llnl / comb Goto Github PK
View Code? Open in Web Editor NEWComb is a communication performance benchmarking tool.
License: MIT License
Comb is a communication performance benchmarking tool.
License: MIT License
Hi,
I want to compile the CUDA version, and I get some trouble during make
.
I have change the script in ./scripts/ubuntu-builds/ubuntu_nvcc10_gcc8.sh
and ./host-configs/ubuntu-builds/nvcc_gcc_X.cmake
respectively according to my system environment. What needs to be mentioned is that I add -std=c++11
to COMB_NVCC_FLAGS
in nvcc_gcc_X.cmake
. Otherwise, it will has the errors as bellow.
/app/compiler/gcc/4.9.4/include/c++/4.9.4/bits/c++0x_warning.h:32:2: error:#error This file requires compiler and library support for the ISO C++ 2011 standard. This support is currently experimental, and must be enabled with the -std=c++11 or -std=gnu++11 compiler options.
#error This file requires compiler and library support for the \
^
make[2]: *** [CMakeFiles/comb.dir/src/comb.cpp.o] error 1
make[1]: *** [CMakeFiles/comb.dir/all] error 2
After that, I run the script and get below.
loading initial cache file ../host-configs/ubuntu-builds/nvcc_gcc_X.cmake
-- The CXX compiler identification is GNU 4.9.4
-- Check for working CXX compiler: /app/compiler/gcc/4.9.4/bin/g++
-- Check for working CXX compiler: /app/compiler/gcc/4.9.4/bin/g++ -- works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- BLT Version: 0.3.0
-- CMake Version: 3.10.2
-- CMake Executable: /GPUFS/app_GPU/application/cmake/3.10.2/bin/cmake
-- Found Git: /usr/bin/git (found version "1.8.3.1")
-- Git Support is ON
-- Git Executable: /usr/bin/git
-- Git Version: 1.8.3.1
-- MPI Support is ON
-- Enable FindMPI: ON
-- Found MPI_CXX: /app/compiler/intel/15.0.1/impi/5.0.2.044/intel64/lib/libmpicxx.so (found version "3.0")
-- Found MPI: TRUE (found version "3.0")
-- BLT MPI Compile Flags:
-- BLT MPI Include Paths: /app/compiler/intel/15.0.1/impi/5.0.2.044/intel64/include
-- BLT MPI Libraries: /app/compiler/intel/15.0.1/impi/5.0.2.044/intel64/lib/libmpicxx.so;/app/compiler/intel/15.0.1/impi/5.0.2.044/intel64/lib/libmpifor
t.so;/app/compiler/intel/15.0.1/impi/5.0.2.044/intel64/lib/release_mt/libmpi.so;/app/compiler/intel/15.0.1/impi/5.0.2.044/intel64/lib/libmpigi.a;/lib64/libdl
.so;/lib64/librt.so;/lib64/libpthread.so
-- BLT MPI Link Flags: -Xlinker --enable-new-dtags -Xlinker -rpath -Xlinker /app/compiler/intel/15.0.1//impi/5.0.2.044/intel64/lib/release_mt -Xlinker -r
path -Xlinker /app/compiler/intel/15.0.1//impi/5.0.2.044/intel64/lib -Xlinker -rpath -Xlinker /opt/intel/mpi-rt/5.0/intel64/lib/release_mt -Xlinker -rpath -X
linker /opt/intel/mpi-rt/5.0/intel64/lib
-- MPI Executable: /app/compiler/intel/15.0.1/impi/5.0.2.044/intel64/bin/mpiexec
-- MPI Num Proc Flag: -n
-- MPI Command Append:
-- OpenMP Support is ON
-- Found OpenMP_CXX: -fopenmp (found version "4.0")
-- Found OpenMP: TRUE (found version "4.0")
