Compilation of Nektar++ 4.4.1 fails with "DSO missing from command line"
Hi all I am trying to install Nektar++ for the first time, on Linux 4.18.6 with CMake version 3.12.1. The compilation fails with: [ 66%] Linking CXX executable CardiacEPSolver /usr/bin/ld: CMakeFiles/CardiacEPSolver.dir/CellModels/CellModel.cpp.o: undefined reference to symbol 'pthread_condattr_setclock@@GLIBC_2.3.3' /usr/bin/ld: /usr/lib/libpthread.so.0: error adding symbols: DSO missing from command line collect2: error: ld returned 1 exit status make[2]: *** [solvers/CardiacEPSolver/CMakeFiles/CardiacEPSolver.dir/build.make:465: solvers/CardiacEPSolver/CardiacEPSolver] Error 1 make[1]: *** [CMakeFiles/Makefile2:2907: solvers/CardiacEPSolver/CMakeFiles/CardiacEPSolver.dir/all] Error 2 make: *** [Makefile:163: all] Error 2 if I disable CardiacEPSolver in the configuration, I get a similar error: [ 70%] Linking CXX executable CompressibleFlowSolver /usr/bin/ld: CMakeFiles/CompressibleFlowSolver.dir/ArtificialDiffusion/ArtificialDiffusion.cpp.o: undefined reference to symbol 'pthread_condattr_setclock@@GLIBC_2.3.3' /usr/bin/ld: /usr/lib/libpthread.so.0: error adding symbols: DSO missing from command line collect2: error: ld returned 1 exit status make[2]: *** [solvers/CompressibleFlowSolver/CMakeFiles/CompressibleFlowSolver.dir/build.make:615: solvers/CompressibleFlowSolver/CompressibleFlowSolver] Error 1 make[1]: *** [CMakeFiles/Makefile2:2925: solvers/CompressibleFlowSolver/CMakeFiles/CompressibleFlowSolver.dir/all] Error 2 make: *** [Makefile:163: all] Error 2 Can someone point me in the right direction? Thanks, Mikkel
Hi Mikkel, Which Linux distribution/version are you using? Could you send the output of running 'make VERBOSE=1'. Cheers, Chris On Tue, 11 Sep 2018 13:05:10 +0100, Mikkel Bue Lykkegaard <mikkelbue@gmail.com> wrote:
Hi all
I am trying to install Nektar++ for the first time, on Linux 4.18.6 with CMake version 3.12.1.
The compilation fails with:
[ 66%] Linking CXX executable CardiacEPSolver /usr/bin/ld: CMakeFiles/CardiacEPSolver.dir/CellModels/CellModel.cpp.o: undefined reference to symbol 'pthread_condattr_setclock@@GLIBC_2.3.3' /usr/bin/ld: /usr/lib/libpthread.so.0: error adding symbols: DSO missing from command line collect2: error: ld returned 1 exit status make[2]: *** [solvers/CardiacEPSolver/CMakeFiles/CardiacEPSolver.dir/build.make:465: solvers/CardiacEPSolver/CardiacEPSolver] Error 1 make[1]: *** [CMakeFiles/Makefile2:2907: solvers/CardiacEPSolver/CMakeFiles/CardiacEPSolver.dir/all] Error 2 make: *** [Makefile:163: all] Error 2
if I disable CardiacEPSolver in the configuration, I get a similar error:
[ 70%] Linking CXX executable CompressibleFlowSolver /usr/bin/ld: CMakeFiles/CompressibleFlowSolver.dir/ArtificialDiffusion/ArtificialDiffusion.cpp.o: undefined reference to symbol 'pthread_condattr_setclock@@GLIBC_2.3.3' /usr/bin/ld: /usr/lib/libpthread.so.0: error adding symbols: DSO missing from command line collect2: error: ld returned 1 exit status make[2]: *** [solvers/CompressibleFlowSolver/CMakeFiles/CompressibleFlowSolver.dir/build.make:615: solvers/CompressibleFlowSolver/CompressibleFlowSolver] Error 1 make[1]: *** [CMakeFiles/Makefile2:2925: solvers/CompressibleFlowSolver/CMakeFiles/CompressibleFlowSolver.dir/all] Error 2 make: *** [Makefile:163: all] Error 2
Can someone point me in the right direction?
Thanks, Mikkel
-- Chris Cantwell Imperial College London South Kensington Campus London SW7 2AZ Email: c.cantwell@imperial.ac.uk www.imperial.ac.uk/people/c.cantwell
Hi Chris Thanks for the reply. Sure, here is the output of 'make VERBOSE=1': https://pastebin.com/AtkuDCCi Also, here are the messages from ccmake (in case it has any relevance): https://pastebin.com/KpKYeDNM I am running Antergos (Arch) Linux with the following package versions: Linux 4.18.6 gcc 8.2.1 CMake 3.12.1 binutils 2.31.1 BLAS 3.8.0 LAPACK 3.8.0 boost 1.67.0 Mikkel On Tue, Sep 11, 2018 at 10:37 PM Chris Cantwell <c.cantwell@imperial.ac.uk> wrote:
Hi Mikkel,
Which Linux distribution/version are you using?
Could you send the output of running 'make VERBOSE=1'.
Cheers, Chris
On Tue, 11 Sep 2018 13:05:10 +0100, Mikkel Bue Lykkegaard < mikkelbue@gmail.com> wrote:
Hi all
I am trying to install Nektar++ for the first time, on Linux 4.18.6 with CMake version 3.12.1.
The compilation fails with:
[ 66%] Linking CXX executable CardiacEPSolver /usr/bin/ld: CMakeFiles/CardiacEPSolver.dir/CellModels/CellModel.cpp.o: undefined reference to symbol 'pthread_condattr_setclock@@GLIBC_2.3.3' /usr/bin/ld: /usr/lib/libpthread.so.0: error adding symbols: DSO missing from command line collect2: error: ld returned 1 exit status make[2]: *** [solvers/CardiacEPSolver/CMakeFiles/CardiacEPSolver.dir/build.make:465: solvers/CardiacEPSolver/CardiacEPSolver] Error 1 make[1]: *** [CMakeFiles/Makefile2:2907: solvers/CardiacEPSolver/CMakeFiles/CardiacEPSolver.dir/all] Error 2 make: *** [Makefile:163: all] Error 2
if I disable CardiacEPSolver in the configuration, I get a similar error:
[ 70%] Linking CXX executable CompressibleFlowSolver /usr/bin/ld:
CMakeFiles/CompressibleFlowSolver.dir/ArtificialDiffusion/ArtificialDiffusion.cpp.o:
undefined reference to symbol 'pthread_condattr_setclock@@GLIBC_2.3.3' /usr/bin/ld: /usr/lib/libpthread.so.0: error adding symbols: DSO missing from command line collect2: error: ld returned 1 exit status make[2]: ***
[solvers/CompressibleFlowSolver/CMakeFiles/CompressibleFlowSolver.dir/build.make:615:
solvers/CompressibleFlowSolver/CompressibleFlowSolver] Error 1 make[1]: *** [CMakeFiles/Makefile2:2925: solvers/CompressibleFlowSolver/CMakeFiles/CompressibleFlowSolver.dir/all] Error 2 make: *** [Makefile:163: all] Error 2
Can someone point me in the right direction?
Thanks, Mikkel
-- Chris Cantwell Imperial College London South Kensington Campus London SW7 2AZ Email: c.cantwell@imperial.ac.uk www.imperial.ac.uk/people/c.cantwell
Hi Mikkel, Thanks for the output. The -pthread flag is missing from the solvers, but not on the library demos, etc. The attached patch should fix the issue to make the solvers consistent with the library executables. We will include this in the next patch release. Please let me know if you still have problems. Cheers, Chris On Wed, 12 Sep 2018 09:17:59 +0100, Mikkel Bue Lykkegaard <mikkelbue@gmail.com> wrote:
Hi Chris
Thanks for the reply. Sure, here is the output of 'make VERBOSE=1': https://pastebin.com/AtkuDCCi
Also, here are the messages from ccmake (in case it has any relevance): https://pastebin.com/KpKYeDNM
I am running Antergos (Arch) Linux with the following package versions: Linux 4.18.6 gcc 8.2.1 CMake 3.12.1 binutils 2.31.1 BLAS 3.8.0 LAPACK 3.8.0 boost 1.67.0
Mikkel
On Tue, Sep 11, 2018 at 10:37 PM Chris Cantwell <c.cantwell@imperial.ac.uk> wrote:
Hi Mikkel,
Which Linux distribution/version are you using?
Could you send the output of running 'make VERBOSE=1'.
Cheers, Chris
On Tue, 11 Sep 2018 13:05:10 +0100, Mikkel Bue Lykkegaard < mikkelbue@gmail.com> wrote:
Hi all
I am trying to install Nektar++ for the first time, on Linux 4.18.6 with CMake version 3.12.1.
