Please use GitHub pull requests for new patches. Phabricator shutdown timeline
Changeset View
Changeset View
Standalone View
Standalone View
llvm/trunk/test/CodeGen/AMDGPU/idot2.ll
Show First 20 Lines • Show All 2,172 Lines • ▼ Show 20 Lines | |||||
; GFX7-NEXT: v_mad_u32_u24 v0, s4, v1, v0 | ; GFX7-NEXT: v_mad_u32_u24 v0, s4, v1, v0 | ||||
; GFX7-NEXT: buffer_store_short v0, off, s[0:3], 0 | ; GFX7-NEXT: buffer_store_short v0, off, s[0:3], 0 | ||||
; GFX7-NEXT: s_endpgm | ; GFX7-NEXT: s_endpgm | ||||
; | ; | ||||
; GFX8-LABEL: udot2_acc16: | ; GFX8-LABEL: udot2_acc16: | ||||
; GFX8: ; %bb.0: ; %entry | ; GFX8: ; %bb.0: ; %entry | ||||
; GFX8-NEXT: s_load_dwordx4 s[4:7], s[0:1], 0x24 | ; GFX8-NEXT: s_load_dwordx4 s[4:7], s[0:1], 0x24 | ||||
; GFX8-NEXT: s_load_dwordx2 s[0:1], s[0:1], 0x34 | ; GFX8-NEXT: s_load_dwordx2 s[0:1], s[0:1], 0x34 | ||||
; GFX8-NEXT: s_mov_b32 s2, 0xffff | |||||
; GFX8-NEXT: s_waitcnt lgkmcnt(0) | ; GFX8-NEXT: s_waitcnt lgkmcnt(0) | ||||
; GFX8-NEXT: v_mov_b32_e32 v0, s0 | ; GFX8-NEXT: v_mov_b32_e32 v0, s0 | ||||
; GFX8-NEXT: v_mov_b32_e32 v1, s1 | ; GFX8-NEXT: v_mov_b32_e32 v1, s1 | ||||
; GFX8-NEXT: flat_load_ushort v2, v[0:1] | ; GFX8-NEXT: flat_load_ushort v2, v[0:1] | ||||
; GFX8-NEXT: s_load_dword s1, s[4:5], 0x0 | ; GFX8-NEXT: s_load_dword s0, s[4:5], 0x0 | ||||
; GFX8-NEXT: s_load_dword s2, s[6:7], 0x0 | ; GFX8-NEXT: s_load_dword s1, s[6:7], 0x0 | ||||
; GFX8-NEXT: s_mov_b32 s0, 0xffff | |||||
; GFX8-NEXT: s_waitcnt lgkmcnt(0) | ; GFX8-NEXT: s_waitcnt lgkmcnt(0) | ||||
; GFX8-NEXT: s_and_b32 s3, s1, s0 | ; GFX8-NEXT: s_and_b32 s3, s1, s2 | ||||
; GFX8-NEXT: s_and_b32 s0, s2, s0 | |||||
; GFX8-NEXT: s_lshr_b32 s2, s2, 16 | |||||
; GFX8-NEXT: s_lshr_b32 s1, s1, 16 | ; GFX8-NEXT: s_lshr_b32 s1, s1, 16 | ||||
; GFX8-NEXT: v_mov_b32_e32 v3, s2 | ; GFX8-NEXT: s_and_b32 s2, s0, s2 | ||||
; GFX8-NEXT: s_lshr_b32 s0, s0, 16 | |||||
; GFX8-NEXT: v_mov_b32_e32 v3, s1 | |||||
; GFX8-NEXT: s_waitcnt vmcnt(0) | ; GFX8-NEXT: s_waitcnt vmcnt(0) | ||||
; GFX8-NEXT: v_mad_u32_u24 v2, s1, v3, v2 | ; GFX8-NEXT: v_mad_u32_u24 v2, s0, v3, v2 | ||||
; GFX8-NEXT: v_mov_b32_e32 v3, s0 | ; GFX8-NEXT: v_mov_b32_e32 v3, s3 | ||||
; GFX8-NEXT: v_mad_u32_u24 v2, s3, v3, v2 | ; GFX8-NEXT: v_mad_u32_u24 v2, s2, v3, v2 | ||||
; GFX8-NEXT: flat_store_short v[0:1], v2 | ; GFX8-NEXT: flat_store_short v[0:1], v2 | ||||
; GFX8-NEXT: s_endpgm | ; GFX8-NEXT: s_endpgm | ||||
; | ; | ||||
; GFX9-NODL-LABEL: udot2_acc16: | ; GFX9-NODL-LABEL: udot2_acc16: | ||||
; GFX9-NODL: ; %bb.0: ; %entry | ; GFX9-NODL: ; %bb.0: ; %entry | ||||