-- OpenMP Compile Flags: $<$<NOT:$<COMPILE_LANGUAGE:CUDA>>:-fopenmp>;$<$<COMPILE_LANGUAGE:CUDA>:-Xcompiler=-fopenmp>
-- OpenMP Link Flags: -fopenmp
-- CUDA Support is ON
-- The CUDA compiler identification is NVIDIA 10.0.130
-- Check for working CUDA compiler: /GPUFS/app_GPU/compiler/CUDA/10.0.1/bin/nvcc
-- Check for working CUDA compiler: /GPUFS/app_GPU/compiler/CUDA/10.0.1/bin/nvcc -- works
-- Detecting CUDA compiler ABI info
-- Detecting CUDA compiler ABI info - done
-- Looking for C++ include pthread.h
-- Looking for C++ include pthread.h - found
-- Looking for pthread_create
-- Looking for pthread_create - not found
-- Looking for pthread_create in pthreads
-- Looking for pthread_create in pthreads - not found
-- Looking for pthread_create in pthread
-- Looking for pthread_create in pthread - found
-- Found Threads: TRUE
-- Found CUDA: /GPUFS/app_GPU/compiler/CUDA/10.0.1 (found version "10.0")
-- CUDA Version: 10.0
-- CUDA Compiler: /GPUFS/app_GPU/compiler/CUDA/10.0.1/bin/nvcc
-- CUDA Host Compiler: /app/compiler/gcc/4.9.4/bin/g++
-- CUDA Include Path: /GPUFS/app_GPU/compiler/CUDA/10.0.1/include
-- CUDA Libraries: /GPUFS/app_GPU/compiler/CUDA/10.0.1/lib64/libcudart_static.a;-lpthread;dl;/usr/lib64/librt.so
-- CUDA Compile Flags:
-- CUDA Link Flags:
-- CUDA Separable Compilation: ON
-- CUDA Link with NVCC:
-- HIP Support is OFF
-- HCC Support is OFF
-- Found Doxygen: /usr/bin/doxygen (found version "1.8.5") found components: doxygen dot
-- Sphinx support is ON
-- Failed to locate Sphinx executable (missing: SPHINX_EXECUTABLE)
-- Valgrind support is ON
-- Found Valgrind: /usr/bin/valgrind
-- Uncrustify support is ON
-- Failed to locate Uncrustify executable (missing: UNCRUSTIFY_EXECUTABLE)
-- AStyle support is ON
-- Failed to locate AStyle executable (missing: ASTYLE_EXECUTABLE)
-- Cppcheck support is ON
-- Failed to locate Cppcheck executable (missing: CPPCHECK_EXECUTABLE)
-- ClangQuery support is ON
-- Failed to locate ClangQuery executable (missing: CLANGQUERY_EXECUTABLE)
-- C Compiler family is GNU
-- Adding optional BLT definitions and compiler flags
-- Enabling all compiler warnings on all targets.
-- Fortran support disabled.
-- CMAKE_C_FLAGS flags are: -Wall -Wextra
-- CMAKE_CXX_FLAGS flags are: -Wall -Wextra
-- CMAKE_EXE_LINKER_FLAGS flags are:
-- Google Test Support is ON
-- Google Mock Support is OFF
-- The C compiler identification is GNU 4.9.4
-- Check for working C compiler: /app/compiler/gcc/4.9.4/bin/gcc
-- Check for working C compiler: /app/compiler/gcc/4.9.4/bin/gcc -- works
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Detecting C compile features
-- Detecting C compile features - done
-- Found PythonInterp: /usr/bin/python (found version "2.7.5")
-- MPI Enabled
-- OpenMP Enabled
-- Cuda Enabled
-- Configuring done
-- Generating done
-- Build files have been written to: /GPUFS/demo/benchmark/comb/build_ubuntu-nvcc10-gcc8
When I make
, I get the errors.