The compilation fails with:
[ 66%] Linking CXX executable CardiacEPSolver /usr/bin/ld: CMakeFiles/CardiacEPSolver.dir/CellModels/CellModel.cpp.o: undefined reference to symbol 'pthread_condattr_setclock@@GLIBC_2.3.3' /usr/bin/ld: /usr/lib/libpthread.so.0: error adding symbols: DSO missing from command line collect2: error: ld returned 1 exit status make[2]: *** [solvers/CardiacEPSolver/CMakeFiles/CardiacEPSolver.dir/build.make:465: solvers/CardiacEPSolver/CardiacEPSolver] Error 1 make[1]: *** [CMakeFiles/Makefile2:2907: solvers/CardiacEPSolver/CMakeFiles/CardiacEPSolver.dir/all] Error 2 make: *** [Makefile:163: all] Error 2
if I disable CardiacEPSolver in the configuration, I get a similar error:
[ 70%] Linking CXX executable CompressibleFlowSolver /usr/bin/ld:
CMakeFiles/CompressibleFlowSolver.dir/ArtificialDiffusion/ArtificialDiffusion.cpp.o:
undefined reference to symbol 'pthread_condattr_setclock@@GLIBC_2.3.3' /usr/bin/ld: /usr/lib/libpthread.so.0: error adding symbols: DSO missing from command line collect2: error: ld returned 1 exit status make[2]: ***
[solvers/CompressibleFlowSolver/CMakeFiles/CompressibleFlowSolver.dir/build.make:615:
solvers/CompressibleFlowSolver/CompressibleFlowSolver] Error 1 make[1]: *** [CMakeFiles/Makefile2:2925: solvers/CompressibleFlowSolver/CMakeFiles/CompressibleFlowSolver.dir/all] Error 2 make: *** [Makefile:163: all] Error 2
Can someone point me in the right direction?
Thanks, Mikkel
-- Chris Cantwell Imperial College London South Kensington Campus London SW7 2AZ Email: c.cantwell@imperial.ac.uk www.imperial.ac.uk/people/c.cantwell
-- Chris Cantwell Imperial College London South Kensington Campus London SW7 2AZ Email: c.cantwell@imperial.ac.uk www.imperial.ac.uk/people/c.cantwell
Hi Chris Thank you. After patching CMakeLists.txt, ccmake now throws the following errors: CMake Error at CMakeLists.txt:10 (GET_TARGET_PROPERTY): GET_TARGET_PROPERTY called with incorrect number of arguments CMake Error at CMakeLists.txt:11 (GET_TARGET_PROPERTY): GET_TARGET_PROPERTY called with incorrect number of arguments Best, Mikkel On Wed, Sep 12, 2018 at 9:49 AM Chris Cantwell <c.cantwell@imperial.ac.uk> wrote:
Hi Mikkel,
Thanks for the output. The -pthread flag is missing from the solvers, but not on the library demos, etc.
The attached patch should fix the issue to make the solvers consistent with the library executables. We will include this in the next patch release.
Please let me know if you still have problems.
Cheers, Chris
On Wed, 12 Sep 2018 09:17:59 +0100, Mikkel Bue Lykkegaard < mikkelbue@gmail.com> wrote:
Hi Chris
Thanks for the reply. Sure, here is the output of 'make VERBOSE=1': https://pastebin.com/AtkuDCCi
Also, here are the messages from ccmake (in case it has any relevance): https://pastebin.com/KpKYeDNM
I am running Antergos (Arch) Linux with the following package versions: Linux 4.18.6 gcc 8.2.1 CMake 3.12.1 binutils 2.31.1 BLAS 3.8.0 LAPACK 3.8.0 boost 1.67.0
Mikkel
On Tue, Sep 11, 2018 at 10:37 PM Chris Cantwell < c.cantwell@imperial.ac.uk> wrote:
Hi Mikkel,
Which Linux distribution/version are you using?
Could you send the output of running 'make VERBOSE=1'.
Cheers, Chris
On Tue, 11 Sep 2018 13:05:10 +0100, Mikkel Bue Lykkegaard < mikkelbue@gmail.com> wrote:
Hi all
I am trying to install Nektar++ for the first time, on Linux 4.18.6 with CMake version 3.12.1.
The compilation fails with:
[ 66%] Linking CXX executable CardiacEPSolver /usr/bin/ld: CMakeFiles/CardiacEPSolver.dir/CellModels/CellModel.cpp.o: undefined reference to symbol 'pthread_condattr_setclock@ @GLIBC_2.3.3' /usr/bin/ld: /usr/lib/libpthread.so.0: error adding symbols: DSO missing from command line collect2: error: ld returned 1 exit status make[2]: ***
[solvers/CardiacEPSolver/CMakeFiles/CardiacEPSolver.dir/build.make:465:
solvers/CardiacEPSolver/CardiacEPSolver] Error 1 make[1]: *** [CMakeFiles/Makefile2:2907: solvers/CardiacEPSolver/CMakeFiles/CardiacEPSolver.dir/all] Error 2 make: *** [Makefile:163: all] Error 2
if I disable CardiacEPSolver in the configuration, I get a similar error:
[ 70%] Linking CXX executable CompressibleFlowSolver /usr/bin/ld:
CMakeFiles/CompressibleFlowSolver.dir/ArtificialDiffusion/ArtificialDiffusion.cpp.o:
undefined reference to symbol 'pthread_condattr_setclock@ @GLIBC_2.3.3' /usr/bin/ld: /usr/lib/libpthread.so.0: error adding symbols: DSO missing from command line collect2: error: ld returned 1 exit status make[2]: ***
[solvers/CompressibleFlowSolver/CMakeFiles/CompressibleFlowSolver.dir/build.make:615:
solvers/CompressibleFlowSolver/CompressibleFlowSolver] Error 1 make[1]: *** [CMakeFiles/Makefile2:2925:
solvers/CompressibleFlowSolver/CMakeFiles/CompressibleFlowSolver.dir/all]
Error 2 make: *** [Makefile:163: all] Error 2
Can someone point me in the right direction?
Thanks, Mikkel
-- Chris Cantwell Imperial College London South Kensington Campus London SW7 2AZ Email: c.cantwell@imperial.ac.uk www.imperial.ac.uk/people/c.cantwell
-- Chris Cantwell Imperial College London South Kensington Campus London SW7 2AZ Email: c.cantwell@imperial.ac.uk www.imperial.ac.uk/people/c.cantwell
Hi Mikkel, That is strange, since it is the same CMake commands as used for the library executables! I am also using CMake 3.12.1, so should not be related to an API change. Please can you confirm: - you are still using an unmodified version of Nektar++ v4.4.1 - you have tried completely wiping the contents of your build directory and configuring from scratch - you are using the default options for the NEKTAR_BUILD_* variables. If you still encounter the error, please could you send me the file sys.txt produced by running cmake --system-information sys.txt and we can see if that sheds any light on the problem. Cheers, Chris On Wed, 12 Sep 2018 10:04:25 +0100, Mikkel Bue Lykkegaard <mikkelbue@gmail.com> wrote:
Hi Chris
Thank you. After patching CMakeLists.txt, ccmake now throws the following errors:
CMake Error at CMakeLists.txt:10 (GET_TARGET_PROPERTY): GET_TARGET_PROPERTY called with incorrect number of arguments
CMake Error at CMakeLists.txt:11 (GET_TARGET_PROPERTY): GET_TARGET_PROPERTY called with incorrect number of arguments
Best, Mikkel
On Wed, Sep 12, 2018 at 9:49 AM Chris Cantwell <c.cantwell@imperial.ac.uk> wrote:
Hi Mikkel,
Thanks for the output. The -pthread flag is missing from the solvers, but not on the library demos, etc.
The attached patch should fix the issue to make the solvers consistent with the library executables. We will include this in the next patch release.
Please let me know if you still have problems.
Cheers, Chris
On Wed, 12 Sep 2018 09:17:59 +0100, Mikkel Bue Lykkegaard < mikkelbue@gmail.com> wrote:
Hi Chris
Thanks for the reply. Sure, here is the output of 'make VERBOSE=1': https://pastebin.com/AtkuDCCi
Also, here are the messages from ccmake (in case it has any relevance): https://pastebin.com/KpKYeDNM
I am running Antergos (Arch) Linux with the following package versions: Linux 4.18.6 gcc 8.2.1 CMake 3.12.1 binutils 2.31.1 BLAS 3.8.0 LAPACK 3.8.0 boost 1.67.0
Mikkel
On Tue, Sep 11, 2018 at 10:37 PM Chris Cantwell < c.cantwell@imperial.ac.uk> wrote:
Hi Mikkel,
Which Linux distribution/version are you using?
Could you send the output of running 'make VERBOSE=1'.
Cheers, Chris
On Tue, 11 Sep 2018 13:05:10 +0100, Mikkel Bue Lykkegaard < mikkelbue@gmail.com> wrote:
Hi all
I am trying to install Nektar++ for the first time, on Linux 4.18.6 with CMake version 3.12.1.
The compilation fails with:
[ 66%] Linking CXX executable CardiacEPSolver /usr/bin/ld: CMakeFiles/CardiacEPSolver.dir/CellModels/CellModel.cpp.o: undefined reference to symbol 'pthread_condattr_setclock@ @GLIBC_2.3.3' /usr/bin/ld: /usr/lib/libpthread.so.0: error adding symbols: DSO missing from command line collect2: error: ld returned 1 exit status make[2]: ***
[solvers/CardiacEPSolver/CMakeFiles/CardiacEPSolver.dir/build.make:465:
solvers/CardiacEPSolver/CardiacEPSolver] Error 1 make[1]: *** [CMakeFiles/Makefile2:2907: solvers/CardiacEPSolver/CMakeFiles/CardiacEPSolver.dir/all] Error 2 make: *** [Makefile:163: all] Error 2
if I disable CardiacEPSolver in the configuration, I get a similar error:
[ 70%] Linking CXX executable CompressibleFlowSolver /usr/bin/ld:
CMakeFiles/CompressibleFlowSolver.dir/ArtificialDiffusion/ArtificialDiffusion.cpp.o:
undefined reference to symbol 'pthread_condattr_setclock@ @GLIBC_2.3.3' /usr/bin/ld: /usr/lib/libpthread.so.0: error adding symbols: DSO missing from command line collect2: error: ld returned 1 exit status make[2]: ***
[solvers/CompressibleFlowSolver/CMakeFiles/CompressibleFlowSolver.dir/build.make:615:
solvers/CompressibleFlowSolver/CompressibleFlowSolver] Error 1 make[1]: *** [CMakeFiles/Makefile2:2925:
solvers/CompressibleFlowSolver/CMakeFiles/CompressibleFlowSolver.dir/all]
Error 2 make: *** [Makefile:163: all] Error 2
Can someone point me in the right direction?
