[PATCH] D99488: [SYCL][Doc] Add address space handling section to SYCL documentation

[Anastasia Stulova via Phabricator via cfe-commits]

Anastasia accepted this revision.
Anastasia added a comment.
This revision is now accepted and ready to land.

LGTM! Thanks for working on this. The expected sematic seems fairly clear now. We might add a few more details while refining the implementation but it should not block the development progress at this point.



================
Comment at: clang/docs/SYCLSupport.rst:255
+to a specific address space. This is described in
+https://www.khronos.org/registry/SYCL/specs/sycl-2020/html/sycl-2020.html#subsec:commonAddressSpace.
+
----------------
I think this should be converted into a hyper link?


================
Comment at: clang/docs/SYCLSupport.rst:258
+This is also in line with the behaviour of CUDA (small example
+https://godbolt.org/z/veqTfo9PK).
+
----------------
same here, the list should be converted into a hyper link.


================
Comment at: clang/docs/SYCLSupport.rst:344
+
+The SPIR target allocates SYCL namespace scope variables in the global address
+space.
----------------
bader wrote:
> Anastasia wrote:
> > Interesting, will this deduction always be target specific or can it be generalized since it is governed by the language semantic already?
> > Interesting, will this deduction always be target specific or can it be generalized since it is governed by the language semantic already?
> 
> It's target specific deduction. CPU targets doesn't require such deduction.
That's right we have a similar situation in OpenCL. If you expect certain logic to be common enough it might make sense to add abstractions at some point.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D99488/new/

https://reviews.llvm.org/D99488