[Openmp-commits] [PATCH] D44992: [OpenMP] enable bc file compilation using the latest clang

[Greg Rodgers via Phabricator via Openmp-commits]

gregrodgers added a comment.

Maybe my search is missing something, but the only place I see CUDARelocatableDeviceCode is in lib/Sema/SemaDeclAttr.cpp to allow for extern __shared__.   How could this be causing slowness?  I would think forcing extern to be __global__ would be slower.

In my search for cuda_rdc, I only see code to forward the option to clang cc1 if set in the driver.  So again, I don't see it affecting anything but SemaDeclAttr.cpp.

Greg


Repository:
  rOMP OpenMP

https://reviews.llvm.org/D44992