[Openmp-commits] [PATCH] D110845: [libomptarget] Apply D110029 to amdgpu

[Jon Chesterfield via Phabricator via Openmp-commits]

JonChesterfield added inline comments.


================
Comment at: openmp/libomptarget/plugins/amdgpu/src/rtl.cpp:1696
       uint16_t TSize;
       uint16_t WG_Size;
     };
----------------
the uint8_t mode here is essentially legacy cruft where rocm used to encode the same information twice, working on getting rid of the rest of that in parallel to this


================
Comment at: openmp/libomptarget/plugins/amdgpu/src/rtl.cpp:1775
     }
 
+    // Read execution mode from global in binary
----------------
Same block of code as above with int8_t replaced with the enum, shame diff doesn't render it as such. Moving it below that block means executing it unconditionally.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D110845/new/

https://reviews.llvm.org/D110845