Thanks, Mikkel
-- Chris Cantwell Imperial College London South Kensington Campus London SW7 2AZ Email: c.cantwell@imperial.ac.uk www.imperial.ac.uk/people/c.cantwell
-- Chris Cantwell Imperial College London South Kensington Campus London SW7 2AZ Email: c.cantwell@imperial.ac.uk www.imperial.ac.uk/people/c.cantwell
-- Chris Cantwell Imperial College London South Kensington Campus London SW7 2AZ Email: c.cantwell@imperial.ac.uk www.imperial.ac.uk/people/c.cantwell
Hi Chris I'm sorry, I think I made a mistake with the patch utility. The above output came because the root-level CMakeLists.txt was patched (running 'patch < pthread.patch'), not solvers/CMakeLists.txt. Running 'patch -p1 < pthread.patch', the correct file was patched. My fault. The solvers are now compiled with no errors. However, the compilation now fails similarly when linking utilities/NekMesh: https://pastebin.com/EfsqEQwR Moreover, when compiling in parallel, the same error occurs when linking Tester (which I assume would be compiled after utilities/NekMesh, if not compiling in parallel): https://pastebin.com/qE8mfKcs Cheers, Mikkel On Wed, Sep 12, 2018 at 10:48 AM Chris Cantwell <c.cantwell@imperial.ac.uk> wrote:
Hi Mikkel,
That is strange, since it is the same CMake commands as used for the library executables! I am also using CMake 3.12.1, so should not be related to an API change.
Please can you confirm: - you are still using an unmodified version of Nektar++ v4.4.1 - you have tried completely wiping the contents of your build directory and configuring from scratch - you are using the default options for the NEKTAR_BUILD_* variables.
If you still encounter the error, please could you send me the file sys.txt produced by running cmake --system-information sys.txt and we can see if that sheds any light on the problem.
Cheers, Chris
On Wed, 12 Sep 2018 10:04:25 +0100, Mikkel Bue Lykkegaard < mikkelbue@gmail.com> wrote:
Hi Chris
Thank you. After patching CMakeLists.txt, ccmake now throws the following errors:
CMake Error at CMakeLists.txt:10 (GET_TARGET_PROPERTY): GET_TARGET_PROPERTY called with incorrect number of arguments
CMake Error at CMakeLists.txt:11 (GET_TARGET_PROPERTY): GET_TARGET_PROPERTY called with incorrect number of arguments
Best, Mikkel
On Wed, Sep 12, 2018 at 9:49 AM Chris Cantwell < c.cantwell@imperial.ac.uk> wrote:
Hi Mikkel,
Thanks for the output. The -pthread flag is missing from the solvers, but not on the library demos, etc.
The attached patch should fix the issue to make the solvers consistent with the library executables. We will include this in the next patch release.
Please let me know if you still have problems.
Cheers, Chris
On Wed, 12 Sep 2018 09:17:59 +0100, Mikkel Bue Lykkegaard < mikkelbue@gmail.com> wrote:
Hi Chris
Thanks for the reply. Sure, here is the output of 'make VERBOSE=1': https://pastebin.com/AtkuDCCi
Also, here are the messages from ccmake (in case it has any relevance): https://pastebin.com/KpKYeDNM
I am running Antergos (Arch) Linux with the following package versions: Linux 4.18.6 gcc 8.2.1 CMake 3.12.1 binutils 2.31.1 BLAS 3.8.0 LAPACK 3.8.0 boost 1.67.0
Mikkel
On Tue, Sep 11, 2018 at 10:37 PM Chris Cantwell < c.cantwell@imperial.ac.uk> wrote:
Hi Mikkel,
Which Linux distribution/version are you using?
Could you send the output of running 'make VERBOSE=1'.
Cheers, Chris
On Tue, 11 Sep 2018 13:05:10 +0100, Mikkel Bue Lykkegaard < mikkelbue@gmail.com> wrote:
Hi all
I am trying to install Nektar++ for the first time, on Linux 4.18.6 with CMake version 3.12.1.
The compilation fails with:
[ 66%] Linking CXX executable CardiacEPSolver /usr/bin/ld: CMakeFiles/CardiacEPSolver.dir/CellModels/CellModel.cpp.o: undefined reference to symbol 'pthread_condattr_setclock@ @GLIBC_2.3.3' /usr/bin/ld: /usr/lib/libpthread.so.0: error adding symbols: DSO missing from command line collect2: error: ld returned 1 exit status make[2]: ***
[solvers/CardiacEPSolver/CMakeFiles/CardiacEPSolver.dir/build.make:465:
solvers/CardiacEPSolver/CardiacEPSolver] Error 1 make[1]: *** [CMakeFiles/Makefile2:2907: solvers/CardiacEPSolver/CMakeFiles/CardiacEPSolver.dir/all] Error 2 make: *** [Makefile:163: all] Error 2
if I disable CardiacEPSolver in the configuration, I get a similar error:
[ 70%] Linking CXX executable CompressibleFlowSolver /usr/bin/ld:
CMakeFiles/CompressibleFlowSolver.dir/ArtificialDiffusion/ArtificialDiffusion.cpp.o:
undefined reference to symbol 'pthread_condattr_setclock@ @GLIBC_2.3.3' /usr/bin/ld: /usr/lib/libpthread.so.0: error adding symbols:
DSO
missing
from command line collect2: error: ld returned 1 exit status make[2]: ***
[solvers/CompressibleFlowSolver/CMakeFiles/CompressibleFlowSolver.dir/build.make:615:
solvers/CompressibleFlowSolver/CompressibleFlowSolver] Error 1 make[1]: *** [CMakeFiles/Makefile2:2925:
solvers/CompressibleFlowSolver/CMakeFiles/CompressibleFlowSolver.dir/all]
Error 2 make: *** [Makefile:163: all] Error 2
Can someone point me in the right direction?
Thanks, Mikkel
-- Chris Cantwell Imperial College London South Kensington Campus London SW7 2AZ Email: c.cantwell@imperial.ac.uk www.imperial.ac.uk/people/c.cantwell
-- Chris Cantwell Imperial College London South Kensington Campus London SW7 2AZ Email: c.cantwell@imperial.ac.uk www.imperial.ac.uk/people/c.cantwell
-- Chris Cantwell Imperial College London South Kensington Campus London SW7 2AZ Email: c.cantwell@imperial.ac.uk www.imperial.ac.uk/people/c.cantwell
Hi Mikkel, A similar update to the corresponding CMakeLists.txt files in Utilities and tests should fix those errors. I will prepare a patch for you later tonight. Tester is not dependent on other parts of Nektar++ so can be built straight away when running the build in parallel. Cheers, Chris On 12 September 2018 12:32:51 BST, Mikkel Bue Lykkegaard <mikkelbue@gmail.com> wrote:
Hi Chris
I'm sorry, I think I made a mistake with the patch utility. The above output came because the root-level CMakeLists.txt was patched (running 'patch < pthread.patch'), not solvers/CMakeLists.txt. Running 'patch -p1 < pthread.patch', the correct file was patched. My fault.
The solvers are now compiled with no errors.
However, the compilation now fails similarly when linking utilities/NekMesh: https://pastebin.com/EfsqEQwR
Moreover, when compiling in parallel, the same error occurs when linking Tester (which I assume would be compiled after utilities/NekMesh, if not compiling in parallel): https://pastebin.com/qE8mfKcs
Cheers, Mikkel
On Wed, Sep 12, 2018 at 10:48 AM Chris Cantwell <c.cantwell@imperial.ac.uk> wrote:
Hi Mikkel,
That is strange, since it is the same CMake commands as used for the library executables! I am also using CMake 3.12.1, so should not be related to an API change.
Please can you confirm: - you are still using an unmodified version of Nektar++ v4.4.1 - you have tried completely wiping the contents of your build directory and configuring from scratch - you are using the default options for the NEKTAR_BUILD_* variables.
If you still encounter the error, please could you send me the file sys.txt produced by running cmake --system-information sys.txt and we can see if that sheds any light on the problem.
Cheers, Chris
On Wed, 12 Sep 2018 10:04:25 +0100, Mikkel Bue Lykkegaard < mikkelbue@gmail.com> wrote:
Hi Chris
Thank you. After patching CMakeLists.txt, ccmake now throws the following errors:
CMake Error at CMakeLists.txt:10 (GET_TARGET_PROPERTY): GET_TARGET_PROPERTY called with incorrect number of arguments
CMake Error at CMakeLists.txt:11 (GET_TARGET_PROPERTY): GET_TARGET_PROPERTY called with incorrect number of arguments
Best, Mikkel
On Wed, Sep 12, 2018 at 9:49 AM Chris Cantwell < c.cantwell@imperial.ac.uk> wrote:
Hi Mikkel,
Thanks for the output. The -pthread flag is missing from the solvers, but not on the library demos, etc.
