missing Cnakefile.txt

[A-DCA]

Hello,

after running install_base/libs.sh successfully, install_peridem.sh failed with error message as follows:

jma@EGIS-PG-105:~/PeriDEM$ ./tools/script/build_scripts/ubuntu-18.04/install_peridem.sh Release
SCRIPTPATH = /home/jma/PeriDEM/tools/script/build_scripts/ubuntu-18.04
SOURCEDIR = /home/jma/PeriDEM/tools/script/build_scripts/ubuntu-18.04/source/
BUILDDIR = /home/jma/PeriDEM/tools/script/build_scripts/ubuntu-18.04/build/
Shell: /bin/sh
<<<<<<<<<<< >>>>>>>>>>>
PERIDEM
<<<<<<<<<<< >>>>>>>>>>>
CMake Error: The source directory "/home/jma/PeriDEM/tools/script/build_scripts/ubuntu-18.04/source/peridem/0.1.0" does not appear to contain CMakeLists.txt.
Specify --help for usage, or press the help button on the CMake GUI.
make: *** No targets specified and no makefile found. Stop.

Any suggestion?

Regards,

Jingsheng Ma

[github-actions]

Hey # thanks for reporting an issue! We will try to get in touch with you soon.

[prashjha]

Hi Jingsheng,

I repeated the steps by creating 'PeriDEM' in my home directory so except 'prashant' in place of 'jma', directories and everything is same. While running 'install_peridem.sh', I got different error than yours:

CMake Error at CMakeLists.txt:83 (find_package):
  By not providing "FindHPX.cmake" in CMAKE_MODULE_PATH this project has
  asked CMake to find a package configuration file provided by "HPX", but
  CMake did not find one.

  Could not find a package configuration file provided by "HPX" with any of
  the following names:

    HPXConfig.cmake
    hpx-config.cmake

  Add the installation prefix of "HPX" to CMAKE_PREFIX_PATH or set "HPX_DIR"
  to a directory containing one of the above files.  If "HPX" provides a
  separate development package or SDK, be sure it has been installed.


-- Configuring incomplete, errors occurred!
See also "/home/prashant/PeriDEM/tools/script/build_scripts/ubuntu-18.04/peridem/CMakeFiles/CMakeOutput.log".

And this error is simply due to the fact that I did not provide the correct HPX path. If I provide HPX_INSTALL_PATH correctly, PeriDEM is built. In summary, the script did not show the error

CMake Error: The source directory "/home/jma/PeriDEM/tools/script/build_scripts/ubuntu-18.04/source/peridem/0.1.0" does not appear to contain CMakeLists.txt.

which suggests that in your case, for some reason,

1. PeriDEM was not cloned properly
2. Or you had an empty PeriDEM source directory so this block of code was not executed

if [ ! -d $PERIDEM_SOURCE_DIR ]; then
    mkdir -p $PERIDEM_SOURCE_DIR
    git clone git@github.com:prashjha/PeriDEM.git $PERIDEM_SOURCE_DIR
fi

I will recommend checking if PeriDEM is cloned to the directory PERIDEM_SOURCE_DIR and if not first delete this directory if it still exists and then rerun the script.

Also, please update your local copy of PeriDEM. In this latest commit '8c5a523', I enhance the install script so that it writes the paths such as HPX path to a file allowing one to copy this path and add it to 'install_peridem.sh'.

[A-DCA]

Hi Prashant,

Thanks for your reply and suggestion.

I managed installing PeriDem. I think the problem stemmed from the wget while in my case I run the install_perimed.sh in the package I downloaded previously. Now I went back the top of level and run cmake there after having associated packages installed.

I run the examples, two_particles, afterward and they all seemed to be ok by looking at logs. But in each out dir, I load output_xxx.vtu using paraview 5.9 but did not see any object though I can see the outline after zooming in. Not sure what I did wrong. I enclosed files for circle_damp here your information - circ_damp.zip

Kind regards

Jingsheng

[prashjha]

Good to know that it now compiles and test problem runs. I looked at the your simulation results and the vtu files are good. You can visualize the results by selecting 'Point Gaussian' view in paraview (click on the button next to 'Solid Color' where by default it paraview has 'Surface', you need to select 'Point Gaussian' instead of 'Surface'). After selecting 'Point Gaussian', you may need to zoom the plot by clicking on the two cross arrows button next to 'Point Gaussian'.

I discuss how to visualize results in README file; see this link.

[A-DCA]

Thanks for pointing this. Visualization works now.

I run compressive_test\n500_circ_hex\run1 2 hours ago and it does not seem to complete the first 10% of calculation. Here are stats from 'top' command.

I tried set threads=1, 2 and 4. No difference it makes.

Thanks

#####################################
top - 22:03:07 up 1:36, 0 users, load average: 0.52, 0.58, 0.59
Tasks: 9 total, 1 running, 8 sleeping, 0 stopped, 0 zombie
%Cpu(s): 8.7 us, 1.4 sy, 0.0 ni, 89.9 id, 0.0 wa, 0.0 hi, 0.0 si, 0.0 st
KiB Mem : 33328424 total, 25505020 free, 7594052 used, 229352 buff/cache
KiB Swap: 31178524 total, 31178524 free, 0 used. 25600640 avail Mem

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
390 jma 20 0 264176 58228 13008 S 100.0 0.2 1:48.88 PeriDEM

[prashjha]

This simulation is very big as it consists of 502 particles with each particle discretized using meshfree. We have nonlocal interaction within each particle and then particle-particle interaction is also nonlocal so the compute time is quite high. For my paper, I used 26 threads to get results quickly.

Try running with a higher number of threads or try changing the setup with a smaller number of particles. Let me know if you need help with creating a new setup similar to this but with smaller particles.

[A-DCA]

On which system did you run your case with 26 threads? how long did it take roughly?

To reduce the number of particles, is it the simplest way by increasing the size of particles or reducing the initial domain in the python script.

[A-DCA]

It would help if you could give some ideas on these:

On which system did you run your case with 26 threads? how long did it take roughly?

To reduce the number of particles, is it the simplest way by increasing the size of particles or reducing the initial domain in the python script?

Thanks,