Closed anshumang closed 9 years ago
can you pls try again with the actual latest release branch, which is release-0.1.0
?
Can you pls post the output of a fresh cmake run with that, too?
I checked out the release-0.1.0 branch but still getting the same error. Please find below the output from the cmake run =>
agoswami@shiva:~/computationalRadiationPhysics/picongpu/build-temp$ ../configure ../examples/KelvinHelmholtz/ cmake command: cmake -DCUDA_ARCH=sm_20 -DCMAKE_INSTALL_PREFIX=../examples/KelvinHelmholtz/ -DPIC_EXTENSION_PATH=../examples/KelvinHelmholtz/ .. -- Found CUDA: /usr/local/cuda-6.0 (found suitable version "6.0", minimum required is "5.0") -- Debug version -- Boost version: 1.54.0 -- Found the following Boost libraries: -- program_options -- regex -- filesystem -- system -- Found CUDA: /usr/local/cuda-6.0 (found version "6.0") -- Could NOT find NVML (missing: NVML_INCLUDE_DIR) -- Boost version: 1.54.0 -- Found the following Boost libraries: -- program_options -- Can NOT find 'adios_config' - set ADIOS_ROOT or check your PATH -- Could NOT find ADIOS (missing: ADIOS_LIBRARIES ADIOS_INCLUDE_DIRS) (Required is at least version "1.6.0") -- Can NOT find libSplash for HDF5 output - set SPLASH_ROOT -- Could NOT find Splash (missing: Splash_LIBRARIES Splash_INCLUDE_DIRS) (Required is at least version "1.2.4") -- Can NOT find PNGwriter - set PNGWRITER_ROOT -- Could NOT find PNGwriter (missing: PNGwriter_LIBRARIES PNGwriter_INCLUDE_DIRS) -- Configuring done -- Generating done -- Build files have been written to: /home/agoswami/computationalRadiationPhysics/picongpu/build-temp
It's gotta be due to some change in my environment because I am getting the same error in my existing work area for PIConGPU, too. I tried re-running cmake and building on another machine and it works fine so I should be able to proceed for now. But I am really interested to find out the cause of this error.
can you pls clean the build directory and run cmake again, the output is very short which means it did take an old set.
can you create a fresh case with createParameterSet
so we can be sure it's not an incompatibility from an update of the code base (with an old param set)? It looks a bit like you are using an old setup/case but updated the PIConGPU main source.
Still getting the error, here are my steps - 1) Download the latest release and extract 2) cd picongpu-beta-rc6 3) mkdir temp-build temp-build/paramSets 4) cd temp-build 5) ../createParameterSet ../examples/KelvinHelmholtz ./paramSets/KH 6) ../configure ./paramSets/KH 7) make -j12
If you go to the source directory where you extraced PIConGPU to, what does
find ./src -name TypeCast*
print?
Can you please try it again with the release-branch of 0.1.0
?
git clone https://github.com/ComputationalRadiationPhysics/picongpu.git
cd picongpu
git checkout -b release-0.1.0 origin/release-0.1.0
what is your compiler and operating system?
It prints - ./src/libPMacc/include/algorithms/TypeCast.hpp
I did a fresh clone and checked out the release branch. And then followed steps 3 to 7 from above. Still the same error :(
host compiler => g++ (Ubuntu 4.8.2-19ubuntu1) 4.8.2 cuda compiler => Cuda compilation tools, release 6.5, V6.5.12 OS => Ubuntu 14.04.1 LTS
found with @psychocoderHPC: wait, you do build in-source!
Step 3 and 4: do not create your build-temp
directory in a sub-folder of the picongpu/
source directory.
Also do not create the param directory in the build directory
also, please avoid building an example directly but run createParameterSet
to clone from it - into an other directory that is again nota sub-directory of picongpu, first.
does that fix it?
Please try to follow the instructions in our INSTALL file, there are many steps mixed and many directories not correctly placed :) Please use the directory structure we suggest in Step 1 :)
Sorry I deviated from the INSTALL.md recommendations slightly .... but the build-temp and param-sets-temp directory placements never created any trouble before...