/GPUFS/demo/benchmark/comb/include/graph_launch.hpp(154): error: identifier "cudaGraphExecKernelNodeSetParams" is undefined
/GPUFS/demo/benchmark/comb/include/graph_launch.hpp(174): error: argument of type "const cudaGraphNode_t *" is incompatible with parameter of type "cudaGraphNode_t *"
2 errors detected in the compilation of "/tmp/tmpxft_0004aa35_00000000-6_comb.cpp1.ii".
make[2]: *** [CMakeFiles/comb.dir/src/comb.cpp.o] error 1
make[1]: *** [CMakeFiles/comb.dir/all] error 2
make: *** [all] error 2
I cannot find cudaGraphExecKernelNodeSetParams
in the code except graph_launch.hpp(154)
. Is it the bug or something I do wrong (like add the -std=c++11
)
error: argument of type "const cudaGraphNode_t *" is incompatible with parameter of type "cudaGraphNode_t *"
is also a problem about the code. How can I fix it?
Thanks for your attention. Looking forward to your reply.
Hi,
I have built this on a system with a single GPU, that I would like to share between two MPI ranks (just for the sake of getting things up and running).
The build basically follows the ubuntu_nvcc10_gcc8
except adjusted for gcc 10.
I built commit e06e54d (the latest at the time of writing).
-- The CXX compiler identification is GNU 10.2.0
-- Check for working CXX compiler: /usr/bin/g++-10
-- Check for working CXX compiler: /usr/bin/g++-10 - works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- BLT Version: 0.3.0
-- CMake Version: 3.17.5
-- CMake Executable: /home/pearson/software/cmake-3.17.5/bin/cmake
-- Found Git: /usr/bin/git (found version "2.28.0")
-- Git Support is ON
-- Git Executable: /usr/bin/git
-- Git Version: 2.28.0
-- MPI Support is ON
-- Enable FindMPI: ON
-- Found MPI_CXX: /home/pearson/software/openmpi-4.0.5/lib/libmpi.so (found version "3.1")
-- Found MPI: TRUE (found version "3.1")
-- BLT MPI Compile Flags: $<$<NOT:$<COMPILE_LANGUAGE:CUDA>>:-pthread>;$<$<COMPILE_LANGUAGE:CUDA>:-Xcompiler=-pthread>
-- BLT MPI Include Paths: /home/pearson/software/openmpi-4.0.5/include
-- BLT MPI Libraries: /home/pearson/software/openmpi-4.0.5/lib/libmpi.so
-- BLT MPI Link Flags: -Wl,-rpath -Wl,/home/pearson/software/openmpi-4.0.5/lib -Wl,--enable-new-dtags -pthread
-- MPI Executable: /home/pearson/software/openmpi-4.0.5/bin/mpiexec
-- MPI Num Proc Flag: -n
-- MPI Command Append:
-- OpenMP Support is OFF
-- CUDA Support is ON
-- The CUDA compiler identification is NVIDIA 11.1.74
-- Check for working CUDA compiler: /usr/local/cuda/bin/nvcc
-- Check for working CUDA compiler: /usr/local/cuda/bin/nvcc - works
-- Detecting CUDA compiler ABI info
-- Detecting CUDA compiler ABI info - done
-- Detecting CUDA compile features
-- Detecting CUDA compile features - done
-- Looking for C++ include pthread.h
-- Looking for C++ include pthread.h - found
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD - Failed
-- Looking for pthread_create in pthreads
-- Looking for pthread_create in pthreads - not found
-- Looking for pthread_create in pthread
-- Looking for pthread_create in pthread - found
-- Found Threads: TRUE
-- Found CUDA: /usr/local/cuda (found version "11.1")
-- CUDA Version: 11.1
-- CUDA Compiler: /usr/local/cuda/bin/nvcc
-- CUDA Host Compiler: /usr/bin/g++-10
-- CUDA Include Path: /usr/local/cuda/include
-- CUDA Libraries: /usr/local/cuda/lib64/libcudart_static.a;Threads::Threads;dl;/usr/lib/x86_64-linux-gnu/librt.so
-- CUDA Compile Flags:
-- CUDA Link Flags:
-- CUDA Separable Compilation: ON
-- CUDA Link with NVCC:
-- HIP Support is OFF
-- HCC Support is OFF
-- Could NOT find Doxygen (missing: DOXYGEN_EXECUTABLE)
-- Sphinx support is ON
-- Failed to locate Sphinx executable (missing: SPHINX_EXECUTABLE)
-- Valgrind support is ON
-- Failed to locate Valgrind executable (missing: VALGRIND_EXECUTABLE)
-- Uncrustify support is ON
-- Failed to locate Uncrustify executable (missing: UNCRUSTIFY_EXECUTABLE)
-- AStyle support is ON
-- Failed to locate AStyle executable (missing: ASTYLE_EXECUTABLE)
-- Cppcheck support is ON
-- Failed to locate Cppcheck executable (missing: CPPCHECK_EXECUTABLE)
-- ClangQuery support is ON
-- Failed to locate ClangQuery executable (missing: CLANGQUERY_EXECUTABLE)
-- C Compiler family is GNU
-- Adding optional BLT definitions and compiler flags
-- Enabling all compiler warnings on all targets.
