a = fneg <2 x half> %a
%neg.b = fneg <2 x half> %b
%cast.neg.a = bitcast <2 x half> %neg.a to <2 x i16>
%cast.neg.b = bitcast <2 x half> %neg.b to <2 x i16>
%add = add <2 x i16> %cast.neg.a, %cast.neg.b
%cast = bitcast <2 x i16> %add to i32
ret i32 %cast
}
define <2 x i16> @add_inline_imm_neg1_0(<2 x i16> %x) {
; GFX7-LABEL: add_inline_imm_neg1_0:
; GFX7: ; %bb.0:
; GFX7-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX7-NEXT: v_add_i32_e32 v0, vcc, -1, v0
; GFX7-NEXT: s_setpc_b64 s[30:31]
;
; GFX9-LABEL: add_inline_imm_neg1_0:
; GFX9: ; %bb.0:
; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX9-NEXT: v_pk_sub_u16 v0, v0, 1
; GFX9-NEXT: s_setpc_b64 s[30:31]
;
; GFX8-LABEL: add_inline_imm_neg1_0:
; GFX8: ; %bb.0:
; GFX8-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX8-NEXT: v_lshrrev_b32_e32 v1, 16, v0
; GFX8-NEXT: v_add_u16_e32 v0, -1, v0
; GFX8-NEXT: v_lshlrev_b32_e32 v1, 16, v1
; GFX8-NEXT: v_or_b32_e32 v0, v0, v1
; GFX8-NEXT: s_setpc_b64 s[30:31]
;
; GFX10-LABEL: add_inline_imm_neg1_0:
; GFX10: ; %bb.0:
; GFX10-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX10-NEXT: v_pk_sub_u16 v0, v0, 1
; GFX10-NEXT: s_setpc_b64 s[30:31]
%y = add <2 x i16> %x, <i16 -1, i16 0>
ret <2 x i16> %y
}
define <2 x i16> @add_inline_imm_1_0(<2 x i16> %x) {
; GFX7-LABEL: add_inline_imm_1_0:
; GFX7: ; %bb.0:
; GFX7-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX7-NEXT: v_add_i32_e32 v0, vcc, 1, v0
; GFX7-NEXT: s_setpc_b64 s[30:31]
;
; GFX9-LABEL: add_inline_imm_1_0:
; GFX9: ; %bb.0:
; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX9-NEXT: v_pk_add_u16 v0, v0, 1
; GFX9-NEXT: s_setpc_b64 s[30:31]
;
; GFX8-LABEL: add_inline_imm_1_0:
; GFX8: ; %bb.0:
; GFX8-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX8-NEXT: v_lshrrev_b32_e32 v1, 16, v0
; GFX8-NEXT: v_add_u16_e32 v0, 1, v0
; GFX8-NEXT: v_lshlrev_b32_e32 v1, 16, v1
; GFX8-NEXT: v_or_b32_e32 v0, v0, v1
; GFX8-NEXT: s_setpc_b64 s[30:31]
;
; GFX10-LABEL: add_inline_imm_1_0:
; GFX10: ; %bb.0:
; GFX10-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX10-NEXT: v_pk_add_u16 v0, v0, 1
; GFX10-NEXT: s_setpc_b64 s[30:31]
%y = add <2 x i16> %x, <i16 1, i16 0>
ret <2 x i16> %y
}