The attached patch should fix the issue to make the solvers consistent with the library executables. We will include this in the next patch release.
Please let me know if you still have problems.
Cheers, Chris
On Wed, 12 Sep 2018 09:17:59 +0100, Mikkel Bue Lykkegaard < mikkelbue@gmail.com> wrote:
Hi Chris
Thanks for the reply. Sure, here is the output of 'make VERBOSE=1': https://pastebin.com/AtkuDCCi
Also, here are the messages from ccmake (in case it has any relevance): https://pastebin.com/KpKYeDNM
I am running Antergos (Arch) Linux with the following package versions: Linux 4.18.6 gcc 8.2.1 CMake 3.12.1 binutils 2.31.1 BLAS 3.8.0 LAPACK 3.8.0 boost 1.67.0
Mikkel
On Tue, Sep 11, 2018 at 10:37 PM Chris Cantwell < c.cantwell@imperial.ac.uk> wrote:
Hi Mikkel,
Which Linux distribution/version are you using?
Could you send the output of running 'make VERBOSE=1'.
Cheers, Chris
On Tue, 11 Sep 2018 13:05:10 +0100, Mikkel Bue Lykkegaard < mikkelbue@gmail.com> wrote: > Hi all > > I am trying to install Nektar++ for the first time, on Linux 4.18.6 with > CMake version 3.12.1. > > The compilation fails with: > > [ 66%] Linking CXX executable CardiacEPSolver > /usr/bin/ld: CMakeFiles/CardiacEPSolver.dir/CellModels/CellModel.cpp.o: > undefined reference to symbol 'pthread_condattr_setclock@ @GLIBC_2.3.3' > /usr/bin/ld: /usr/lib/libpthread.so.0: error adding symbols: DSO missing > from command line > collect2: error: ld returned 1 exit status > make[2]: *** >
[solvers/CardiacEPSolver/CMakeFiles/CardiacEPSolver.dir/build.make:465:
> solvers/CardiacEPSolver/CardiacEPSolver] Error 1 > make[1]: *** [CMakeFiles/Makefile2:2907: > solvers/CardiacEPSolver/CMakeFiles/CardiacEPSolver.dir/all] Error 2 > make: *** [Makefile:163: all] Error 2 > > if I disable CardiacEPSolver in the configuration, I get a similar error: > > [ 70%] Linking CXX executable CompressibleFlowSolver > /usr/bin/ld: >
CMakeFiles/CompressibleFlowSolver.dir/ArtificialDiffusion/ArtificialDiffusion.cpp.o:
> undefined reference to symbol 'pthread_condattr_setclock@ @GLIBC_2.3.3' > /usr/bin/ld: /usr/lib/libpthread.so.0: error adding
symbols: DSO
missing
> from command line > collect2: error: ld returned 1 exit status > make[2]: *** >
[solvers/CompressibleFlowSolver/CMakeFiles/CompressibleFlowSolver.dir/build.make:615:
> solvers/CompressibleFlowSolver/CompressibleFlowSolver]
Error 1
> make[1]: *** [CMakeFiles/Makefile2:2925: >
solvers/CompressibleFlowSolver/CMakeFiles/CompressibleFlowSolver.dir/all]
> Error 2 > make: *** [Makefile:163: all] Error 2 > > Can someone point me in the right direction? > > Thanks, Mikkel
-- Chris Cantwell Imperial College London South Kensington Campus London SW7 2AZ Email: c.cantwell@imperial.ac.uk www.imperial.ac.uk/people/c.cantwell
-- Chris Cantwell Imperial College London South Kensington Campus London SW7 2AZ Email: c.cantwell@imperial.ac.uk www.imperial.ac.uk/people/c.cantwell
-- Chris Cantwell Imperial College London South Kensington Campus London SW7 2AZ Email: c.cantwell@imperial.ac.uk www.imperial.ac.uk/people/c.cantwell
-- Chris Cantwell Imperial College London South Kensington Campus London SW7 2AZ Email: c.cantwell@imperial.ac.uk www.imperial.ac.uk/people/c.cantwell
Hi Mikkel, Please try the attached patch. Cheers, Chris On Wed, 12 Sep 2018 17:48:39 +0100, Chris Cantwell <c.cantwell@imperial.ac.uk> wrote:
Hi Mikkel,
A similar update to the corresponding CMakeLists.txt files in Utilities and tests should fix those errors. I will prepare a patch for you later tonight.
Tester is not dependent on other parts of Nektar++ so can be built straight away when running the build in parallel.
Cheers, Chris
On 12 September 2018 12:32:51 BST, Mikkel Bue Lykkegaard <mikkelbue@gmail.com> wrote:
Hi Chris
I'm sorry, I think I made a mistake with the patch utility. The above output came because the root-level CMakeLists.txt was patched (running 'patch < pthread.patch'), not solvers/CMakeLists.txt. Running 'patch -p1 < pthread.patch', the correct file was patched. My fault.
The solvers are now compiled with no errors.
However, the compilation now fails similarly when linking utilities/NekMesh: https://pastebin.com/EfsqEQwR
Moreover, when compiling in parallel, the same error occurs when linking Tester (which I assume would be compiled after utilities/NekMesh, if not compiling in parallel): https://pastebin.com/qE8mfKcs
Cheers, Mikkel
On Wed, Sep 12, 2018 at 10:48 AM Chris Cantwell <c.cantwell@imperial.ac.uk> wrote:
Hi Mikkel,
That is strange, since it is the same CMake commands as used for the library executables! I am also using CMake 3.12.1, so should not be related to an API change.
Please can you confirm: - you are still using an unmodified version of Nektar++ v4.4.1 - you have tried completely wiping the contents of your build directory and configuring from scratch - you are using the default options for the NEKTAR_BUILD_* variables.
If you still encounter the error, please could you send me the file sys.txt produced by running cmake --system-information sys.txt and we can see if that sheds any light on the problem.
Cheers, Chris
On Wed, 12 Sep 2018 10:04:25 +0100, Mikkel Bue Lykkegaard < mikkelbue@gmail.com> wrote:
Hi Chris
Thank you. After patching CMakeLists.txt, ccmake now throws the following errors:
CMake Error at CMakeLists.txt:10 (GET_TARGET_PROPERTY): GET_TARGET_PROPERTY called with incorrect number of arguments
CMake Error at CMakeLists.txt:11 (GET_TARGET_PROPERTY): GET_TARGET_PROPERTY called with incorrect number of arguments
Best, Mikkel
On Wed, Sep 12, 2018 at 9:49 AM Chris Cantwell < c.cantwell@imperial.ac.uk> wrote:
Hi Mikkel,
Thanks for the output. The -pthread flag is missing from the solvers, but not on the library demos, etc.
The attached patch should fix the issue to make the solvers consistent with the library executables. We will include this in the next patch release.
Please let me know if you still have problems.
Cheers, Chris
On Wed, 12 Sep 2018 09:17:59 +0100, Mikkel Bue Lykkegaard < mikkelbue@gmail.com> wrote:
Hi Chris
Thanks for the reply. Sure, here is the output of 'make VERBOSE=1': https://pastebin.com/AtkuDCCi
Also, here are the messages from ccmake (in case it has any relevance): https://pastebin.com/KpKYeDNM
I am running Antergos (Arch) Linux with the following package versions: Linux 4.18.6 gcc 8.2.1 CMake 3.12.1 binutils 2.31.1 BLAS 3.8.0 LAPACK 3.8.0 boost 1.67.0
Mikkel
On Tue, Sep 11, 2018 at 10:37 PM Chris Cantwell < c.cantwell@imperial.ac.uk> wrote:
> Hi Mikkel, > > Which Linux distribution/version are you using? > > Could you send the output of running 'make VERBOSE=1'. > > Cheers, > Chris > > > On Tue, 11 Sep 2018 13:05:10 +0100, Mikkel Bue Lykkegaard < > mikkelbue@gmail.com> wrote: > > Hi all > > > > I am trying to install Nektar++ for the first time, on Linux 4.18.6 with > > CMake version 3.12.1. > > > > The compilation fails with: > > > > [ 66%] Linking CXX executable CardiacEPSolver > > /usr/bin/ld: CMakeFiles/CardiacEPSolver.dir/CellModels/CellModel.cpp.o: > > undefined reference to symbol 'pthread_condattr_setclock@ @GLIBC_2.3.3' > > /usr/bin/ld: /usr/lib/libpthread.so.0: error adding symbols: DSO missing > > from command line > > collect2: error: ld returned 1 exit status > > make[2]: *** > >
[solvers/CardiacEPSolver/CMakeFiles/CardiacEPSolver.dir/build.make:465:
> > solvers/CardiacEPSolver/CardiacEPSolver] Error 1 > > make[1]: *** [CMakeFiles/Makefile2:2907: > > solvers/CardiacEPSolver/CMakeFiles/CardiacEPSolver.dir/all] Error 2 > > make: *** [Makefile:163: all] Error 2 > > > > if I disable CardiacEPSolver in the configuration, I get a similar error: > > > > [ 70%] Linking CXX executable CompressibleFlowSolver > > /usr/bin/ld: > > >
CMakeFiles/CompressibleFlowSolver.dir/ArtificialDiffusion/ArtificialDiffusion.cpp.o:
> > undefined reference to symbol 'pthread_condattr_setclock@ @GLIBC_2.3.3' > > /usr/bin/ld: /usr/lib/libpthread.so.0: error adding
symbols: DSO
missing
> > from command line > > collect2: error: ld returned 1 exit status > > make[2]: *** > > >
[solvers/CompressibleFlowSolver/CMakeFiles/CompressibleFlowSolver.dir/build.make:615:
> > solvers/CompressibleFlowSolver/CompressibleFlowSolver]
Error 1
> > make[1]: *** [CMakeFiles/Makefile2:2925: > >
solvers/CompressibleFlowSolver/CMakeFiles/CompressibleFlowSolver.dir/all]
> > Error 2 > > make: *** [Makefile:163: all] Error 2 > > > > Can someone point me in the right direction? > > > > Thanks, Mikkel > > > -- > Chris Cantwell > Imperial College London > South Kensington Campus > London SW7 2AZ > Email: c.cantwell@imperial.ac.uk > www.imperial.ac.uk/people/c.cantwell >
-- Chris Cantwell Imperial College London South Kensington Campus London SW7 2AZ Email: c.cantwell@imperial.ac.uk www.imperial.ac.uk/people/c.cantwell
-- Chris Cantwell Imperial College London South Kensington Campus London SW7 2AZ Email: c.cantwell@imperial.ac.uk www.imperial.ac.uk/people/c.cantwell
-- Chris Cantwell Imperial College London South Kensington Campus London SW7 2AZ Email: c.cantwell@imperial.ac.uk www.imperial.ac.uk/people/c.cantwell
-- Chris Cantwell Imperial College London South Kensington Campus London SW7 2AZ Email: c.cantwell@imperial.ac.uk www.imperial.ac.uk/people/c.cantwell
Hi Chris Thank you for this. Tester is now linked succesfully. However, the patch for utilities doesn't seem to work. Linking NekMesh still fails the same way as before: https://pastebin.com/zAn6cSer Please find attached the patched utilities/CMakeLists.txt to confirm that it has been patched as intended. Cheers, MIkkel On Wed, Sep 12, 2018 at 10:56 PM Chris Cantwell <c.cantwell@imperial.ac.uk> wrote:
Hi Mikkel,
Please try the attached patch.