-- Fortran support disabled.
-- CMAKE_C_FLAGS flags are: -Wall -Wextra
-- CMAKE_CXX_FLAGS flags are: -Wall -Wextra
-- CMAKE_EXE_LINKER_FLAGS flags are:
-- Google Test Support is ON
-- Google Mock Support is OFF
-- The C compiler identification is GNU 10.2.0
-- Check for working C compiler: /usr/bin/gcc-10
-- Check for working C compiler: /usr/bin/gcc-10 - works
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Detecting C compile features
-- Detecting C compile features - done
-- Found PythonInterp: /usr/bin/python3.8 (found version "3.8.6")
-- MPI Enabled
-- Cuda Enabled
-- Configuring done
-- Generating done
-- Build files have been written to: /home/pearson/repos/Comb/build_debian-nvcc11-gcc10
I tried to run it with the following:
~/software/openmpi-4.0.5/bin/mpirun -n 2 bin/comb 10_10_10 -divide 2_1_1 -cuda_aware_mpi -comm enable mpi -exec enable mpi_type -memory enable cuda_device
but I get the following error:
Comb version 0.2.0
Args bin/comb;10_10_10;-divide;2_1_1;-cuda_aware_mpi;-comm;enable;mpi;-exec;enable;mpi_type;-memory;enable;cuda_device
Started rank 0 of 2
Node deneb
Compiler "/usr/bin/g++-10"
Cuda compiler "/usr/local/cuda/bin/nvcc"
Cuda driver version 11010
Cuda runtime version 11010
GPU 0 visible undefined
Cart coords 0 0 0
Message policy cutoff 200
Post Recv using wait_all method
Post Send using wait_all method
Wait Recv using wait_all method
Wait Send using wait_all method
Num cycles 5
Num vars 1
ghost_widths 1 1 1
sizes 10 10 10
divisions 2 1 1
periodic 0 0 0
division map
map 0 0 0
map 5 10 10
map 10
Starting test memcpy seq dst Host src Host
Starting test Comm mock Mesh seq Host Buffers seq Host seq Host
Starting test Comm mock Mesh seq Host Buffers mpi_type Host mpi_type Host
comb: /home/pearson/repos/Comb/include/comm_pol_mock.hpp:948: void detail::MessageGroup<detail::MessageBase::Kind::recv, mock_pol, mpi_type_pol>::Irecv(detail::MessageGroup<detail::MessageBase::Kind::recv, mock_pol, mpi_type_pol>::context_type&, detail::MessageGroup<detail::MessageBase::Kind::recv, mock_pol, mpi_type_pol>::communicator_type&, detail::MessageGroup<detail::MessageBase::Kind::recv, mock_pol, mpi_type_pol>::message_type**, IdxT, detail::MessageGroup<detail::MessageBase::Kind::recv, mock_pol, mpi_type_pol>::request_type*): Assertion `buf != nullptr' failed.
comb: /home/pearson/repos/Comb/include/comm_pol_mock.hpp:948: void detail::MessageGroup<detail::MessageBase::Kind::recv, mock_pol, mpi_type_pol>::Irecv(detail::MessageGroup<detail::MessageBase::Kind::recv, mock_pol, mpi_type_pol>::context_type&, detail::MessageGroup<detail::MessageBase::Kind::recv, mock_pol, mpi_type_pol>::communicator_type&, detail::MessageGroup<detail::MessageBase::Kind::recv, mock_pol, mpi_type_pol>::message_type**, IdxT, detail::MessageGroup<detail::MessageBase::Kind::recv, mock_pol, mpi_type_pol>::request_type*): Assertion `buf != nullptr' failed.
