diff options
| author | Matt Arsenault <Matthew.Arsenault@amd.com> | 2019-09-05 23:40:14 +0000 |
|---|---|---|
| committer | Matt Arsenault <Matthew.Arsenault@amd.com> | 2019-09-05 23:40:14 +0000 |
| commit | 59ff77ee38c6902ffa4b4bc71e80ab01032babec (patch) | |
| tree | 0fb3b583d9e5cfa82d24ecde9bda472db42e787c /llvm/test/CodeGen/AMDGPU | |
| parent | 9dd453ce8d6b0d0970ed11a93d5180f74f620078 (diff) | |
| download | bcm5719-llvm-59ff77ee38c6902ffa4b4bc71e80ab01032babec.tar.gz bcm5719-llvm-59ff77ee38c6902ffa4b4bc71e80ab01032babec.zip | |
AMDGPU: Fix emitting multiple stack loads for stack passed workitems
The same stack is loaded for each workitem ID, and each use. Nothing
prevents you from creating multiple fixed stack objects with the same
offsets, so this was creating a load for each unique frame index,
despite them being the same offset. Re-use the same frame index so the
loads are CSEable.
llvm-svn: 371148
Diffstat (limited to 'llvm/test/CodeGen/AMDGPU')
| -rw-r--r-- | llvm/test/CodeGen/AMDGPU/callee-special-input-vgprs.ll | 28 |
1 files changed, 17 insertions, 11 deletions
diff --git a/llvm/test/CodeGen/AMDGPU/callee-special-input-vgprs.ll b/llvm/test/CodeGen/AMDGPU/callee-special-input-vgprs.ll index 2df3ad9cdef..d5211ceede3 100644 --- a/llvm/test/CodeGen/AMDGPU/callee-special-input-vgprs.ll +++ b/llvm/test/CodeGen/AMDGPU/callee-special-input-vgprs.ll @@ -543,19 +543,25 @@ define void @func_call_too_many_args_use_workitem_id_x_byval() #1 { ret void } +; Only one stack load should be emitted for all 3 values. ; GCN-LABEL: {{^}}too_many_args_use_workitem_id_xyz: -; GCN: buffer_store_dword v32, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill -; GCN: buffer_load_dword v32, off, s[0:3], s32{{$}} -; GCN: v_and_b32_e32 v32, 0x3ff, v32 -; GCN: {{flat|global}}_store_dword v{{\[[0-9]+:[0-9]+]}}, v32 -; GCN: buffer_load_dword v32, off, s[0:3], s32{{$}} -; GCN: v_bfe_u32 v32, v32, 10, 10 -; GCN: {{flat|global}}_store_dword v{{\[[0-9]+:[0-9]+]}}, v32 +; GCN: buffer_store_dword v32, off, s[0:3], s32 offset:8 ; 4-byte Folded Spill +; GCN: buffer_store_dword v33, off, s[0:3], s32 offset:4 ; 4-byte Folded Spill +; GCN-NOT: buffer_load_dword v{{[0-9]+}}, off, s[0:3], s32{{$}} ; GCN: buffer_load_dword v32, off, s[0:3], s32{{$}} -; GCN: v_bfe_u32 v32, v32, 20, 10 -; GCN: {{flat|global}}_store_dword v{{\[[0-9]+:[0-9]+]}}, v32 - -; GCN: buffer_load_dword v32, off, s[0:3], s32 offset:4 ; 4-byte Folded Reload +; GCN-NOT: buffer_load_dword + +; GCN: v_and_b32_e32 [[AND_X:v[0-9]+]], 0x3ff, v32 +; GCN-NOT: buffer_load_dword +; GCN: {{flat|global}}_store_dword v{{\[[0-9]+:[0-9]+]}}, [[AND_X]] +; GCN-NOT: buffer_load_dword +; GCN: v_bfe_u32 [[BFE_Y:v[0-9]+]], v32, 10, 10 +; GCN-NEXT: v_bfe_u32 [[BFE_Z:v[0-9]+]], v32, 20, 10 +; GCN-NEXT: {{flat|global}}_store_dword v{{\[[0-9]+:[0-9]+]}}, [[BFE_Y]] +; GCN-NEXT: {{flat|global}}_store_dword v{{\[[0-9]+:[0-9]+]}}, [[BFE_Z]] + +; GCN: buffer_load_dword v33, off, s[0:3], s32 offset:4 ; 4-byte Folded Reload +; GCN: buffer_load_dword v32, off, s[0:3], s32 offset:8 ; 4-byte Folded Reload ; GCN-NEXT: s_waitcnt ; GCN-NEXT: s_setpc_b64 define void @too_many_args_use_workitem_id_xyz( |
