[PATCH] D124385: AMDGPU: Special case divergence analysis for wave ID computation

Stanislav Mekhanoshin via Phabricator via llvm-commits llvm-commits at lists.llvm.org
Fri Nov 18 13:19:42 PST 2022


rampitec added a comment.

Description speaks about 1D, while code targets multidimensional grids. In fact for 1D you would not need a wg to be uniform, it is only needed for 2D and 3D.



================
Comment at: llvm/lib/Target/AMDGPU/AMDGPUTargetTransformInfo.cpp:943
+  return match(PtrBase, m_Intrinsic<Intrinsic::amdgcn_dispatch_ptr>()) &&
+         Offset == 4 + 2 * GroupIdx;
+}
----------------
Can this constant be defined somewhere?


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D124385/new/

https://reviews.llvm.org/D124385



More information about the llvm-commits mailing list