So I created two directories, build-temp and param-sets-temp under the same top-level directory where I cloned (and checked out the release branch of) picongpu. This is the directory structure => top-level |--- picongpu |--- build-temp |--- param-sets-temp
cd build-temp ../picongpu/createParameterSet ../picongpu/examples/KelvinHelmholtz ../param-sets-temp/KH
../picongpu/configure ../param-sets-temp/KH
Output=>
cmake command: cmake -DCUDA_ARCH=sm_20 -DCMAKE_INSTALL_PREFIX=../param-sets-temp/KH/ -DPIC_EXTENSION_PATH=../param-sets-temp/KH/ ../picongpu
-- The C compiler identification is GNU 4.8.2
-- The CXX compiler identification is GNU 4.8.2
-- Check for working C compiler: /usr/bin/cc
-- Check for working C compiler: /usr/bin/cc -- works
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Check for working CXX compiler: /usr/bin/c++
-- Check for working CXX compiler: /usr/bin/c++ -- works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Found CUDA: /usr/local/cuda (found suitable version "6.5", minimum required is "5.0")
-- Debug version
-- Found MPI_C: /usr/lib/libmpi.so;/usr/lib/x86_64-linux-gnu/libdl.so;/usr/lib/x86_64-linux-gnu/libhwloc.so
-- Found MPI_CXX: /usr/lib/libmpi_cxx.so;/usr/lib/libmpi.so;/usr/lib/x86_64-linux-gnu/libdl.so;/usr/lib/x86_64-linux-gnu/libhwloc.so
-- Looking for include file pthread.h
-- Looking for include file 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
-- Boost version: 1.54.0
-- Found the following Boost libraries:
-- program_options
-- regex
-- filesystem
-- system
-- Found CUDA: /usr/local/cuda (found version "6.5")
-- Could NOT find NVML (missing: NVML_INCLUDE_DIR)
-- Boost version: 1.54.0
-- Found the following Boost libraries:
-- program_options
-- Can NOT find 'adios_config' - set ADIOS_ROOT or check your PATH
-- Could NOT find ADIOS (missing: ADIOS_LIBRARIES ADIOS_INCLUDE_DIRS) (Required is at least version "1.6.0")
-- Can NOT find libSplash for HDF5 output - set SPLASH_ROOT
-- Could NOT find Splash (missing: Splash_LIBRARIES Splash_INCLUDE_DIRS) (Required is at least version "1.2.4")
-- Found ZLIB: /usr/lib/x86_64-linux-gnu/libz.so (found version "1.2.8")
-- Found PNG: /usr/lib/x86_64-linux-gnu/libpng.so (found suitable version "1.2.50", minimum required is "1.2.9")
-- Can NOT find PNGwriter - set PNGWRITER_ROOT
-- Could NOT find PNGwriter (missing: PNGwriter_LIBRARIES PNGwriter_INCLUDE_DIRS)
-- Configuring done
-- Generating done
-- Build files have been written to: /home/agoswami/computationalRadiationPhysics/release-branch/build-temp
make -j12
