diff options
Diffstat (limited to 'llvm/test/CodeGen/AMDGPU/vgpr-tuple-allocation.ll')
| -rw-r--r-- | llvm/test/CodeGen/AMDGPU/vgpr-tuple-allocation.ll | 144 |
1 files changed, 24 insertions, 120 deletions
diff --git a/llvm/test/CodeGen/AMDGPU/vgpr-tuple-allocation.ll b/llvm/test/CodeGen/AMDGPU/vgpr-tuple-allocation.ll index 5164b072a6dd..ed0de729dafd 100644 --- a/llvm/test/CodeGen/AMDGPU/vgpr-tuple-allocation.ll +++ b/llvm/test/CodeGen/AMDGPU/vgpr-tuple-allocation.ll @@ -14,7 +14,6 @@ define <4 x float> @non_preserved_vgpr_tuple8(<8 x i32> %rsrc, <4 x i32> %samp, ; GFX9-NEXT: s_or_saveexec_b64 s[4:5], -1 ; GFX9-NEXT: buffer_store_dword v40, off, s[0:3], s32 offset:16 ; 4-byte Folded Spill ; GFX9-NEXT: s_mov_b64 exec, s[4:5] -; GFX9-NEXT: s_mov_b32 s4, 0 ; GFX9-NEXT: v_writelane_b32 v40, s33, 2 ; GFX9-NEXT: s_mov_b32 s33, s32 ; GFX9-NEXT: v_mov_b32_e32 v36, v16 @@ -22,13 +21,6 @@ define <4 x float> @non_preserved_vgpr_tuple8(<8 x i32> %rsrc, <4 x i32> %samp, ; GFX9-NEXT: v_mov_b32_e32 v34, v14 ; GFX9-NEXT: v_mov_b32_e32 v33, v13 ; GFX9-NEXT: v_mov_b32_e32 v32, v12 -; GFX9-NEXT: s_mov_b32 s5, s4 -; GFX9-NEXT: s_mov_b32 s6, s4 -; GFX9-NEXT: s_mov_b32 s7, s4 -; GFX9-NEXT: s_mov_b32 s8, s4 -; GFX9-NEXT: s_mov_b32 s9, s4 -; GFX9-NEXT: s_mov_b32 s10, s4 -; GFX9-NEXT: s_mov_b32 s11, s4 ; GFX9-NEXT: buffer_store_dword v41, off, s[0:3], s33 offset:12 ; 4-byte Folded Spill ; GFX9-NEXT: buffer_store_dword v42, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill ; GFX9-NEXT: buffer_store_dword v43, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill @@ -82,16 +74,8 @@ define <4 x float> @non_preserved_vgpr_tuple8(<8 x i32> %rsrc, <4 x i32> %samp, ; GFX10-NEXT: v_mov_b32_e32 v34, v14 ; GFX10-NEXT: v_mov_b32_e32 v33, v13 ; GFX10-NEXT: v_mov_b32_e32 v32, v12 -; GFX10-NEXT: s_mov_b32 s4, 0 ; GFX10-NEXT: v_writelane_b32 v40, s33, 2 ; GFX10-NEXT: s_mov_b32 s33, s32 -; GFX10-NEXT: s_mov_b32 s5, s4 -; GFX10-NEXT: s_mov_b32 s6, s4 -; GFX10-NEXT: s_mov_b32 s7, s4 -; GFX10-NEXT: s_mov_b32 s8, s4 -; GFX10-NEXT: s_mov_b32 s9, s4 -; GFX10-NEXT: s_mov_b32 s10, s4 -; GFX10-NEXT: s_mov_b32 s11, s4 ; GFX10-NEXT: buffer_store_dword v41, off, s[0:3], s33 offset:12 ; 4-byte Folded Spill ; GFX10-NEXT: buffer_store_dword v42, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill ; GFX10-NEXT: buffer_store_dword v43, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill @@ -145,16 +129,8 @@ define <4 x float> @non_preserved_vgpr_tuple8(<8 x i32> %rsrc, <4 x i32> %samp, ; GFX11-NEXT: v_dual_mov_b32 v36, v16 :: v_dual_mov_b32 v35, v15 ; GFX11-NEXT: v_dual_mov_b32 v34, v14 :: v_dual_mov_b32 v33, v13 ; GFX11-NEXT: v_mov_b32_e32 v32, v12 -; GFX11-NEXT: s_mov_b32 s0, 0 ; GFX11-NEXT: v_writelane_b32 v40, s33, 2 ; GFX11-NEXT: s_mov_b32 s33, s32 -; GFX11-NEXT: s_mov_b32 s1, s0 -; GFX11-NEXT: s_mov_b32 s2, s0 -; GFX11-NEXT: s_mov_b32 s3, s0 -; GFX11-NEXT: s_mov_b32 s4, s0 -; GFX11-NEXT: s_mov_b32 s5, s0 -; GFX11-NEXT: s_mov_b32 s6, s0 -; GFX11-NEXT: s_mov_b32 s7, s0 ; GFX11-NEXT: s_clause 0x3 ; GFX11-NEXT: scratch_store_b32 off, v41, s33 offset:12 ; GFX11-NEXT: scratch_store_b32 off, v42, s33 offset:8 @@ -225,65 +201,41 @@ define <4 x float> @call_preserved_vgpr_tuple8(<8 x i32> %rsrc, <4 x i32> %samp, ; GFX9-NEXT: s_or_saveexec_b64 s[4:5], -1 ; GFX9-NEXT: buffer_store_dword v40, off, s[0:3], s32 offset:20 ; 4-byte Folded Spill ; GFX9-NEXT: s_mov_b64 exec, s[4:5] -; GFX9-NEXT: v_writelane_b32 v40, s33, 10 -; GFX9-NEXT: v_writelane_b32 v40, s30, 0 -; GFX9-NEXT: v_writelane_b32 v40, s31, 1 -; GFX9-NEXT: v_writelane_b32 v40, s36, 2 -; GFX9-NEXT: v_writelane_b32 v40, s37, 3 -; GFX9-NEXT: v_writelane_b32 v40, s38, 4 -; GFX9-NEXT: v_writelane_b32 v40, s39, 5 -; GFX9-NEXT: v_writelane_b32 v40, s40, 6 -; GFX9-NEXT: v_writelane_b32 v40, s41, 7 +; GFX9-NEXT: v_writelane_b32 v40, s33, 2 ; GFX9-NEXT: s_mov_b32 s33, s32 -; GFX9-NEXT: v_writelane_b32 v40, s42, 8 -; GFX9-NEXT: s_mov_b32 s36, 0 ; GFX9-NEXT: buffer_store_dword v41, off, s[0:3], s33 offset:16 ; 4-byte Folded Spill ; GFX9-NEXT: buffer_store_dword v42, off, s[0:3], s33 offset:12 ; 4-byte Folded Spill ; GFX9-NEXT: buffer_store_dword v43, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill ; GFX9-NEXT: buffer_store_dword v44, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill ; GFX9-NEXT: buffer_store_dword v45, off, s[0:3], s33 ; 4-byte Folded Spill -; GFX9-NEXT: v_writelane_b32 v40, s43, 9 ; GFX9-NEXT: v_mov_b32_e32 v45, v16 ; GFX9-NEXT: v_mov_b32_e32 v44, v15 ; GFX9-NEXT: v_mov_b32_e32 v43, v14 ; GFX9-NEXT: v_mov_b32_e32 v42, v13 ; GFX9-NEXT: v_mov_b32_e32 v41, v12 -; GFX9-NEXT: s_mov_b32 s37, s36 -; GFX9-NEXT: s_mov_b32 s38, s36 -; GFX9-NEXT: s_mov_b32 s39, s36 -; GFX9-NEXT: s_mov_b32 s40, s36 -; GFX9-NEXT: s_mov_b32 s41, s36 -; GFX9-NEXT: s_mov_b32 s42, s36 -; GFX9-NEXT: s_mov_b32 s43, s36 -; GFX9-NEXT: image_gather4_c_b_cl v[0:3], v[41:45], s[36:43], s[4:7] dmask:0x1 +; GFX9-NEXT: image_gather4_c_b_cl v[0:3], v[41:45], s[4:11], s[4:7] dmask:0x1 ; GFX9-NEXT: s_addk_i32 s32, 0x800 ; GFX9-NEXT: s_getpc_b64 s[4:5] ; GFX9-NEXT: s_add_u32 s4, s4, extern_func@gotpcrel32@lo+4 ; GFX9-NEXT: s_addc_u32 s5, s5, extern_func@gotpcrel32@hi+12 ; GFX9-NEXT: s_load_dwordx2 s[4:5], s[4:5], 0x0 +; GFX9-NEXT: v_writelane_b32 v40, s30, 0 +; GFX9-NEXT: v_writelane_b32 v40, s31, 1 ; GFX9-NEXT: s_waitcnt vmcnt(0) ; GFX9-NEXT: global_store_dwordx4 v[0:1], v[0:3], off ; GFX9-NEXT: s_waitcnt lgkmcnt(0) ; GFX9-NEXT: s_swappc_b64 s[30:31], s[4:5] -; GFX9-NEXT: image_gather4_c_b_cl v[0:3], v[41:45], s[36:43], s[4:7] dmask:0x1 +; GFX9-NEXT: image_gather4_c_b_cl v[0:3], v[41:45], s[4:11], s[4:7] dmask:0x1 ; GFX9-NEXT: s_nop 0 ; GFX9-NEXT: buffer_load_dword v45, off, s[0:3], s33 ; 4-byte Folded Reload ; GFX9-NEXT: buffer_load_dword v44, off, s[0:3], s33 offset:4 ; 4-byte Folded Reload ; GFX9-NEXT: buffer_load_dword v43, off, s[0:3], s33 offset:8 ; 4-byte Folded Reload ; GFX9-NEXT: buffer_load_dword v42, off, s[0:3], s33 offset:12 ; 4-byte Folded Reload ; GFX9-NEXT: buffer_load_dword v41, off, s[0:3], s33 offset:16 ; 4-byte Folded Reload -; GFX9-NEXT: v_readlane_b32 s43, v40, 9 -; GFX9-NEXT: v_readlane_b32 s42, v40, 8 -; GFX9-NEXT: v_readlane_b32 s41, v40, 7 -; GFX9-NEXT: v_readlane_b32 s40, v40, 6 -; GFX9-NEXT: v_readlane_b32 s39, v40, 5 -; GFX9-NEXT: v_readlane_b32 s38, v40, 4 -; GFX9-NEXT: v_readlane_b32 s37, v40, 3 -; GFX9-NEXT: v_readlane_b32 s36, v40, 2 ; GFX9-NEXT: v_readlane_b32 s31, v40, 1 ; GFX9-NEXT: v_readlane_b32 s30, v40, 0 ; GFX9-NEXT: s_addk_i32 s32, 0xf800 -; GFX9-NEXT: v_readlane_b32 s33, v40, 10 +; GFX9-NEXT: v_readlane_b32 s33, v40, 2 ; GFX9-NEXT: s_or_saveexec_b64 s[4:5], -1 ; GFX9-NEXT: buffer_load_dword v40, off, s[0:3], s32 offset:20 ; 4-byte Folded Reload ; GFX9-NEXT: s_mov_b64 exec, s[4:5] @@ -298,66 +250,42 @@ define <4 x float> @call_preserved_vgpr_tuple8(<8 x i32> %rsrc, <4 x i32> %samp, ; GFX10-NEXT: buffer_store_dword v40, off, s[0:3], s32 offset:20 ; 4-byte Folded Spill ; GFX10-NEXT: s_waitcnt_depctr 0xffe3 ; GFX10-NEXT: s_mov_b32 exec_lo, s4 -; GFX10-NEXT: v_writelane_b32 v40, s33, 10 +; GFX10-NEXT: v_writelane_b32 v40, s33, 2 ; GFX10-NEXT: s_mov_b32 s33, s32 ; GFX10-NEXT: buffer_store_dword v41, off, s[0:3], s33 offset:16 ; 4-byte Folded Spill ; GFX10-NEXT: buffer_store_dword v42, off, s[0:3], s33 offset:12 ; 4-byte Folded Spill ; GFX10-NEXT: buffer_store_dword v43, off, s[0:3], s33 offset:8 ; 4-byte Folded Spill ; GFX10-NEXT: buffer_store_dword v44, off, s[0:3], s33 offset:4 ; 4-byte Folded Spill ; GFX10-NEXT: buffer_store_dword v45, off, s[0:3], s33 ; 4-byte Folded Spill +; GFX10-NEXT: image_gather4_c_b_cl v[0:3], v[12:16], s[4:11], s[4:7] dmask:0x1 dim:SQ_RSRC_IMG_2D ; GFX10-NEXT: s_addk_i32 s32, 0x400 +; GFX10-NEXT: s_waitcnt_depctr 0xffe3 +; GFX10-NEXT: s_getpc_b64 s[4:5] +; GFX10-NEXT: s_add_u32 s4, s4, extern_func@gotpcrel32@lo+4 +; GFX10-NEXT: s_addc_u32 s5, s5, extern_func@gotpcrel32@hi+12 ; GFX10-NEXT: v_writelane_b32 v40, s30, 0 +; GFX10-NEXT: s_load_dwordx2 s[4:5], s[4:5], 0x0 ; GFX10-NEXT: v_mov_b32_e32 v41, v16 ; GFX10-NEXT: v_mov_b32_e32 v42, v15 ; GFX10-NEXT: v_mov_b32_e32 v43, v14 -; GFX10-NEXT: v_mov_b32_e32 v44, v13 ; GFX10-NEXT: v_writelane_b32 v40, s31, 1 +; GFX10-NEXT: v_mov_b32_e32 v44, v13 ; GFX10-NEXT: v_mov_b32_e32 v45, v12 -; GFX10-NEXT: v_writelane_b32 v40, s36, 2 -; GFX10-NEXT: s_mov_b32 s36, 0 -; GFX10-NEXT: v_writelane_b32 v40, s37, 3 -; GFX10-NEXT: s_mov_b32 s37, s36 -; GFX10-NEXT: v_writelane_b32 v40, s38, 4 -; GFX10-NEXT: s_mov_b32 s38, s36 -; GFX10-NEXT: v_writelane_b32 v40, s39, 5 -; GFX10-NEXT: s_mov_b32 s39, s36 -; GFX10-NEXT: v_writelane_b32 v40, s40, 6 -; GFX10-NEXT: s_mov_b32 s40, s36 -; GFX10-NEXT: v_writelane_b32 v40, s41, 7 -; GFX10-NEXT: s_mov_b32 s41, s36 -; GFX10-NEXT: v_writelane_b32 v40, s42, 8 -; GFX10-NEXT: s_mov_b32 s42, s36 -; GFX10-NEXT: v_writelane_b32 v40, s43, 9 -; GFX10-NEXT: s_mov_b32 s43, s36 -; GFX10-NEXT: image_gather4_c_b_cl v[0:3], v[12:16], s[36:43], s[4:7] dmask:0x1 dim:SQ_RSRC_IMG_2D -; GFX10-NEXT: s_waitcnt_depctr 0xffe3 -; GFX10-NEXT: s_getpc_b64 s[4:5] -; GFX10-NEXT: s_add_u32 s4, s4, extern_func@gotpcrel32@lo+4 -; GFX10-NEXT: s_addc_u32 s5, s5, extern_func@gotpcrel32@hi+12 -; GFX10-NEXT: s_load_dwordx2 s[4:5], s[4:5], 0x0 ; GFX10-NEXT: s_waitcnt vmcnt(0) ; GFX10-NEXT: global_store_dwordx4 v[0:1], v[0:3], off ; GFX10-NEXT: s_waitcnt lgkmcnt(0) ; GFX10-NEXT: s_swappc_b64 s[30:31], s[4:5] -; GFX10-NEXT: image_gather4_c_b_cl v[0:3], [v45, v44, v43, v42, v41], s[36:43], s[4:7] dmask:0x1 dim:SQ_RSRC_IMG_2D +; GFX10-NEXT: