[Paraview] Query on Paraview CPU vs GPU acceleration

[Bishwajit Dutta]

Hi All,


I am trying to compile and run Paraview 5.0.1 with both CPU and GPU
acceleration on ubuntu server with NVIDIA cards. I am able to successfully
run a pvpython script.

However in both below cases I am able to see CPU power getting consumed
when the script is run.
I am confused on the accelerator being used.

For GPU:
=======
I compiled with the below options:

cmake ../ -DBUILD_TESTING=OFF -DCMAKE_BUILD_TYPE=Release
-DPARAVIEW_ENABLE_CATALYST=ON -DPARAVIEW_ENABLE_PYTHON=ON
-DPYTHON_LIBRARY=$PYTHON_LIBRARY -DPYTHON_INCLUDE_DIR=$PYTHON_INCLUDE_DIR
-DPARAVIEW_USE_MPI=ON -DMPI_HEADER_PATH=$MPI_HEADER_PATH
-DCMAKE_INSTALL_PREFIX=$BD_INSTALL_PATH -DPARAVIEW_BUILD_QT_GUI=OFF
-DVTK_USE_X=OFF

For CPU:
========
I compiled mesa, glu libs etc. and the compiled Paraview with the below
options

I run the script with
#pvpython --mesa-llvm <script_name>

cmake ../ -DBUILD_TESTING=OFF -DCMAKE_BUILD_TYPE=Release
-DPARAVIEW_ENABLE_CATALYST=ON -DPARAVIEW_ENABLE_PYTHON=ON
-DPYTHON_LIBRARY=$PYTHON_LIBRARY -DPYTHON_INCLUDE_DIR=$PYTHON_INCLUDE_DIR
-DPARAVIEW_USE_MPI=ON -DMPI_HEADER_PATH=$MPI_HEADER_PATH
-DCMAKE_INSTALL_PREFIX=$BD_INSTALL_PATH -DPARAVIEW_BUILD_QT_GUI=OFF
-DVTK_USE_X=OFF -DOPENGL_INCLUDE_DIR=$BD_INSTALL_PATH/include
-DOPENGL_gl_LIBRARY=$BD_INSTALL_PATH/lib/libOSMesa.so
-DOPENGL_glu_LIBRARY=$BD_INSTALL_PATH/lib/libGLU.so
-DVTK_OPENGL_HAS_OSMESA=ON -DOSMESA_INCLUDE_DIR=$BD_INSTALL_PATH/include
-DOSMESA_LIBRARY=$BD_INSTALL_PATH/lib/libOSMesa.so

Thanks in advance for help on this.

If this all is running in CPU then how do I compile it for GPU without
x-windows (i.e. on server)


BR,
Bishwajit
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/paraview/attachments/20170227/d3768f54/attachment.html>