[deneb:23991] *** Process received signal ***
[deneb:23991] Signal: Aborted (6)
[deneb:23991] Signal code: (-6)
[deneb:23992] *** Process received signal ***
[deneb:23992] Signal: Aborted (6)
[deneb:23992] Signal code: (-6)
[deneb:23991] [ 0] /lib/x86_64-linux-gnu/libpthread.so.0(+0x14140)[0x7ffb77f91140]
[deneb:23992] [ 0] /lib/x86_64-linux-gnu/libpthread.so.0(+0x14140)[0x7fd0a2014140]
[deneb:23992] [ 1] [deneb:23991] [ 1] /lib/x86_64-linux-gnu/libc.so.6(gsignal+0x141)[0x7ffb77ac6db1]
[deneb:23991] [ 2] /lib/x86_64-linux-gnu/libc.so.6(gsignal+0x141)[0x7fd0a1b49db1]
[deneb:23992] [ 2] /lib/x86_64-linux-gnu/libc.so.6(abort+0x123)[0x7ffb77ab0537]
[deneb:23991] [ 3] /lib/x86_64-linux-gnu/libc.so.6(abort+0x123)[0x7fd0a1b33537]
[deneb:23992] [ 3] /lib/x86_64-linux-gnu/libc.so.6(+0x2540f)[0x7fd0a1b3340f]
[deneb:23992] [ 4] /lib/x86_64-linux-gnu/libc.so.6(+0x2540f)[0x7ffb77ab040f]
[deneb:23991] [ 4] /lib/x86_64-linux-gnu/libc.so.6(+0x345b2)[0x7fd0a1b425b2]
[deneb:23992] [ 5] bin/comb(+0x32b1d)[0x55baa8e21b1d]
[deneb:23992] [ 6] bin/comb(+0x43508)[0x55baa8e32508]
/lib/x86_64-linux-gnu/libc.so.6(+0x345b2)[0x7ffb77abf5b2]
[deneb:23991] [ 5] bin/comb(+0x32b1d)[0x56455c96ab1d]
[deneb:23991] [ 6] bin/comb(+0x43508)[0x56455c97b508]
[deneb:23991] [ 7] bin/comb(+0x468c6)[0x56455c97e8c6]
[deneb:23991] [ 8] bin/comb(+0x5cdc5)[0x56455c994dc5]
[deneb:23991] [ 9] bin/comb(+0x63303)[0x56455c99b303]
[deneb:23991] [10] bin/comb(+0x635a7)[0x56455c99b5a7]
[deneb:23991] [11] bin/comb(+0xe184)[0x56455c946184]
[deneb:23991] [12] [deneb:23992] [ 7] bin/comb(+0x468c6)[0x55baa8e358c6]
[deneb:23992] [ 8] bin/comb(+0x5cdc5)[0x55baa8e4bdc5]
[deneb:23992] [ 9] bin/comb(+0x63303)[0x55baa8e52303]
[deneb:23992] [10] bin/comb(+0x635a7)[0x55baa8e525a7]
[deneb:23992] [11] bin/comb(+0xe184)[0x55baa8dfd184]
[deneb:23992] [12] /lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xea)[0x7fd0a1b34cca]
[deneb:23992] [13] bin/comb(+0xf41a)[0x55baa8dfe41a]
/lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xea)[0x7ffb77ab1cca]
[deneb:23991] [13] bin/comb(+0xf41a)[0x56455c94741a]
[deneb:23991] *** End of error message ***
[deneb:23992] *** End of error message ***
--------------------------------------------------------------------------
Primary job terminated normally, but 1 process returned
a non-zero exit code. Per user-direction, the job has been aborted.