Output=> [ 16%] [ 33%] [ 50%] Scanning dependencies of target mpiInfo Building NVCC (Device) object build_picongpu/CMakeFiles/picongpu.dir//./picongpu_generated_main.cu.o Building NVCC (Device) object build_picongpu/build_cuda_memtest/CMakeFiles/cuda_memtest.dir//./cuda_memtest_generated_cuda_memtest.cu.o Building NVCC (Device) object build_picongpu/build_cuda_memtest/CMakeFiles/cuda_memtest.dir//./cuda_memtest_generated_tests.cu.o [ 66%] Building CXX object build_picongpu/build_mpiInfo/CMakeFiles/mpiInfo.dir/main.cpp.o In file included from /home/agoswami/computationalRadiationPhysics/release-branch/picongpu/src/picongpu/include/simulation_defines.hpp:26:0, from /home/agoswami/computationalRadiationPhysics/release-branch/picongpu/src/picongpu/main.cu:32: /home/agoswami/computationalRadiationPhysics/release-branch/picongpu/src/picongpu/include/simulation_types.hpp:25:35: fatal error: algorithms/TypeCast.hpp: No such file or directory
^
compilation terminated. CMake Error at picongpu_generated_main.cu.o.cmake:206 (message): Error generating /home/agoswami/computationalRadiationPhysics/release-branch/build-temp/build_picongpu/CMakeFiles/picongpu.dir//./picongpu_generated_main.cu.o
make[2]: * [build_picongpu/CMakeFiles/picongpu.dir/./picongpu_generated_main.cu.o] Error 1 make[1]: * [build_picongpu/CMakeFiles/picongpu.dir/all] Error 2 make[1]: * Waiting for unfinished jobs.... Linking CXX executable mpiInfo [ 66%] Built target mpiInfo Scanning dependencies of target cuda_memtest [ 83%] Building CXX object build_picongpu/build_cuda_memtest/CMakeFiles/cuda_memtest.dir/misc.cpp.o Linking CXX executable cuda_memtest [ 83%] Built target cuda_memtest make: * [all] Error 2
Still the same error !!! I am at a loss :(
Please run cmake -L
in your build folder, git status
in the picongpu folder and post the output.
Please also run make VERBOSE=1
from your build folder.
Thanks @psychocoderHPC for the pointers....sorry I was unable to respond earlier... Here are the steps and their corresponding outputs => git status On branch release-0.1.0 Your branch is up-to-date with 'origin/release-0.1.0'.
Untracked files:
(use "git add
temp-build/
nothing added to commit but untracked files present (use "git add" to track)
cmake -L => CMake Error: The source directory "/home/agoswami/computationalRadiationPhysics/release-branch/build-temp" does not appear to contain CMakeLists.txt. Specify --help for usage, or press the help button on the CMake GUI. -- Cache values ADIOS_CONFIG:FILEPATH=ADIOS_CONFIG-NOTFOUND CMAKE_BUILD_TYPE:STRING= CMAKE_INSTALL_PREFIX:PATH=/home/agoswami/computationalRadiationPhysics/release-branch/param-sets-temp/KH CUDA_ARCH:STRING=sm_20 CUDA_BUILD_CUBIN:BOOL=OFF CUDA_BUILD_EMULATION:BOOL=OFF CUDA_FTZ:STRING=--ftz=false CUDA_HOST_COMPILER:FILEPATH=/usr/bin/cc CUDA_KEEP_FILES:BOOL=OFF CUDA_MATH:STRING=--use_fast_math CUDA_MEMTEST_DIR:PATH=/home/agoswami/computationalRadiationPhysics/release-branch/picongpu/src/cuda_memtest CUDA_MEMTEST_RELEASE:BOOL=ON CUDA_SDK_ROOT_DIR:PATH=CUDA_SDK_ROOT_DIR-NOTFOUND CUDA_SEPARABLE_COMPILATION:BOOL=OFF CUDA_SHOW_CODELINES:BOOL=OFF CUDA_SHOW_REGISTER:BOOL=OFF