Skip to content

Latest commit

 

History

History
181 lines (129 loc) · 7.01 KB

common_hpc_dependencies.rst

File metadata and controls

181 lines (129 loc) · 7.01 KB
Raw

Common HPC Dependencies

BLT creates named targets for the common HPC dependencies that most HPC projects need, such as MPI, CUDA, HIP, and OpenMP. Something BLT assists it's users with is getting these dependencies to interoperate within the same library or executable.

As previously mentioned in :ref:`AddingTests`, BLT also provides bundled versions of GoogleTest, GoogleMock, GoogleBenchmark, and FRUIT. Not only are the source for these included, we provide named CMake targets for them as well.

BLT's mpi, cuda, cuda_runtime, blt::hip, blt::hip_runtime,and openmp targets are all defined via the :ref:`blt_import_library` macro. This creates a true CMake imported target that is inherited properly through the CMake's dependency graph.

Note

BLT also supports exporting its third-party targets via the BLT_EXPORT_THIRDPARTY option. See :ref:`ExportingTargets` for more information.

You have already seen one use of DEPENDS_ON for a BLT dependency, gtest, in test_1:

.. literalinclude:: calc_pi/CMakeLists.txt
   :start-after: _blt_tutorial_calcpi_test1_executable_start
   :end-before:  _blt_tutorial_calcpi_test1_executable_end
   :language: cmake

MPI

Our next example, test_2, builds and tests the calc_pi_mpi library, which uses MPI to parallelize the calculation over the integration intervals.

To enable MPI, we set ENABLE_MPI, MPI_C_COMPILER, and MPI_CXX_COMPILER in our host config file. Here is a snippet with these settings for LLNL's Lassen Cluster:

.. literalinclude:: ../../host-configs/llnl/toss_3_x86_64_ib/gcc@8.3.1.cmake
   :start-after: _blt_tutorial_mpi_config_start
   :end-before:  _blt_tutorial_mpi_config_end
   :language: cmake

Here, you can see how calc_pi_mpi and test_2 use DEPENDS_ON:

.. literalinclude:: calc_pi/CMakeLists.txt
   :start-after: _blt_tutorial_calcpi_test2_executable_start
   :end-before:  _blt_tutorial_calcpi_test2_executable_end
   :language: cmake

For MPI unit tests, you also need to specify the number of MPI Tasks to launch. We use the NUM_MPI_TASKS argument to :ref:`blt_add_test` macro.

.. literalinclude:: calc_pi/CMakeLists.txt
   :start-after: _blt_tutorial_calcpi_test2_test_start
   :end-before:  _blt_tutorial_calcpi_test2_test_end
   :language: cmake

As mentioned in :ref:`AddingTests`, GoogleTest provides a default main() driver that will execute all unit tests defined in the source. To test MPI code, we need to create a main that initializes and finalizes MPI in addition to Google Test. test_2.cpp provides an example driver for MPI with GoogleTest.

.. literalinclude:: calc_pi/test_2.cpp
   :start-after: _blt_tutorial_calcpi_test2_main_start
   :end-before:  _blt_tutorial_calcpi_test2_main_end
   :language: cpp

Note

While we have tried to ensure that BLT chooses the correct setup information for MPI, there are several niche cases where the default behavior is insufficient. We have provided several available override variables:

  • BLT_MPI_COMPILE_FLAGS
  • BLT_MPI_INCLUDES
  • BLT_MPI_LIBRARIES
  • BLT_MPI_LINK_FLAGS

BLT also has the variable ENABLE_FIND_MPI which turns off all CMake's FindMPI logic and then uses the MPI wrapper directly when you provide them as the default compilers.

CUDA

Finally, test_3 builds and tests the calc_pi_cuda library, which uses CUDA to parallelize the calculation over the integration intervals.

To enable CUDA, we set ENABLE_CUDA, CMAKE_CUDA_COMPILER, and CUDA_TOOLKIT_ROOT_DIR in our host config file. Also before enabling the CUDA language in CMake, you need to set CMAKE_CUDA_HOST_COMPILER in CMake 3.9+ or CUDA_HOST_COMPILER in previous versions. If you do not call enable_language(CUDA), BLT will set the appropriate host compiler variable for you and enable the CUDA language.

Here is a snippet with these settings for LLNL's Lassen Cluster:

.. literalinclude:: ../../host-configs/llnl/blueos_3_ppc64le_ib_p9/clang@upstream_nvcc_c++17.cmake
   :start-after: _blt_tutorial_cuda_config_start
   :end-before:  _blt_tutorial_cuda_config_end
   :language: cmake

Here, you can see how calc_pi_cuda and test_3 use DEPENDS_ON:

.. literalinclude:: calc_pi/CMakeLists.txt
   :start-after: _blt_tutorial_calcpi_cuda_start
   :end-before:  _blt_tutorial_calcpi_cuda_end
   :language: cmake

The cuda dependency for calc_pi_cuda is a little special, along with adding the normal CUDA library and headers to your library or executable, it also tells BLT that this target's C/C++/CUDA source files need to be compiled via nvcc or cuda-clang. If this is not a requirement, you can use the dependency cuda_runtime which also adds the CUDA runtime library and headers but will not compile each source file with nvcc.

Some other useful CUDA flags are:

.. literalinclude:: ../../host-configs/llnl/blueos_3_ppc64le_ib_p9/clang@upstream_nvcc_xlf.cmake
   :start-after: _blt_tutorial_useful_cuda_flags_start
   :end-before:  _blt_tutorial_useful_cuda_flags_end
   :language: cmake

OpenMP

To enable OpenMP, set ENABLE_OPENMP in your host-config file or before loading SetupBLT.cmake. Once OpenMP is enabled, simply add openmp to your library executable's DEPENDS_ON list.

Here is an example of how to add an OpenMP enabled executable:

.. literalinclude:: ../../tests/smoke/CMakeLists.txt
   :start-after: _blt_tutorial_openmp_executable_start
   :end-before:  _blt_tutorial_openmp_executable_end
   :language: cmake

Note

While we have tried to ensure that BLT chooses the correct compile and link flags for OpenMP, there are several niche cases where the default options are insufficient. For example, linking with NVCC requires to link in the OpenMP libraries directly instead of relying on the compile and link flags returned by CMake's FindOpenMP package. An example of this is in host-configs/llnl/blueos_3_ppc64le_ib_p9/clang@upstream_link_with_nvcc.cmake. We provide two variables to override BLT's OpenMP flag logic:

  • BLT_OPENMP_COMPILE_FLAGS
  • BLT_OPENMP_LINK_FLAGS

Here is an example of how to add an OpenMP enabled test that sets the amount of threads used:

.. literalinclude:: ../../tests/smoke/CMakeLists.txt
   :start-after: _blt_tutorial_openmp_test_start
   :end-before:  _blt_tutorial_openmp_test_end
   :language: cmake

HIP

HIP tutorial coming soon!

BLT also supports AMD's HIP via a mechanism very similar to our CUDA support.

Important Setup Variables

  • ENABLE_HIP : Enables HIP support in BLT
  • HIP_ROOT_DIR : Root directory for HIP installation

BLT Targets

  • blt::hip : Adds include directories, hip runtime libraries, and compiles source with hipcc
  • blt::hip_runtime : Adds include directories and hip runtime libraries