Cheers, Chris
On Wed, 12 Sep 2018 17:48:39 +0100, Chris Cantwell < c.cantwell@imperial.ac.uk> wrote:
Hi Mikkel,
A similar update to the corresponding CMakeLists.txt files in Utilities and tests should fix those errors. I will prepare a patch for you later tonight.
Tester is not dependent on other parts of Nektar++ so can be built straight away when running the build in parallel.
Cheers, Chris
On 12 September 2018 12:32:51 BST, Mikkel Bue Lykkegaard < mikkelbue@gmail.com> wrote:
Hi Chris
I'm sorry, I think I made a mistake with the patch utility. The above output came because the root-level CMakeLists.txt was patched (running 'patch < pthread.patch'), not solvers/CMakeLists.txt. Running 'patch -p1 < pthread.patch', the correct file was patched. My fault.
The solvers are now compiled with no errors.
However, the compilation now fails similarly when linking utilities/NekMesh: https://pastebin.com/EfsqEQwR
Moreover, when compiling in parallel, the same error occurs when linking Tester (which I assume would be compiled after utilities/NekMesh, if not compiling in parallel): https://pastebin.com/qE8mfKcs
Cheers, Mikkel
On Wed, Sep 12, 2018 at 10:48 AM Chris Cantwell <c.cantwell@imperial.ac.uk> wrote:
Hi Mikkel,
That is strange, since it is the same CMake commands as used for the library executables! I am also using CMake 3.12.1, so should not be related to an API change.
Please can you confirm: - you are still using an unmodified version of Nektar++ v4.4.1 - you have tried completely wiping the contents of your build directory and configuring from scratch - you are using the default options for the NEKTAR_BUILD_* variables.
If you still encounter the error, please could you send me the file sys.txt produced by running cmake --system-information sys.txt and we can see if that sheds any light on the problem.
Cheers, Chris
On Wed, 12 Sep 2018 10:04:25 +0100, Mikkel Bue Lykkegaard < mikkelbue@gmail.com> wrote:
Hi Chris
Thank you. After patching CMakeLists.txt, ccmake now throws the following errors:
CMake Error at CMakeLists.txt:10 (GET_TARGET_PROPERTY): GET_TARGET_PROPERTY called with incorrect number of arguments
CMake Error at CMakeLists.txt:11 (GET_TARGET_PROPERTY): GET_TARGET_PROPERTY called with incorrect number of arguments
Best, Mikkel
On Wed, Sep 12, 2018 at 9:49 AM Chris Cantwell < c.cantwell@imperial.ac.uk> wrote:
Hi Mikkel,
Thanks for the output. The -pthread flag is missing from the solvers, but not on the library demos, etc.
The attached patch should fix the issue to make the solvers consistent with the library executables. We will include this in the next patch release.
Please let me know if you still have problems.
Cheers, Chris
On Wed, 12 Sep 2018 09:17:59 +0100, Mikkel Bue Lykkegaard < mikkelbue@gmail.com> wrote: > Hi Chris > > Thanks for the reply. Sure, here is the output of 'make VERBOSE=1': > https://pastebin.com/AtkuDCCi > > Also, here are the messages from ccmake (in case it has any relevance): > https://pastebin.com/KpKYeDNM > > I am running Antergos (Arch) Linux with the following package versions: > Linux 4.18.6 > gcc 8.2.1 > CMake 3.12.1 > binutils 2.31.1 > BLAS 3.8.0 > LAPACK 3.8.0 > boost 1.67.0 > > Mikkel > > On Tue, Sep 11, 2018 at 10:37 PM Chris Cantwell < c.cantwell@imperial.ac.uk> > wrote: > > > Hi Mikkel, > > > > Which Linux distribution/version are you using? > > > > Could you send the output of running 'make VERBOSE=1'. > > > > Cheers, > > Chris > > > > > > On Tue, 11 Sep 2018 13:05:10 +0100, Mikkel Bue Lykkegaard < > > mikkelbue@gmail.com> wrote: > > > Hi all > > > > > > I am trying to install Nektar++ for the first time, on Linux 4.18.6 with > > > CMake version 3.12.1. > > > > > > The compilation fails with: > > > > > > [ 66%] Linking CXX executable CardiacEPSolver > > > /usr/bin/ld: CMakeFiles/CardiacEPSolver.dir/CellModels/CellModel.cpp.o: > > > undefined reference to symbol 'pthread_condattr_setclock@ @GLIBC_2.3.3' > > > /usr/bin/ld: /usr/lib/libpthread.so.0: error adding symbols: DSO missing > > > from command line > > > collect2: error: ld returned 1 exit status > > > make[2]: *** > > >
[solvers/CardiacEPSolver/CMakeFiles/CardiacEPSolver.dir/build.make:465:
> > > solvers/CardiacEPSolver/CardiacEPSolver] Error 1 > > > make[1]: *** [CMakeFiles/Makefile2:2907: > > > solvers/CardiacEPSolver/CMakeFiles/CardiacEPSolver.dir/all] Error 2 > > > make: *** [Makefile:163: all] Error 2 > > > > > > if I disable CardiacEPSolver in the configuration, I get a similar error: > > > > > > [ 70%] Linking CXX executable CompressibleFlowSolver > > > /usr/bin/ld: > > > > >
CMakeFiles/CompressibleFlowSolver.dir/ArtificialDiffusion/ArtificialDiffusion.cpp.o:
> > > undefined reference to symbol 'pthread_condattr_setclock@ @GLIBC_2.3.3' > > > /usr/bin/ld: /usr/lib/libpthread.so.0: error adding
symbols: DSO
missing > > > from command line > > > collect2: error: ld returned 1 exit status > > > make[2]: *** > > > > >
[solvers/CompressibleFlowSolver/CMakeFiles/CompressibleFlowSolver.dir/build.make:615:
> > > solvers/CompressibleFlowSolver/CompressibleFlowSolver]
Error 1
> > > make[1]: *** [CMakeFiles/Makefile2:2925: > > >
solvers/CompressibleFlowSolver/CMakeFiles/CompressibleFlowSolver.dir/all]
> > > Error 2 > > > make: *** [Makefile:163: all] Error 2 > > > > > > Can someone point me in the right direction? > > > > > > Thanks, Mikkel > > > > > > -- > > Chris Cantwell > > Imperial College London > > South Kensington Campus > > London SW7 2AZ > > Email: c.cantwell@imperial.ac.uk > > www.imperial.ac.uk/people/c.cantwell > >
-- Chris Cantwell Imperial College London South Kensington Campus London SW7 2AZ Email: c.cantwell@imperial.ac.uk www.imperial.ac.uk/people/c.cantwell
-- Chris Cantwell Imperial College London South Kensington Campus London SW7 2AZ Email: c.cantwell@imperial.ac.uk www.imperial.ac.uk/people/c.cantwell
-- Chris Cantwell Imperial College London South Kensington Campus London SW7 2AZ Email: c.cantwell@imperial.ac.uk www.imperial.ac.uk/people/c.cantwell
-- Chris Cantwell Imperial College London South Kensington Campus London SW7 2AZ Email: c.cantwell@imperial.ac.uk www.imperial.ac.uk/people/c.cantwell
Hi Mikkel, Sorry, as I cannot actually reproduce your problem, I cannot properly test the patches. Please include the error message regarding Utilities. Cheers, Chris On Thu, 13 Sep 2018 09:07:33 +0100, Mikkel Bue Lykkegaard <mikkelbue@gmail.com> wrote:
Hi Chris
Thank you for this. Tester is now linked succesfully. However, the patch for utilities doesn't seem to work. Linking NekMesh still fails the same way as before: https://pastebin.com/zAn6cSer
Please find attached the patched utilities/CMakeLists.txt to confirm that it has been patched as intended.