CUDA_TOOLKIT_ROOT_DIR:PATH=/usr/local/cuda CUDA_VERBOSE_BUILD:BOOL=OFF MPI_EXTRA_LIBRARY:STRING=/usr/lib/libmpi.so;/usr/lib/x86_64-linux-gnu/libdl.so;/usr/lib/x86_64-linux-gnu/libhwloc.so MPI_INFO_DIR:PATH=/home/agoswami/computationalRadiationPhysics/release-branch/picongpu/src/mpiInfo MPI_LIBRARY:FILEPATH=/usr/lib/libmpi_cxx.so PIC_COPY_ON_INSTALL:STRING=include/simulation_defines;submit PIC_ENABLE_INSITU_VOLVIS:BOOL=OFF PIC_EXTENSION_PATH:PATH=/home/agoswami/computationalRadiationPhysics/release-branch/param-sets-temp/KH PIC_RELEASE:BOOL=OFF PIC_VERBOSE:STRING=1 PMACC_BLOCKING_KERNEL:BOOL=OFF PMACC_ROOT_DIR:PATH=/usr/local PMACC_VERBOSE:STRING=0 PNGwriter_ROOT_DIR:PATH=PNGwriter_ROOT_DIR-NOTFOUND SCOREP_ENABLE:BOOL=OFF Splash_ROOT_DIR:PATH=Splash_ROOT_DIR-NOTFOUND VAMPIR_ENABLE:BOOL=OFF VT_INST_FILE_FILTER:STRING=stl,usr/include,libgpugrid,vector_types.h,Vector.hpp,DeviceBuffer.hpp,DeviceBufferIntern.hpp,Buffer.hpp,StrideMapping.hpp,StrideMappingMethods.hpp,MappingDescription.hpp,AreaMapping.hpp,AreaMappingMethods.hpp,ExchangeMapping.hpp,ExchangeMappingMethods.hpp,DataSpace.hpp,Manager.hpp,Manager.tpp,Transaction.hpp,Transaction.tpp,TransactionManager.hpp,TransactionManager.tpp,Vector.tpp,Mask.hpp,ITask.hpp,EventTask.hpp,EventTask.tpp,StandartAccessor.hpp,StandartNavigator.hpp,HostBuffer.hpp,HostBufferIntern.hpp VT_INST_FUNC_FILTER:STRING=vector,Vector,dim3,GPUGrid,execute,allocator,Task,Manager,Transaction,Mask,operator,DataSpace,PitchedBox,Event,new,getGridDim,GetCurrentDataSpaces,MappingDescription,getOffset,getParticlesBuffer,getDataSpace,getInstance
make VERBOSE=1 =>
/usr/bin/cmake -H/home/agoswami/computationalRadiationPhysics/release-branch/picongpu -B/home/agoswami/computationalRadiationPhysics/release-branch/build-temp --check-build-system CMakeFiles/Makefile.cmake 0
/usr/bin/cmake -E cmake_progress_start /home/agoswami/computationalRadiationPhysics/release-branch/build-temp/CMakeFiles /home/agoswami/computationalRadiationPhysics/release-branch/build-temp/CMakeFiles/progress.marks
make -f CMakeFiles/Makefile2 all
make[1]: Entering directory /home/agoswami/computationalRadiationPhysics/release-branch/build-temp' make -f build_picongpu/CMakeFiles/picongpu.dir/build.make build_picongpu/CMakeFiles/picongpu.dir/depend make[2]: Entering directory
/home/agoswami/computationalRadiationPhysics/release-branch/build-temp'
/usr/bin/cmake -E cmake_progress_report /home/agoswami/computationalRadiationPhysics/release-branch/build-temp/CMakeFiles 5
[ 16%] Building NVCC (Device) object build_picongpu/CMakeFiles/picongpu.dir//./picongpu_generated_main.cu.o
cd /home/agoswami/computationalRadiationPhysics/release-branch/build-temp/build_picongpu/CMakeFiles/picongpu.dir && /usr/bin/cmake -E make_directory /home/agoswami/computationalRadiationPhysics/release-branch/build-temp/build_picongpu/CMakeFiles/picongpu.dir//.
