[Openmp-commits] [llvm] [openmp] [offload] - Remove standalone build in favor of 'runtimes' (PR #170693)

[via Openmp-commits]

================
@@ -34,6 +34,22 @@ OpenMP in Flang Technical Call
 FAQ
 ---
 
+For a standalone build, users are expected to use the LLVM runtimes directory.
+
+.. code-block:: sh
+
+  $> cd llvm-project  # The llvm-project checkout
+  $> mkdir build
+  $> cd build
+  $> cmake ../runtimes -G Ninja                                                 \
+     -DLLVM_BINARY_DIR="${PATH_TO_LLVM}" \ # Point to build directory if lit testing
+     -DLLVM_ENABLE_RUNTIMES="openmp;offload" \ # offload has a dependency on openmp
+     -DCMAKE_BUILD_TYPE=<Debug|Release>   \ # Select build type
+     -DCMAKE_INSTALL_PREFIX=<PATH>        \ # Where the libraries will live
+  $> ninja install
+
+A separate cmake invocation to build the openmp deviceRTL via ``-DLLVM_ENABLE_RUNTIMES=openmp`` and ``-DLLVM_DEFAULT_TARGET_TRIPLE=<amdgcn-amd-amdhsa|nvptx64-nvidia-cuda>`` is needed.
----------------
estewart08 wrote:

Thanks for the pointer I will remove this change.

https://github.com/llvm/llvm-project/pull/170693