Cheers, MIkkel
On Wed, Sep 12, 2018 at 10:56 PM Chris Cantwell <c.cantwell@imperial.ac.uk> wrote:
Hi Mikkel,
Please try the attached patch.
Cheers, Chris
On Wed, 12 Sep 2018 17:48:39 +0100, Chris Cantwell < c.cantwell@imperial.ac.uk> wrote:
Hi Mikkel,
A similar update to the corresponding CMakeLists.txt files in Utilities and tests should fix those errors. I will prepare a patch for you later tonight.
Tester is not dependent on other parts of Nektar++ so can be built straight away when running the build in parallel.
Cheers, Chris
On 12 September 2018 12:32:51 BST, Mikkel Bue Lykkegaard < mikkelbue@gmail.com> wrote:
Hi Chris
I'm sorry, I think I made a mistake with the patch utility. The above output came because the root-level CMakeLists.txt was patched (running 'patch < pthread.patch'), not solvers/CMakeLists.txt. Running 'patch -p1 < pthread.patch', the correct file was patched. My fault.
The solvers are now compiled with no errors.
However, the compilation now fails similarly when linking utilities/NekMesh: https://pastebin.com/EfsqEQwR
Moreover, when compiling in parallel, the same error occurs when linking Tester (which I assume would be compiled after utilities/NekMesh, if not compiling in parallel): https://pastebin.com/qE8mfKcs
Cheers, Mikkel
On Wed, Sep 12, 2018 at 10:48 AM Chris Cantwell <c.cantwell@imperial.ac.uk> wrote:
Hi Mikkel,
That is strange, since it is the same CMake commands as used for the library executables! I am also using CMake 3.12.1, so should not be related to an API change.
Please can you confirm: - you are still using an unmodified version of Nektar++ v4.4.1 - you have tried completely wiping the contents of your build directory and configuring from scratch - you are using the default options for the NEKTAR_BUILD_* variables.
If you still encounter the error, please could you send me the file sys.txt produced by running cmake --system-information sys.txt and we can see if that sheds any light on the problem.
Cheers, Chris
On Wed, 12 Sep 2018 10:04:25 +0100, Mikkel Bue Lykkegaard < mikkelbue@gmail.com> wrote:
Hi Chris
Thank you. After patching CMakeLists.txt, ccmake now throws the following errors:
CMake Error at CMakeLists.txt:10 (GET_TARGET_PROPERTY): GET_TARGET_PROPERTY called with incorrect number of arguments
CMake Error at CMakeLists.txt:11 (GET_TARGET_PROPERTY): GET_TARGET_PROPERTY called with incorrect number of arguments
Best, Mikkel
On Wed, Sep 12, 2018 at 9:49 AM Chris Cantwell < c.cantwell@imperial.ac.uk> wrote:
> Hi Mikkel, > > Thanks for the output. The -pthread flag is missing from the solvers, but > not on the library demos, etc. > > The attached patch should fix the issue to make the solvers consistent > with the library executables. We will include this in the next patch > release. > > Please let me know if you still have problems. > > Cheers, > Chris > > > On Wed, 12 Sep 2018 09:17:59 +0100, Mikkel Bue Lykkegaard < > mikkelbue@gmail.com> wrote: > > Hi Chris > > > > Thanks for the reply. Sure, here is the output of 'make VERBOSE=1': > > https://pastebin.com/AtkuDCCi > > > > Also, here are the messages from ccmake (in case it has any relevance): > > https://pastebin.com/KpKYeDNM > > > > I am running Antergos (Arch) Linux with the following package versions: > > Linux 4.18.6 > > gcc 8.2.1 > > CMake 3.12.1 > > binutils 2.31.1 > > BLAS 3.8.0 > > LAPACK 3.8.0 > > boost 1.67.0 > > > > Mikkel > > > > On Tue, Sep 11, 2018 at 10:37 PM Chris Cantwell < > c.cantwell@imperial.ac.uk> > > wrote: > > > > > Hi Mikkel, > > > > > > Which Linux distribution/version are you using? > > > > > > Could you send the output of running 'make VERBOSE=1'. > > > > > > Cheers, > > > Chris > > > > > > > > > On Tue, 11 Sep 2018 13:05:10 +0100, Mikkel Bue Lykkegaard < > > > mikkelbue@gmail.com> wrote: > > > > Hi all > > > > > > > > I am trying to install Nektar++ for the first time, on Linux 4.18.6 > with > > > > CMake version 3.12.1. > > > > > > > > The compilation fails with: > > > > > > > > [ 66%] Linking CXX executable CardiacEPSolver > > > > /usr/bin/ld: > CMakeFiles/CardiacEPSolver.dir/CellModels/CellModel.cpp.o: > > > > undefined reference to symbol 'pthread_condattr_setclock@ > @GLIBC_2.3.3' > > > > /usr/bin/ld: /usr/lib/libpthread.so.0: error adding symbols: DSO > missing > > > > from command line > > > > collect2: error: ld returned 1 exit status > > > > make[2]: *** > > > > >
[solvers/CardiacEPSolver/CMakeFiles/CardiacEPSolver.dir/build.make:465:
> > > > solvers/CardiacEPSolver/CardiacEPSolver] Error 1 > > > > make[1]: *** [CMakeFiles/Makefile2:2907: > > > > solvers/CardiacEPSolver/CMakeFiles/CardiacEPSolver.dir/all] Error 2 > > > > make: *** [Makefile:163: all] Error 2 > > > > > > > > if I disable CardiacEPSolver in the configuration, I get a similar > error: > > > > > > > > [ 70%] Linking CXX executable CompressibleFlowSolver > > > > /usr/bin/ld: > > > > > > > >
CMakeFiles/CompressibleFlowSolver.dir/ArtificialDiffusion/ArtificialDiffusion.cpp.o:
> > > > > undefined reference to symbol 'pthread_condattr_setclock@ > @GLIBC_2.3.3' > > > > /usr/bin/ld: /usr/lib/libpthread.so.0: error adding symbols: DSO > missing > > > > from command line > > > > collect2: error: ld returned 1 exit status > > > > make[2]: *** > > > > > > > >
[solvers/CompressibleFlowSolver/CMakeFiles/CompressibleFlowSolver.dir/build.make:615:
> > > > > solvers/CompressibleFlowSolver/CompressibleFlowSolver] Error 1 > > > > make[1]: *** [CMakeFiles/Makefile2:2925: > > > > >
solvers/CompressibleFlowSolver/CMakeFiles/CompressibleFlowSolver.dir/all]
> > > > Error 2 > > > > make: *** [Makefile:163: all] Error 2 > > > > > > > > Can someone point me in the right direction? > > > > > > > > Thanks, Mikkel > > > > > > > > > -- > > > Chris Cantwell > > > Imperial College London > > > South Kensington Campus > > > London SW7 2AZ > > > Email: c.cantwell@imperial.ac.uk > > > www.imperial.ac.uk/people/c.cantwell > > > > > > -- > Chris Cantwell > Imperial College London > South Kensington Campus > London SW7 2AZ > Email: c.cantwell@imperial.ac.uk > www.imperial.ac.uk/people/c.cantwell >
-- Chris Cantwell Imperial College London South Kensington Campus London SW7 2AZ Email: c.cantwell@imperial.ac.uk www.imperial.ac.uk/people/c.cantwell
-- Chris Cantwell Imperial College London South Kensington Campus London SW7 2AZ Email: c.cantwell@imperial.ac.uk www.imperial.ac.uk/people/c.cantwell
-- Chris Cantwell Imperial College London South Kensington Campus London SW7 2AZ Email: c.cantwell@imperial.ac.uk www.imperial.ac.uk/people/c.cantwell
-- Chris Cantwell Imperial College London South Kensington Campus London SW7 2AZ Email: c.cantwell@imperial.ac.uk www.imperial.ac.uk/people/c.cantwell
Hi Chris I'm sorry, I thought I had linked to the error. Here it is: https://pastebin.com/zAn6cSer In the meantime, however, I found a fix :) ... NekMesh has its own CMakeList.txt (i.e. utilities/NekMesh/CMakeList.txt), which doesn't make use of the macro defined in the upper level /utilities/CMakeList.txt .. So I removed the changes from /utilities/CMakeList.txt and modified utilities/NekMesh/CMakeList.txt instead, and now it compiles without errors. Please find attached the modified utilities/NekMesh/CMakeList.txt, my changes are in lines 97-116. Thank you so much for your help. Best, Mikkel On Thu, Sep 13, 2018 at 10:05 AM Chris Cantwell <c.cantwell@imperial.ac.uk> wrote:
Hi Mikkel,
Sorry, as I cannot actually reproduce your problem, I cannot properly test the patches.
Please include the error message regarding Utilities.
Cheers, Chris
On Thu, 13 Sep 2018 09:07:33 +0100, Mikkel Bue Lykkegaard < mikkelbue@gmail.com> wrote:
Hi Chris
Thank you for this. Tester is now linked succesfully. However, the patch for utilities doesn't seem to work. Linking NekMesh still fails the same way as before: https://pastebin.com/zAn6cSer
Please find attached the patched utilities/CMakeLists.txt to confirm that it has been patched as intended.