cd /home/agoswami/computationalRadiationPhysics/release-branch/build-temp/build_picongpu/CMakeFiles/picongpu.dir && /usr/bin/cmake -D verbose:BOOL=1 -D build_configuration:STRING=Debug -D generated_file:STRING=/home/agoswami/computationalRadiationPhysics/release-branch/build-temp/build_picongpu/CMakeFiles/picongpu.dir//./picongpu_generated_main.cu.o -D generated_cubin_file:STRING=/home/agoswami/computationalRadiationPhysics/release-branch/build-temp/build_picongpu/CMakeFiles/picongpu.dir//./picongpu_generated_main.cu.o.cubin.txt -P /home/agoswami/computationalRadiationPhysics/release-branch/build-temp/build_picongpu/CMakeFiles/picongpu.dir//picongpu_generated_main.cu.o.cmake
-- Removing /home/agoswami/computationalRadiationPhysics/release-branch/build-temp/build_picongpu/CMakeFiles/picongpu.dir//./picongpu_generated_main.cu.o
/usr/bin/cmake -E remove /home/agoswami/computationalRadiationPhysics/release-branch/build-temp/build_picongpu/CMakeFiles/picongpu.dir//./picongpu_generated_main.cu.o
-- Generating dependency file: /home/agoswami/computationalRadiationPhysics/release-branch/build-temp/build_picongpu/CMakeFiles/picongpu.dir//picongpu_generated_main.cu.o.NVCC-depend
/usr/local/cuda/bin/nvcc -M -DCUDACC /home/agoswami/computationalRadiationPhysics/release-branch/picongpu/src/picongpu/main.cu -o /home/agoswami/computationalRadiationPhysics/release-branch/build-temp/build_picongpu/CMakeFiles/picongpu.dir//picongpu_generated_main.cu.o.NVCC-depend -ccbin /usr/bin/cc -m64 -DPMACC_VERBOSE_LVL=0 -DPIC_VERBOSE_LVL=1 -Xcompiler ,\"-Wall\",\"-Wno-unknown-pragmas\",\"-Wextra\",\"-Wno-unused-parameter\",\"-Wno-unused-local-typedefs\",\"-g\" -arch=sm_20 --use_fast_math --ftz=false -g -Xcompiler=-g,-pthread -DNVCC -I/usr/local/cuda/include -I/home/agoswami/computationalRadiationPhysics/release-branch/param-sets-temp/KH/include -I/usr/lib/openmpi/include -I/usr/lib/openmpi/include/openmpi -I/usr/include -I/usr/local/include -I/home/agoswami/computationalRadiationPhysics/release-branch/picongpu/src/picongpu/include -I/usr/local/cuda/include
In file included from /home/agoswami/computationalRadiationPhysics/release-branch/picongpu/src/picongpu/include/simulation_defines.hpp:26:0,
from /home/agoswami/computationalRadiationPhysics/release-branch/picongpu/src/picongpu/main.cu:32:
/home/agoswami/computationalRadiationPhysics/release-branch/picongpu/src/picongpu/include/simulation_types.hpp:25:35: fatal error: algorithms/TypeCast.hpp: No such file or directory
^
compilation terminated. CMake Error at picongpu_generated_main.cu.o.cmake:206 (message): Error generating /home/agoswami/computationalRadiationPhysics/release-branch/build-temp/build_picongpu/CMakeFiles/picongpu.dir//./picongpu_generated_main.cu.o
make[2]: * [build_picongpu/CMakeFiles/picongpu.dir/./picongpu_generated_main.cu.o] Error 1 make[2]: Leaving directory `/home/agoswami/computationalRadiationPhysics/release-branch/build-temp' make[1]: * [build_picongpu/CMakeFiles/picongpu.dir/all] Error 2 make[1]: Leaving directory`/home/agoswami/computationalRadiationPhysics/release-branch/build-temp' make: *\ [all] Error 2
Shouldn't there also be a "-I/home/agoswami/computationalRadiationPhysics/release-branch/picongpu/src/libPMacc/include" in the nvcc flags?
Thanks for your help.
Yes you are right the PMacc
path is missing/wrong.
cmake -L .
showed that the path is set to wrong directory.
...
PIC_VERBOSE:STRING=1
PMACC_BLOCKING_KERNEL:BOOL=OFF
PMACC_ROOT_DIR:PATH=/usr/local
PMACC_VERBOSE:STRING=0
...
I will check why this can happened. A workaround for the moment:
../picongpu/configure -c"-DPMACC_ROOT_DIR=../picongpu/src/libPMacc" ../param-sets-temp/KH
We should check the cmake version: please run cmake --version
Ok it is a BUG here
Please run ls -la /usr/local/include/types.h
to verify that I am right.
