diff options
| author | Guozhi Wei <carrot@google.com> | 2019-08-22 16:21:32 +0000 |
|---|---|---|
| committer | Guozhi Wei <carrot@google.com> | 2019-08-22 16:21:32 +0000 |
| commit | 51f48295cbe8fa3a44db263b528dd9f7bae7bf9a (patch) | |
| tree | 28614c4e7fca7431bf287b07453443bacab0f39b /llvm/test/CodeGen/AMDGPU/indirect-addressing-si.ll | |
| parent | 95cf66de7cc186253c59a009bb795da1cbe9d6f4 (diff) | |
| download | bcm5719-llvm-51f48295cbe8fa3a44db263b528dd9f7bae7bf9a.tar.gz bcm5719-llvm-51f48295cbe8fa3a44db263b528dd9f7bae7bf9a.zip | |
[MBP] Disable aggressive loop rotate in plain mode
Patch https://reviews.llvm.org/D43256 introduced more aggressive loop layout optimization which depends on profile information. If profile information is not available, the statically estimated profile information(generated by BranchProbabilityInfo.cpp) is used. If user program doesn't behave as BranchProbabilityInfo.cpp expected, the layout may be worse.
To be conservative this patch restores the original layout algorithm in plain mode. But user can still try the aggressive layout optimization with -force-precise-rotation-cost=true.
Differential Revision: https://reviews.llvm.org/D65673
llvm-svn: 369664
Diffstat (limited to 'llvm/test/CodeGen/AMDGPU/indirect-addressing-si.ll')
| -rw-r--r-- | llvm/test/CodeGen/AMDGPU/indirect-addressing-si.ll | 12 |
1 files changed, 6 insertions, 6 deletions
diff --git a/llvm/test/CodeGen/AMDGPU/indirect-addressing-si.ll b/llvm/test/CodeGen/AMDGPU/indirect-addressing-si.ll index bde1cd5c435..dc7f495c42f 100644 --- a/llvm/test/CodeGen/AMDGPU/indirect-addressing-si.ll +++ b/llvm/test/CodeGen/AMDGPU/indirect-addressing-si.ll @@ -630,7 +630,12 @@ define amdgpu_kernel void @insertelement_v16f32_or_index(<16 x float> addrspace( ; GCN-LABEL: {{^}}broken_phi_bb: ; GCN: v_mov_b32_e32 [[PHIREG:v[0-9]+]], 8 -; GCN: [[BB2:BB[0-9]+_[0-9]+]]: +; GCN: s_branch [[BB2:BB[0-9]+_[0-9]+]] + +; GCN: {{^BB[0-9]+_[0-9]+}}: +; GCN: s_mov_b64 exec, + +; GCN: [[BB2]]: ; GCN: v_cmp_le_i32_e32 vcc, s{{[0-9]+}}, [[PHIREG]] ; GCN: buffer_load_dword @@ -642,11 +647,6 @@ define amdgpu_kernel void @insertelement_v16f32_or_index(<16 x float> addrspace( ; IDXMODE: s_set_gpr_idx_off ; GCN: s_cbranch_execnz [[REGLOOP]] - -; GCN: {{^; %bb.[0-9]}}: -; GCN: s_mov_b64 exec, -; GCN: s_branch [[BB2]] - define amdgpu_kernel void @broken_phi_bb(i32 %arg, i32 %arg1) #0 { bb: br label %bb2 |