--------------------------------------------------------------------------
--------------------------------------------------------------------------
mpirun noticed that process rank 1 with PID 0 on node deneb exited on signal 6 (Aborted).
--------------------------------------------------------------------------
I also managed ot run the focused tests:
cd <build>/bin
../../scripts/run_tests.bash 1 ../../scripts/focused_tests.bash
which appears to have worked with the following output:
mpirun -np 1 /home/pearson/repos/Comb/build_debian-nvcc11-gcc10/bin/comb -comm post_recv wait_any -comm post_send wait_any -comm wait_recv wait_any -comm wait_send wait_any 100_100_100 -divide 1_1_1 -periodic 1_1_1 -ghost 1_1_1 -vars 3 -cycles 25 -comm cutoff 250 -omp_threads 10 -exec disable seq -exec enable cuda -memory disable host -memory enable cuda_managed -comm enable mock -comm enable mpi
Comb version 0.2.0
Args /home/pearson/repos/Comb/build_debian-nvcc11-gcc10/bin/comb;-comm;post_recv;wait_any;-comm;post_send;wait_any;-comm;wait_recv;wait_any;-comm;wait_send;wait_any;100_100_100;-divide;1_1_1;-periodic;1_1_1;-ghost;1_1_1;-vars;3;-cycles;25;-comm;cutoff;250;-omp_threads;10;-exec;disable;seq;-exec;enable;cuda;-memory;disable;host;-memory;enable;cuda_managed;-comm;enable;mock;-comm;enable;mpi
Started rank 0 of 1
Node deneb
Compiler "/usr/bin/g++-10"
Cuda compiler "/usr/local/cuda/bin/nvcc"
Cuda driver version 11010
Cuda runtime version 11010
GPU 0 visible undefined
Not built with openmp, ignoring -omp_threads 10.
Cart coords 0 0 0
Message policy cutoff 250
Post Recv using wait_any method
Post Send using wait_any method
Wait Recv using wait_any method
Wait Send using wait_any method
Num cycles 25
Num vars 3
ghost_widths 1 1 1
sizes 100 100 100
divisions 1 1 1
periodic 1 1 1
division map
map 0 0 0
map 100 100 100
Starting test memcpy cuda dst Managed src HostPinned
Starting test memcpy cuda dst Managed src Device
Starting test Comm mock Mesh cuda Managed Buffers cuda HostPinned cuda HostPinned
Starting test Comm mpi Mesh cuda Managed Buffers cuda HostPinned cuda HostPinned
done
real 0m1.475s
user 0m1.244s
sys 0m0.146s
Is device memory + MPI + MPI_Type a supported configuration at this time? If so, any advice?
Thanks!
Create a spack package.
Apologies if this is a CMake n00b problem.
I've been trying to configure on my system with this script (based on the ones in your source tree):
#!/bin/bash
BUILD_SUFFIX=rndv
rm -rf build_${BUILD_SUFFIX} 2>/dev/null
mkdir build_${BUILD_SUFFIX}
cd build_${BUILD_SUFFIX}
PATH=/path/to/mpi/bin:$PATH \
LD_LIBRARY_PATH=/path/to/mpi/lib:$LD_LIBRARY_PATH \
cmake \
-DCMAKE_BUILD_TYPE=Release \
-DENABLE_OPENMP=OFF \
-DENABLE_CUDA=OFF \
-DCMAKE_C_FLAGS=-pthread \
-DCMAKE_CXX_FLAGS=-pthread \
-DCMAKE_EXE_LINKER_FLAGS=-pthread \
-C ../host-configs/ansklnvram01/gcc_${BUILD_SUFFIX}.cmake \
-DCMAKE_INSTALL_PREFIX=../install_${BUILD_SUFFIX} \
"$@" \
-Wno-dev \
..
When I didn't include the -pthread
flags, I kept getting errors about not being able to find the pthread symbols.
