[Openmp-commits] [PATCH] D114270: [openmp][amdgpu][nfc] Simplify implicit args handling

Jon Chesterfield via Phabricator via Openmp-commits openmp-commits at lists.llvm.org
Mon Jan 10 11:08:36 PST 2022


JonChesterfield added inline comments.


================
Comment at: openmp/libomptarget/plugins/amdgpu/impl/internal.h:45
 } impl_implicit_args_t;
+static_assert(sizeof(impl_implicit_args_t) == 56, "");
 
----------------
JonChesterfield wrote:
> arsenm wrote:
> > brooks wrote:
> > > This breaks builds on 32-bit platforms (e.g., i386). If they aren't supported that's fine, but then cmake adjustments are required to prevent it from trying to build.
> > Just use explicit uint64_t, the named types randomly changing sizes is one of the dumbest things about C
> I don't know if i386 will work in general, but those ^ have to be 64 bit regardless, will fix shortly
Patch at D116963


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D114270/new/

https://reviews.llvm.org/D114270



More information about the Openmp-commits mailing list