; GFX9-NODL-NEXT: s_load_dwordx4 s[4:7], s[0:1], 0x24 | ; GFX9-NODL-NEXT: s_load_dwordx4 s[4:7], s[0:1], 0x24 | ||||
; GFX9-NODL-NEXT: s_load_dwordx2 s[0:1], s[0:1], 0x34 | ; GFX9-NODL-NEXT: s_load_dwordx2 s[0:1], s[0:1], 0x34 | ||||
; GFX9-NODL-NEXT: s_mov_b32 s2, 0xffff | |||||
; GFX9-NODL-NEXT: s_waitcnt lgkmcnt(0) | ; GFX9-NODL-NEXT: s_waitcnt lgkmcnt(0) | ||||
; GFX9-NODL-NEXT: v_mov_b32_e32 v0, s0 | ; GFX9-NODL-NEXT: v_mov_b32_e32 v0, s0 | ||||
; GFX9-NODL-NEXT: v_mov_b32_e32 v1, s1 | ; GFX9-NODL-NEXT: v_mov_b32_e32 v1, s1 | ||||
; GFX9-NODL-NEXT: global_load_ushort v2, v[0:1], off | ; GFX9-NODL-NEXT: global_load_ushort v2, v[0:1], off | ||||
; GFX9-NODL-NEXT: s_load_dword s1, s[4:5], 0x0 | ; GFX9-NODL-NEXT: s_load_dword s0, s[4:5], 0x0 | ||||
; GFX9-NODL-NEXT: s_load_dword s2, s[6:7], 0x0 | ; GFX9-NODL-NEXT: s_load_dword s1, s[6:7], 0x0 | ||||
; GFX9-NODL-NEXT: s_mov_b32 s0, 0xffff | |||||
; GFX9-NODL-NEXT: s_waitcnt lgkmcnt(0) | ; GFX9-NODL-NEXT: s_waitcnt lgkmcnt(0) | ||||
; GFX9-NODL-NEXT: s_and_b32 s3, s1, s0 | ; GFX9-NODL-NEXT: s_and_b32 s3, s1, s2 | ||||
; GFX9-NODL-NEXT: s_and_b32 s0, s2, s0 | |||||
; GFX9-NODL-NEXT: s_lshr_b32 s2, s2, 16 | |||||
; GFX9-NODL-NEXT: s_lshr_b32 s1, s1, 16 | ; GFX9-NODL-NEXT: s_lshr_b32 s1, s1, 16 | ||||
; GFX9-NODL-NEXT: v_mov_b32_e32 v3, s2 | ; GFX9-NODL-NEXT: s_and_b32 s2, s0, s2 | ||||
; GFX9-NODL-NEXT: s_lshr_b32 s0, s0, 16 | |||||
; GFX9-NODL-NEXT: v_mov_b32_e32 v3, s1 | |||||
; GFX9-NODL-NEXT: s_waitcnt vmcnt(0) | ; GFX9-NODL-NEXT: s_waitcnt vmcnt(0) | ||||
; GFX9-NODL-NEXT: v_mad_u32_u24 v2, s1, v3, v2 | ; GFX9-NODL-NEXT: v_mad_u32_u24 v2, s0, v3, v2 | ||||
; GFX9-NODL-NEXT: v_mov_b32_e32 v3, s0 | ; GFX9-NODL-NEXT: v_mov_b32_e32 v3, s3 | ||||
; GFX9-NODL-NEXT: v_mad_u32_u24 v2, s3, v3, v2 | ; GFX9-NODL-NEXT: v_mad_u32_u24 v2, s2, v3, v2 | ||||
; GFX9-NODL-NEXT: global_store_short v[0:1], v2, off | ; GFX9-NODL-NEXT: global_store_short v[0:1], v2, off | ||||
; GFX9-NODL-NEXT: s_endpgm | ; GFX9-NODL-NEXT: s_endpgm | ||||
; | ; | ||||
; GFX9-DL-LABEL: udot2_acc16: | ; GFX9-DL-LABEL: udot2_acc16: | ||||
; GFX9-DL: ; %bb.0: ; %entry | ; GFX9-DL: ; %bb.0: ; %entry | ||||
; GFX9-DL-NEXT: s_load_dwordx4 s[4:7], s[0:1], 0x24 | ; GFX9-DL-NEXT: s_load_dwordx4 s[4:7], s[0:1], 0x24 | ||||
; GFX9-DL-NEXT: s_load_dwordx2 s[0:1], s[0:1], 0x34 | ; GFX9-DL-NEXT: s_load_dwordx2 s[0:1], s[0:1], 0x34 | ||||
; GFX9-DL-NEXT: s_waitcnt lgkmcnt(0) | ; GFX9-DL-NEXT: s_waitcnt lgkmcnt(0) | ||||
▲ Show 20 Lines • Show All 167 Lines • Show Last 20 Lines |