[PATCH] D101265: [OpenMP][CMake] Use in-project clang as CUDA->IR compiler.

[Joachim Protze via Phabricator via llvm-commits]

protze.joachim added a comment.

In D101265#2724364 <https://reviews.llvm.org/D101265#2724364>, @tianshilei1992 wrote:

> In D101265#2718317 <https://reviews.llvm.org/D101265#2718317>, @Meinersbur wrote:
>
>> In D101265#2715759 <https://reviews.llvm.org/D101265#2715759>, @tianshilei1992 wrote:
>>
>>> In D101265#2715749 <https://reviews.llvm.org/D101265#2715749>, @Meinersbur wrote:
>>>
>>>> In D101265#2715714 <https://reviews.llvm.org/D101265#2715714>, @tianshilei1992 wrote:
>>>>
>>>>> If OpenMP is built via `LLVM_ENABLE_PROJECTS`, it is *intentional* to build OpenMP w/o using the clang in recent build.
>>>>
>>>> Is this documented somewhere?
>>>
>>> You could refer to https://llvm.org/docs/CMake.html for `LLVM_ENABLE_PROJECTS` and https://llvm.org/docs/BuildingADistribution.html for `LLVM_ENABLE_RUNTIMES`.
>>
>> The first link does not mention what is intended to build with what compiler.
>> The second link is for creating a distributable package, which does not apply here.
>
> The two links show the difference between `LLVM_ENABLE_RUNTIMES` and `LLVM_ENABLE_PROJECTS`. It does mention that projects in `LLVM_ENABLE_RUNTIMES` are built by the recent built clang.

I think, what Michael really meant with "Is this documented somewhere?": 
Is there user documentation, that LLVM_ENABLE_PROJECTS=openmp was bricked by llvm-12 and you shouldn't expect libomptarget to work properly unless moving to LLVM_ENABLE_RUNTIMES=openmp or doing two-stage build? Since this is crucial information, this should be BOLD in the OpenMP release notes.

Building LLVM_ENABLE_PROJECTS=openmp with the previous clang release used to work fine with single-stage build.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D101265/new/

https://reviews.llvm.org/D101265