I will fix it on monday.
well spotted you two!
@psychocoderHPC dah, I should have realized that earlier! Very well spotted!
We should either just set the path directly without find_
since we know it or choose a more significant file... it is also possible to exclude system path's, but that only postpones the problem.
an even better thing would be to add a top-folder such as PMacc/
since an install to /usr
of libPMacc
in the future will always risk naming collisions.
@anshumang thank you tremendously for your help! You did a great job!
@ax3l @psychocoderHPC => Thanks both of you for following up and :+1: for tracing down the bug...
@psychocoderHPC cmake version 2.8.12.2
@psychocoderHPC Indeed, the difference between the two boxes are ls -la /usr/local/include/types.h - the one where it runs does not have such a file
fixed in dev
now and will be fixed in the next release, too (0.1.0).
I am getting the following error on the latest release beta-rc6 -
[ 16%] [ 33%] [ 50%] Scanning dependencies of target mpiInfo Building NVCC (Device) object build_picongpu/CMakeFiles/picongpu.dir//./picongpu_generated_main.cu.o Building NVCC (Device) object build_picongpu/build_cuda_memtest/CMakeFiles/cuda_memtest.dir//./cuda_memtest_generated_tests.cu.o Building NVCC (Device) object build_picongpu/build_cuda_memtest/CMakeFiles/cuda_memtest.dir//./cuda_memtest_generated_cuda_memtest.cu.o [ 66%] Building CXX object build_picongpu/build_mpiInfo/CMakeFiles/mpiInfo.dir/main.cpp.o In file included from /home/agoswami/computationalRadiationPhysics/picongpu-beta-rc6/src/picongpu/include/simulation_defines.hpp:26:0, from /home/agoswami/computationalRadiationPhysics/picongpu-beta-rc6/src/picongpu/main.cu:32: /home/agoswami/computationalRadiationPhysics/picongpu-beta-rc6/src/picongpu/include/simulation_types.hpp:25:35: fatal error: algorithms/TypeCast.hpp: No such file or directory
include "algorithms/TypeCast.hpp"
compilation terminated. CMake Error at picongpu_generated_main.cu.o.cmake:206 (message): Error generating /home/agoswami/computationalRadiationPhysics/picongpu-beta-rc6/temp-build/build_picongpu/CMakeFiles/picongpu.dir//./picongpu_generated_main.cu.o
make[2]: * [build_picongpu/CMakeFiles/picongpu.dir/./picongpu_generated_main.cu.o] Error 1 make[1]: * [build_picongpu/CMakeFiles/picongpu.dir/all] Error 2 make[1]: ** Waiting for unfinished jobs.... /home/agoswami/computationalRadiationPhysics/picongpu-beta-rc6/src/cuda_memtest/cuda_memtest.cu: In function ‘void display_deviceinfo(cudaDeviceProp)’: /home/agoswami/computationalRadiationPhysics/picongpu-beta-rc6/src/cuda_memtest/cuda_memtest.cu:113:212: warning: format ‘%llu’ expects argument of type ‘long long unsigned int’, but argument 7 has type ‘size_t {aka long unsigned int}’ [-Wformat=] PRINTF("Device name=%s, global memory size=%llu\n", prop->name, prop->totalGlobalMem); ^ /home/agoswami/computationalRadiationPhysics/picongpu-beta-rc6/src/cuda_memtest/cuda_memtest.cu:113:413: warning: format ‘%llu’ expects argument of type ‘long long unsigned int’, but argument 6 has type ‘size_t {aka long unsigned int}’ [-Wformat=] PRINTF("Device name=%s, global memory size=%llu\n", prop->name, prop->totalGlobalMem); ^ Scanning dependencies of target cuda_memtest [ 83%] Building CXX object build_picongpu/build_cuda_memtest/CMakeFiles/cuda_memtest.dir/misc.cpp.o Linking CXX executable cuda_memtest [ 83%] Built target cuda_memtest Linking CXX executable mpiInfo [ 83%] Built target mpiInfo make: _* [all] Error 2