Cheers, MIkkel
On Wed, Sep 12, 2018 at 10:56 PM Chris Cantwell < c.cantwell@imperial.ac.uk> wrote:
Hi Mikkel,
Please try the attached patch.
Cheers, Chris
On Wed, 12 Sep 2018 17:48:39 +0100, Chris Cantwell < c.cantwell@imperial.ac.uk> wrote:
Hi Mikkel,
A similar update to the corresponding CMakeLists.txt files in Utilities and tests should fix those errors. I will prepare a patch for you later tonight.
Tester is not dependent on other parts of Nektar++ so can be built straight away when running the build in parallel.
Cheers, Chris
On 12 September 2018 12:32:51 BST, Mikkel Bue Lykkegaard < mikkelbue@gmail.com> wrote:
Hi Chris
I'm sorry, I think I made a mistake with the patch utility. The above output came because the root-level CMakeLists.txt was patched (running 'patch < pthread.patch'), not solvers/CMakeLists.txt. Running 'patch -p1 < pthread.patch', the correct file was patched. My fault.
The solvers are now compiled with no errors.
However, the compilation now fails similarly when linking utilities/NekMesh: https://pastebin.com/EfsqEQwR
Moreover, when compiling in parallel, the same error occurs when linking Tester (which I assume would be compiled after utilities/NekMesh, if not compiling in parallel): https://pastebin.com/qE8mfKcs
Cheers, Mikkel
On Wed, Sep 12, 2018 at 10:48 AM Chris Cantwell <c.cantwell@imperial.ac.uk> wrote:
Hi Mikkel,
That is strange, since it is the same CMake commands as used for the library executables! I am also using CMake 3.12.1, so should not be related to an API change.
Please can you confirm: - you are still using an unmodified version of Nektar++ v4.4.1 - you have tried completely wiping the contents of your build directory and configuring from scratch - you are using the default options for the NEKTAR_BUILD_* variables.
If you still encounter the error, please could you send me the file sys.txt produced by running cmake --system-information sys.txt and we can see if that sheds any light on the problem.
Cheers, Chris
On Wed, 12 Sep 2018 10:04:25 +0100, Mikkel Bue Lykkegaard < mikkelbue@gmail.com> wrote: > Hi Chris > > Thank you. After patching CMakeLists.txt, ccmake now throws the following > errors: > > CMake Error at CMakeLists.txt:10 (GET_TARGET_PROPERTY): > GET_TARGET_PROPERTY called with incorrect number of arguments > > CMake Error at CMakeLists.txt:11 (GET_TARGET_PROPERTY): > GET_TARGET_PROPERTY called with incorrect number of arguments > > Best, Mikkel > > On Wed, Sep 12, 2018 at 9:49 AM Chris Cantwell < c.cantwell@imperial.ac.uk> > wrote: > > > Hi Mikkel, > > > > Thanks for the output. The -pthread flag is missing from the solvers, but > > not on the library demos, etc. > > > > The attached patch should fix the issue to make the solvers consistent > > with the library executables. We will include this in the next patch > > release. > > > > Please let me know if you still have problems. > > > > Cheers, > > Chris > > > > > > On Wed, 12 Sep 2018 09:17:59 +0100, Mikkel Bue Lykkegaard < > > mikkelbue@gmail.com> wrote: > > > Hi Chris > > > > > > Thanks for the reply. Sure, here is the output of 'make VERBOSE=1': > > > https://pastebin.com/AtkuDCCi > > > > > > Also, here are the messages from ccmake (in case it has any relevance): > > > https://pastebin.com/KpKYeDNM > > > > > > I am running Antergos (Arch) Linux with the following package versions: > > > Linux 4.18.6 > > > gcc 8.2.1 > > > CMake 3.12.1 > > > binutils 2.31.1 > > > BLAS 3.8.0 > > > LAPACK 3.8.0 > > > boost 1.67.0 > > > > > > Mikkel > > > > > > On Tue, Sep 11, 2018 at 10:37 PM Chris Cantwell < > > c.cantwell@imperial.ac.uk> > > > wrote: > > > > > > > Hi Mikkel, > > > > > > > > Which Linux distribution/version are you using? > > > > > > > > Could you send the output of running 'make VERBOSE=1'. > > > > > > > > Cheers, > > > > Chris > > > > > > > > > > > > On Tue, 11 Sep 2018 13:05:10 +0100, Mikkel Bue Lykkegaard < > > > > mikkelbue@gmail.com> wrote: > > > > > Hi all > > > > > > > > > > I am trying to install Nektar++ for the first time, on Linux 4.18.6 > > with > > > > > CMake version 3.12.1. > > > > > > > > > > The compilation fails with: > > > > > > > > > > [ 66%] Linking CXX executable CardiacEPSolver > > > > > /usr/bin/ld: > > CMakeFiles/CardiacEPSolver.dir/CellModels/CellModel.cpp.o: > > > > > undefined reference to symbol 'pthread_condattr_setclock@ > > @GLIBC_2.3.3' > > > > > /usr/bin/ld: /usr/lib/libpthread.so.0: error adding symbols: DSO > > missing > > > > > from command line > > > > > collect2: error: ld returned 1 exit status > > > > > make[2]: *** > > > > > > >
[solvers/CardiacEPSolver/CMakeFiles/CardiacEPSolver.dir/build.make:465:
> > > > > solvers/CardiacEPSolver/CardiacEPSolver] Error 1 > > > > > make[1]: *** [CMakeFiles/Makefile2:2907: > > > > > solvers/CardiacEPSolver/CMakeFiles/CardiacEPSolver.dir/all] Error 2 > > > > > make: *** [Makefile:163: all] Error 2 > > > > > > > > > > if I disable CardiacEPSolver in the configuration, I get a similar > > error: > > > > > > > > > > [ 70%] Linking CXX executable CompressibleFlowSolver > > > > > /usr/bin/ld: > > > > > > > > > > >
CMakeFiles/CompressibleFlowSolver.dir/ArtificialDiffusion/ArtificialDiffusion.cpp.o:
> > > > > > > undefined reference to symbol
'pthread_condattr_setclock@
> > @GLIBC_2.3.3' > > > > > /usr/bin/ld: /usr/lib/libpthread.so.0: error adding symbols: DSO > > missing > > > > > from command line > > > > > collect2: error: ld returned 1 exit status > > > > > make[2]: *** > > > > > > > > > > >
[solvers/CompressibleFlowSolver/CMakeFiles/CompressibleFlowSolver.dir/build.make:615:
> > > > > > > solvers/CompressibleFlowSolver/CompressibleFlowSolver] Error 1 > > > > > make[1]: *** [CMakeFiles/Makefile2:2925: > > > > > > >
solvers/CompressibleFlowSolver/CMakeFiles/CompressibleFlowSolver.dir/all]
> > > > > Error 2 > > > > > make: *** [Makefile:163: all] Error 2 > > > > > > > > > > Can someone point me in the right direction? > > > > > > > > > > Thanks, Mikkel > > > > > > > > > > > > -- > > > > Chris Cantwell > > > > Imperial College London > > > > South Kensington Campus > > > > London SW7 2AZ > > > > Email: c.cantwell@imperial.ac.uk > > > > www.imperial.ac.uk/people/c.cantwell > > > > > > > > > > -- > > Chris Cantwell > > Imperial College London > > South Kensington Campus > > London SW7 2AZ > > Email: c.cantwell@imperial.ac.uk > > www.imperial.ac.uk/people/c.cantwell > >
-- Chris Cantwell Imperial College London South Kensington Campus London SW7 2AZ Email: c.cantwell@imperial.ac.uk www.imperial.ac.uk/people/c.cantwell
-- Chris Cantwell Imperial College London South Kensington Campus London SW7 2AZ Email: c.cantwell@imperial.ac.uk www.imperial.ac.uk/people/c.cantwell
-- Chris Cantwell Imperial College London South Kensington Campus London SW7 2AZ Email: c.cantwell@imperial.ac.uk www.imperial.ac.uk/people/c.cantwell
-- Chris Cantwell Imperial College London South Kensington Campus London SW7 2AZ Email: c.cantwell@imperial.ac.uk www.imperial.ac.uk/people/c.cantwell
Hi Mikkel, Ah, I see, the utilities error was the NekMesh error. Glad you have it fixed. I will get these changes merged into the next patch release although, due to other changes in the set up of the build system, they should not be needed for v4.5 onwards. Cheers, Chris On Thu, 13 Sep 2018 11:32:23 +0100, Mikkel Bue Lykkegaard <mikkelbue@gmail.com> wrote:
Hi Chris
I'm sorry, I thought I had linked to the error. Here it is: https://pastebin.com/zAn6cSer
In the meantime, however, I found a fix :) ... NekMesh has its own CMakeList.txt (i.e. utilities/NekMesh/CMakeList.txt), which doesn't make use of the macro defined in the upper level /utilities/CMakeList.txt .. So I removed the changes from /utilities/CMakeList.txt and modified utilities/NekMesh/CMakeList.txt instead, and now it compiles without errors.
Please find attached the modified utilities/NekMesh/CMakeList.txt, my changes are in lines 97-116.
Thank you so much for your help.
Best, Mikkel
On Thu, Sep 13, 2018 at 10:05 AM Chris Cantwell <c.cantwell@imperial.ac.uk> wrote:
Hi Mikkel,
Sorry, as I cannot actually reproduce your problem, I cannot properly test the patches.
