Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

MueLu: Hip compilation errors with LWGraph_kokkos test, rocm/5.6.0 #12691

Closed
ndellingwood opened this issue Jan 24, 2024 · 3 comments
Closed
Assignees
Labels
pkg: MueLu type: bug The primary issue is a bug in Trilinos code or tests

Comments

@ndellingwood
Copy link
Contributor

Bug Report

@trilinos/muelu

The LWGraph_kokkos.cpp fails to compile in Hip builds with rocm/5.6.0 with output:

22:39:48 /home/jenkins/caraway-new/workspace/Trilinos_Caraway_Hip_Serial_Rocm5_6_0_MI210/Trilinos/packages/muelu/test/unit_tests_kokkos/LWGraph_kokkos.cpp:158:25: error: reference to __host__ function 'operator->' in __host__ __device__ function
22:39:48           if (lclLWGraph->GetNodeNumVertices() != numrows)

This follows merge of these commits:

Changes:
Xpetra: Add getLocalGraph* to EpetraCrsGraph (detail)
MueLu Utilities: Refactor DetectDirichletRows and ApplyRowSumCriterion (detail)
MueLu: Refactor LWGraphs (detail)
MueLu Viz test: Use test ETI (detail)
MueLu RefMaxwell: Fix spelling (detail)
MueLu: Add H2D regression test (detail)
Teko: clang-format (detail)
MueLu: Add H2D count specifically for aggregation (detail)

Likely related to PR #12629 ?

Steps to Reproduce

  1. SHA1: [insert here]
  2. Configure script: Caraway testbed, MI210 queue
module load python rocm/5.6.0 cmake/3.25.2 mpi/openmpi-x86_64 openblas/0.3.20/rocm ninja/1.7.2
module list
export OMPI_CXX=$ROCM_PATH/bin/hipcc
export TPETRA_ASSUME_GPU_AWARE_MPI=0

cmake \
-G"Ninja" \
-DCMAKE_INSTALL_PREFIX=$PWD/install \
-DCMAKE_CXX_STANDARD="17" \
-DCMAKE_CXX_COMPILER="`which mpicxx`" \
-DCMAKE_C_COMPILER="`which mpicc`" \
-DCMAKE_FORTRAN_COMPILER="`which mpifort`" \
-DCMAKE_BUILD_TYPE="RELEASE" \
-DBUILD_SHARED_LIBS="OFF" \
\
-DTrilinos_ENABLE_ALL_PACKAGES=OFF \
-DTrilinos_ENABLE_ALL_OPTIONAL_PACKAGES=OFF \
-DTrilinos_ENABLE_EXPLICIT_INSTANTIATION=ON \
-DTrilinos_ASSERT_MISSING_PACKAGES=OFF \
-DTrilinos_ALLOW_NO_PACKAGES=OFF \
-DTrilinos_ENABLE_OpenMP=OFF \
-DTrilinos_ENABLE_TESTS=ON \
\
-DTrilinos_ENABLE_Amesos2=ON \
 -DAmesos2_ENABLE_SuperLU=OFF \
 -DAmesos2_ENABLE_KLU2=ON \
-DTrilinos_ENABLE_Belos=ON \
-DTrilinos_ENABLE_Ifpack2=ON \
-DTrilinos_ENABLE_Kokkos=ON \
 -DKokkos_ARCH_VEGA90A=ON \
 -DKokkos_ENABLE_CUDA=OFF \
 -DKokkos_ENABLE_HIP=ON \
 -DKokkos_ENABLE_OPENMP=OFF \
-DTrilinos_ENABLE_KokkosKernels=ON \
-DTrilinos_ENABLE_MueLu=ON \
 -DMueLu_ENABLE_Kokkos_Refactor=ON \
-DTrilinos_ENABLE_Tpetra=ON \
 -DTpetra_ENABLE_CUDA=OFF \
 -DTpetra_INST_HIP=ON \
 -DTpetra_INST_SERIAL=OFF \
 -DTpetra_INST_OPENMP=OFF \
 -DTpetra_INST_DOUBLE=ON \
-DTrilinos_ENABLE_Gtest=ON \
-DTrilinos_ENABLE_Teuchos=ON \
-DTrilinos_ENABLE_Xpetra=ON \
 -DXpetra_ENABLE_Kokkos_Refactor=ON \
-DTrilinos_ENABLE_Zoltan2=ON \
-DTrilinos_ENABLE_Panzer=ON \
-DTPL_ENABLE_BLAS=ON \
  -D BLAS_LIBRARY_DIRS:FILEPATH="${OPENBLAS_ROOT}/lib" \
  -D BLAS_LIBRARY_NAMES:STRING="openblas" \
-DTPL_ENABLE_LAPACK=ON \
  -D LAPACK_INCLUDE_DIRS:FILEPATH="${OPENBLAS_ROOT}/include" \
  -D LAPACK_LIBRARY_DIRS:FILEPATH="${OPENBLAS_ROOT}/lib" \
  -D LAPACK_LIBRARY_NAMES:STRING="openblas" \
-DTPL_ENABLE_Netcdf=OFF \
-DTPL_ENABLE_MPI=ON \
-DMPI_USE_COMPILER_WRAPPERS=ON \
-DMPI_EXEC="mpirun" \
-DMPI_EXEC_NUMPROCS_FLAG="-np" \
-DMPI_EXEC_POST_NUMPROCS_FLAGS:STRING="-bind-to;none" \
\
$TRILINOS_DIR
@ndellingwood ndellingwood added type: bug The primary issue is a bug in Trilinos code or tests pkg: MueLu labels Jan 24, 2024
Copy link

Automatic mention of the @trilinos/muelu team

1 similar comment
Copy link

Automatic mention of the @trilinos/muelu team

@ndellingwood
Copy link
Contributor Author

Thanks @cgcgcg !

@jhux2 jhux2 added this to MueLu Aug 12, 2024
@jhux2 jhux2 moved this to Done in MueLu Aug 12, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
pkg: MueLu type: bug The primary issue is a bug in Trilinos code or tests
Projects
Status: Done
Development

No branches or pull requests

2 participants