Now my output looks like this:
$ ./scripts/my-builds/gcc_rndv.sh
loading initial cache file ../host-configs/ansklnvram01/gcc_rndv.cmake
CMake Error: Error processing file: ../host-configs/ansklnvram01/gcc_rndv.cmake
-- The CXX compiler identification is GNU 7.3.1
-- Check for working CXX compiler: /opt/rh/devtoolset-7/root/usr/bin/c++
-- Check for working CXX compiler: /opt/rh/devtoolset-7/root/usr/bin/c++ -- works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Found Git: /nfs/site/home/wbland/home/tools/x86/bin/git (found version "2.19.2")
-- Git Support is ON
-- Git Executable: /nfs/site/home/wbland/home/tools/x86/bin/git
-- Git Version: 2.19.2
-- MPI Support is ON
-- Found MPI_CXX: /path/to/mpi/lib/libmpicxx.so (found version "3.1")
-- Found MPI: TRUE (found version "3.1")
-- MPI C Compile Flags:
-- MPI C Include Path:
-- MPI C Link Flags:
-- MPI C Libraries:
-- MPI CXX Compile Flags:
-- MPI CXX Include Path: /path/to/mpi/include;/path/to/mpi/include
-- MPI CXX Link Flags: -Wl,-rpath -Wl,/path/to/mpi/lib -Wl,--enable-new-dtags -L/usr/lib64 -L/path/to/mpi/lib
-- MPI CXX Libraries: /path/to/mpi/lib/libmpi.so
-- MPI Executable: /path/to/mpi/bin/mpiexec
-- MPI Num Proc Flag: -n
-- MPI Command Append:
-- CUDA Support is OFF
-- ROCM Support is OFF
-- Found Doxygen: /usr/bin/doxygen (found version "1.8.5") found components: doxygen dot
-- Sphinx support is ON
-- Failed to locate Sphinx executable (missing: SPHINX_EXECUTABLE)
-- Valgrind support is ON
-- Found Valgrind: /opt/rh/devtoolset-7/root/usr/bin/valgrind
-- Uncrustify support is ON
-- Failed to locate Uncrustify executable (missing: UNCRUSTIFY_EXECUTABLE)
-- AStyle support is ON
-- Failed to locate AStyle executable (missing: ASTYLE_EXECUTABLE)
-- Cppcheck support is ON
-- Failed to locate Cppcheck executable (missing: CPPCHECK_EXECUTABLE)
-- C Compiler family is GNU
-- OpenMP Support is OFF
-- Adding optional BLT definitions and compiler flags
-- Standard C++11 selected
-- Enabling all compiler warnings on all targets.
-- Fortran support disabled.
-- CMAKE_C_FLAGS flags are: -pthread -Wall -Wextra
-- CMAKE_CXX_FLAGS flags are: -pthread -Wall -Wextra
-- CMAKE_EXE_LINKER_FLAGS flags are: -pthread
-- Google Test Support is ON
-- Google Mock Support is OFF
-- The C compiler identification is GNU 7.3.1
-- Check for working C compiler: /opt/rh/devtoolset-7/root/usr/bin/cc
-- Check for working C compiler: /opt/rh/devtoolset-7/root/usr/bin/cc -- works
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Detecting C compile features
-- Detecting C compile features - done
-- Found PythonInterp: /usr/bin/python (found version "2.7.5")
-- Looking for pthread.h
-- Looking for pthread.h - found
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD - Success
-- Found Threads: TRUE
-- MPI Enabled
-- Configuring incomplete, errors occurred!
See also "/path/to/Comb/build_rndv/CMakeFiles/CMakeOutput.log".
So it's telling me there's errors, but I don't actually see any errors and nothing is showing up in that log.
I'm assuming this is probably caused by me manually requiring the pthreads flag, but I'd love it if you could give me a pointer where to look.
Add an output file in csv format.
You can't use arbitrary memory spaces with buffers, currently the code will only allows using one or two memory spaces if the execution space is host or device. For example with cuda you can only use device or pinned memory for communication buffers.
Add some kind of input file to specify the tests to run.
I'm not sure if this should be in addition to or instead of command line input.
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.