Please include the error message regarding Utilities.
Cheers, Chris
On Thu, 13 Sep 2018 09:07:33 +0100, Mikkel Bue Lykkegaard < mikkelbue@gmail.com> wrote:
Hi Chris
Thank you for this. Tester is now linked succesfully. However, the patch for utilities doesn't seem to work. Linking NekMesh still fails the same way as before: https://pastebin.com/zAn6cSer
Please find attached the patched utilities/CMakeLists.txt to confirm that it has been patched as intended.
Cheers, MIkkel
On Wed, Sep 12, 2018 at 10:56 PM Chris Cantwell < c.cantwell@imperial.ac.uk> wrote:
Hi Mikkel,
Please try the attached patch.
Cheers, Chris
On Wed, 12 Sep 2018 17:48:39 +0100, Chris Cantwell < c.cantwell@imperial.ac.uk> wrote:
Hi Mikkel,
A similar update to the corresponding CMakeLists.txt files in Utilities and tests should fix those errors. I will prepare a patch for you later tonight.
Tester is not dependent on other parts of Nektar++ so can be built straight away when running the build in parallel.
Cheers, Chris
On 12 September 2018 12:32:51 BST, Mikkel Bue Lykkegaard < mikkelbue@gmail.com> wrote:
Hi Chris
I'm sorry, I think I made a mistake with the patch utility. The above output came because the root-level CMakeLists.txt was patched (running 'patch < pthread.patch'), not solvers/CMakeLists.txt. Running 'patch -p1 < pthread.patch', the correct file was patched. My fault.
The solvers are now compiled with no errors.
However, the compilation now fails similarly when linking utilities/NekMesh: https://pastebin.com/EfsqEQwR
Moreover, when compiling in parallel, the same error occurs when linking Tester (which I assume would be compiled after utilities/NekMesh, if not compiling in parallel): https://pastebin.com/qE8mfKcs
Cheers, Mikkel
On Wed, Sep 12, 2018 at 10:48 AM Chris Cantwell <c.cantwell@imperial.ac.uk> wrote:
> Hi Mikkel, > > That is strange, since it is the same CMake commands as used for the > library executables! I am also using CMake 3.12.1, so should not be related > to an API change. > > Please can you confirm: > - you are still using an unmodified version of Nektar++ v4.4.1 > - you have tried completely wiping the contents of your build directory > and configuring from scratch > - you are using the default options for the NEKTAR_BUILD_* variables. > > If you still encounter the error, please could you send me the file > sys.txt produced by running > cmake --system-information sys.txt > and we can see if that sheds any light on the problem. > > Cheers, > Chris > > > On Wed, 12 Sep 2018 10:04:25 +0100, Mikkel Bue Lykkegaard < > mikkelbue@gmail.com> wrote: > > Hi Chris > > > > Thank you. After patching CMakeLists.txt, ccmake now throws the following > > errors: > > > > CMake Error at CMakeLists.txt:10 (GET_TARGET_PROPERTY): > > GET_TARGET_PROPERTY called with incorrect number of arguments > > > > CMake Error at CMakeLists.txt:11 (GET_TARGET_PROPERTY): > > GET_TARGET_PROPERTY called with incorrect number of arguments > > > > Best, Mikkel > > > > On Wed, Sep 12, 2018 at 9:49 AM Chris Cantwell < > c.cantwell@imperial.ac.uk> > > wrote: > > > > > Hi Mikkel, > > > > > > Thanks for the output. The -pthread flag is missing from the solvers, > but > > > not on the library demos, etc. > > > > > > The attached patch should fix the issue to make the solvers consistent > > > with the library executables. We will include this in the next patch > > > release. > > > > > > Please let me know if you still have problems. > > > > > > Cheers, > > > Chris > > > > > > > > > On Wed, 12 Sep 2018 09:17:59 +0100, Mikkel Bue Lykkegaard < > > > mikkelbue@gmail.com> wrote: > > > > Hi Chris > > > > > > > > Thanks for the reply. Sure, here is the output of 'make VERBOSE=1': > > > > https://pastebin.com/AtkuDCCi > > > > > > > > Also, here are the messages from ccmake (in case it has any > relevance): > > > > https://pastebin.com/KpKYeDNM > > > > > > > > I am running Antergos (Arch) Linux with the following package > versions: > > > > Linux 4.18.6 > > > > gcc 8.2.1 > > > > CMake 3.12.1 > > > > binutils 2.31.1 > > > > BLAS 3.8.0 > > > > LAPACK 3.8.0 > > > > boost 1.67.0 > > > > > > > > Mikkel > > > > > > > > On Tue, Sep 11, 2018 at 10:37 PM Chris Cantwell < > > > c.cantwell@imperial.ac.uk> > > > > wrote: > > > > > > > > > Hi Mikkel, > > > > > > > > > > Which Linux distribution/version are you using? > > > > > > > > > > Could you send the output of running 'make VERBOSE=1'. > > > > > > > > > > Cheers, > > > > > Chris > > > > > > > > > > > > > > > On Tue, 11 Sep 2018 13:05:10 +0100, Mikkel Bue Lykkegaard < > > > > > mikkelbue@gmail.com> wrote: > > > > > > Hi all > > > > > > > > > > > > I am trying to install Nektar++ for the first time, on Linux > 4.18.6 > > > with > > > > > > CMake version 3.12.1. > > > > > > > > > > > > The compilation fails with: > > > > > > > > > > > > [ 66%] Linking CXX executable CardiacEPSolver > > > > > > /usr/bin/ld: > > > CMakeFiles/CardiacEPSolver.dir/CellModels/CellModel.cpp.o: > > > > > > undefined reference to symbol 'pthread_condattr_setclock@ > > > @GLIBC_2.3.3' > > > > > > /usr/bin/ld: /usr/lib/libpthread.so.0: error adding symbols: > DSO > > > missing > > > > > > from command line > > > > > > collect2: error: ld returned 1 exit status > > > > > > make[2]: *** > > > > > > > > > >
[solvers/CardiacEPSolver/CMakeFiles/CardiacEPSolver.dir/build.make:465:
> > > > > > solvers/CardiacEPSolver/CardiacEPSolver] Error 1 > > > > > > make[1]: *** [CMakeFiles/Makefile2:2907: > > > > > > solvers/CardiacEPSolver/CMakeFiles/CardiacEPSolver.dir/all] > Error 2 > > > > > > make: *** [Makefile:163: all] Error 2 > > > > > > > > > > > > if I disable CardiacEPSolver in the configuration, I get a > similar > > > error: > > > > > > > > > > > > [ 70%] Linking CXX executable CompressibleFlowSolver > > > > > > /usr/bin/ld: > > > > > > > > > > > > > > >
CMakeFiles/CompressibleFlowSolver.dir/ArtificialDiffusion/ArtificialDiffusion.cpp.o:
> > > > > > > > > undefined reference to symbol
'pthread_condattr_setclock@
> > > @GLIBC_2.3.3' > > > > > > /usr/bin/ld: /usr/lib/libpthread.so.0: error adding symbols: > DSO > > > missing > > > > > > from command line > > > > > > collect2: error: ld returned 1 exit status > > > > > > make[2]: *** > > > > > > > > > > > > > > >
[solvers/CompressibleFlowSolver/CMakeFiles/CompressibleFlowSolver.dir/build.make:615:
> > > > > > > > > solvers/CompressibleFlowSolver/CompressibleFlowSolver] Error 1 > > > > > > make[1]: *** [CMakeFiles/Makefile2:2925: > > > > > > > > > >
solvers/CompressibleFlowSolver/CMakeFiles/CompressibleFlowSolver.dir/all]
> > > > > > Error 2 > > > > > > make: *** [Makefile:163: all] Error 2 > > > > > > > > > > > > Can someone point me in the right direction? > > > > > > > > > > > > Thanks, Mikkel > > > > > > > > > > > > > > > -- > > > > > Chris Cantwell > > > > > Imperial College London > > > > > South Kensington Campus > > > > > London SW7 2AZ > > > > > Email: c.cantwell@imperial.ac.uk > > > > > www.imperial.ac.uk/people/c.cantwell > > > > > > > > > > > > > > -- > > > Chris Cantwell > > > Imperial College London > > > South Kensington Campus > > > London SW7 2AZ > > > Email: c.cantwell@imperial.ac.uk > > > www.imperial.ac.uk/people/c.cantwell > > > > > > -- > Chris Cantwell > Imperial College London > South Kensington Campus > London SW7 2AZ > Email: c.cantwell@imperial.ac.uk > www.imperial.ac.uk/people/c.cantwell >
-- Chris Cantwell Imperial College London South Kensington Campus London SW7 2AZ Email: c.cantwell@imperial.ac.uk www.imperial.ac.uk/people/c.cantwell
-- Chris Cantwell Imperial College London South Kensington Campus London SW7 2AZ Email: c.cantwell@imperial.ac.uk www.imperial.ac.uk/people/c.cantwell
-- Chris Cantwell Imperial College London South Kensington Campus London SW7 2AZ Email: c.cantwell@imperial.ac.uk www.imperial.ac.uk/people/c.cantwell
-- Chris Cantwell Imperial College London South Kensington Campus London SW7 2AZ Email: c.cantwell@imperial.ac.uk www.imperial.ac.uk/people/c.cantwell
participants (2)
-
Chris Cantwell
-
Mikkel Bue Lykkegaard