image_gather4_c_b_cl v[0:3], [v45, v44, v43, v42, v41], s[4:11], s[4:7] dmask:0x1 dim:SQ_RSRC_IMG_2D ; GFX10-NEXT: s_clause 0x4 ; GFX10-NEXT: buffer_load_dword v45, off, s[0:3], s33 ; GFX10-NEXT: buffer_load_dword v44, off, s[0:3], s33 offset:4 ; GFX10-NEXT: buffer_load_dword v43, off, s[0:3], s33 offset:8 ; GFX10-NEXT: buffer_load_dword v42, off, s[0:3], s33 offset:12 ; GFX10-NEXT: buffer_load_dword v41, off, s[0:3], s33 offset:16 -; GFX10-NEXT: v_readlane_b32 s43, v40, 9 -; GFX10-NEXT: v_readlane_b32 s42, v40, 8 -; GFX10-NEXT: v_readlane_b32 s41, v40, 7 -; GFX10-NEXT: v_readlane_b32 s40, v40, 6 -; GFX10-NEXT: v_readlane_b32 s39, v40, 5 -; GFX10-NEXT: v_readlane_b32 s38, v40, 4 -; GFX10-NEXT: v_readlane_b32 s37, v40, 3 -; GFX10-NEXT: v_readlane_b32 s36, v40, 2 ; GFX10-NEXT: v_readlane_b32 s31, v40, 1 ; GFX10-NEXT: v_readlane_b32 s30, v40, 0 ; GFX10-NEXT: s_addk_i32 s32, 0xfc00 -; GFX10-NEXT: v_readlane_b32 s33, v40, 10 +; GFX10-NEXT: v_readlane_b32 s33, v40, 2 ; GFX10-NEXT: s_or_saveexec_b32 s4, -1 ; GFX10-NEXT: buffer_load_dword v40, off, s[0:3], s32 offset:20 ; 4-byte Folded Reload ; GFX10-NEXT: s_waitcnt_depctr 0xffe3 @@ -372,7 +300,7 @@ define <4 x float> @call_preserved_vgpr_tuple8(<8 x i32> %rsrc, <4 x i32> %samp, ; GFX11-NEXT: s_or_saveexec_b32 s0, -1 ; GFX11-NEXT: scratch_store_b32 off, v40, s32 offset:20 ; 4-byte Folded Spill ; GFX11-NEXT: s_mov_b32 exec_lo, s0 -; GFX11-NEXT: v_writelane_b32 v40, s33, 10 +; GFX11-NEXT: v_writelane_b32 v40, s33, 2 ; GFX11-NEXT: s_mov_b32 s33, s32 ; GFX11-NEXT: s_clause 0x4 ; GFX11-NEXT: scratch_store_b32 off, v41, s33 offset:16 @@ -380,56 +308,32 @@ define <4 x float> @call_preserved_vgpr_tuple8(<8 x i32> %rsrc, <4 x i32> %samp, ; GFX11-NEXT: scratch_store_b32 off, v43, s33 offset:8 ; GFX11-NEXT: scratch_store_b32 off, v44, s33 offset:4 ; GFX11-NEXT: scratch_store_b32 off, v45, s33 +; GFX11-NEXT: image_gather4_c_b_cl v[0:3], v[12:16], s[0:7], s[0:3] dmask:0x1 dim:SQ_RSRC_IMG_2D ; GFX11-NEXT: s_add_i32 s32, s32, 32 +; GFX11-NEXT: s_getpc_b64 s[0:1] +; GFX11-NEXT: s_add_u32 s0, s0, extern_func@gotpcrel32@lo+4 +; GFX11-NEXT: s_addc_u32 s1, s1, extern_func@gotpcrel32@hi+12 ; GFX11-NEXT: v_writelane_b32 v40, s30, 0 +; GFX11-NEXT: s_load_b64 s[0:1], s[0:1], 0x0 ; GFX11-NEXT: v_dual_mov_b32 v41, v16 :: v_dual_mov_b32 v42, v15 ; GFX11-NEXT: v_dual_mov_b32 v43, v14 :: v_dual_mov_b32 v44, v13 ; GFX11-NEXT: v_writelane_b32 v40, s31, 1 ; GFX11-NEXT: v_mov_b32_e32 v45, v12 -; GFX11-NEXT: v_writelane_b32 v40, s36, 2 -; GFX11-NEXT: s_mov_b32 s36, 0 -; GFX11-NEXT: v_writelane_b32 v40, s37, 3 -; GFX11-NEXT: s_mov_b32 s37, s36 -; GFX11-NEXT: v_writelane_b32 v40, s38, 4 -; GFX11-NEXT: s_mov_b32 s38, s36 -; GFX11-NEXT: v_writelane_b32 v40, s39, 5 -; GFX11-NEXT: s_mov_b32 s39, s36 -; GFX11-NEXT: v_writelane_b32 v40, s40, 6 -; GFX11-NEXT: s_mov_b32 s40, s36 -; GFX11-NEXT: v_writelane_b32 v40, s41, 7 -; GFX11-NEXT: s_mov_b32 s41, s36 -; GFX11-NEXT: v_writelane_b32 v40, s42, 8 -; GFX11-NEXT: s_mov_b32 s42, s36 -; GFX11-NEXT: v_writelane_b32 v40, s43, 9 -; GFX11-NEXT: s_mov_b32 s43, s36 -; GFX11-NEXT: image_gather4_c_b_cl v[0:3], v[12:16], s[36:43], s[0:3] dmask:0x1 dim:SQ_RSRC_IMG_2D -; GFX11-NEXT: s_getpc_b64 s[0:1] -; GFX11-NEXT: s_add_u32 s0, s0, extern_func@gotpcrel32@lo+4 -; GFX11-NEXT: s_addc_u32 s1, s1, extern_func@gotpcrel32@hi+12 -; GFX11-NEXT: s_load_b64 s[0:1], s[0:1], 0x0 ; GFX11-NEXT: s_waitcnt vmcnt(0) ; GFX11-NEXT: global_store_b128 v[0:1], v[0:3], off ; GFX11-NEXT: s_waitcnt lgkmcnt(0) ; GFX11-NEXT: s_swappc_b64 s[30:31], s[0:1] -; GFX11-NEXT: image_gather4_c_b_cl v[0:3], [v45, v44, v43, v42, v41], s[36:43], s[0:3] dmask:0x1 dim:SQ_RSRC_IMG_2D +; GFX11-NEXT: image_gather4_c_b_cl v[0:3], [v45, v44, v43, v42, v41], s[0:7], s[0:3] dmask:0x1 dim:SQ_RSRC_IMG_2D ; GFX11-NEXT: s_clause 0x4 ; GFX11-NEXT: scratch_load_b32 v45, off, s33 ; GFX11-NEXT: scratch_load_b32 v44, off, s33 offset:4 ; GFX11-NEXT: scratch_load_b32 v43, off, s33 offset:8 ; GFX11-NEXT: scratch_load_b32 v42, off, s33 offset:12 ; GFX11-NEXT: scratch_load_b32 v41, off, s33 offset:16 -; GFX11-NEXT: v_readlane_b32 s43, v40, 9 -; GFX11-NEXT: v_readlane_b32 s42, v40, 8 -; GFX11-NEXT: v_readlane_b32 s41, v40, 7 -; GFX11-NEXT: v_readlane_b32 s40, v40, 6 -; GFX11-NEXT: v_readlane_b32 s39, v40, 5 -; GFX11-NEXT: v_readlane_b32 s38, v40, 4 -; GFX11-NEXT: v_readlane_b32 s37, v40, 3 -; GFX11-NEXT: v_readlane_b32 s36, v40, 2 ; GFX11-NEXT: v_readlane_b32 s31, v40, 1 ; GFX11-NEXT: v_readlane_b32 s30, v40, 0 ; GFX11-NEXT: s_addk_i32 s32, 0xffe0 -; GFX11-NEXT: v_readlane_b32 s33, v40, 10 +; GFX11-NEXT: v_readlane_b32 s33, v40, 2 ; GFX11-NEXT: s_or_saveexec_b32 s0, -1 ; GFX11-NEXT: scratch_load_b32 v40, off, s32 offset:20 ; 4-byte Folded Reload ; GFX11-NEXT: s_mov_b32 exec_lo, s0 |