The SENSEI project takes aim at a set of research challenges for enabling scientific knowledge discovery within the context of in situ processing at extreme-scale concurrency. This work is motivated by a widening gap between FLOPs and I/O capacity which will make full-resolution, I/O-intensive post hoc analysis prohibitively expensive, if not impossible.

We focus on new algorithms for analysis, and visualization - topological, geometric, statistical analysis, flow field analysis, pattern detection and matching - suitable for use in an in situ context aimed specifically at enabling scientific knowledge discovery in several exemplar application areas of importance to DOE. Complementary to the in situ algorithmic work, we focus on several leading in situ infrastructures, and tackle research questions germane to enabling new algorithms to run at scale across a diversity of existing in situ implementations.

Our intent is to move the field of in situ processing in a direction where it may ultimately be possible to write an algorithm once, then have it execute in one of several different in situ software implementations. The combination of algorithmic and infrastructure work is grounded in direct interactions with specific application code teams, all of which are engaged in their own R&D aimed at evolving to the exascale.

The SENSEI library contains core base classes that declare the AnalysisAdaptor API which is used to interface to in situ infrastructures and implement custom analyses; the DataAdaptor API which AnalysisAdaptors use to access simulation data in a consistent way; and a number of implementations of both. For more information see our SC16 paper.

SENSEI ships with a number of mini-apps that demonstrate use of the SENSEI library with custom analyses and the supported in situ infrastructures. When the SENSEI library is enabled the mini-apps will make use of any of the supported in situ infrastructures that are enabled. When the SENSEI library is disabled mini-apps are restricted to the custom analysis such as histogram and autocorrelation.

More information on each mini-app is provided in the coresponding README in the mini-app's source directory.

• Oscillators The miniapp from year II generates time varying data on a uniform mesh and demonstrates usage with in situ infrasturctures, histogram, and autocorrelation analyses.

• Newton This Python n-body miniapp demonstrates usage of in situ infrastructures and custom analyses from Python.

End points are programs that receive and analyze simulation data through transport layers such as ADIOS and LibIS. The SENSEI end point uses the transport's data adaptor to read data being serialized by the transport's analysis adaptor and pass it back into a SENSEI analysis for further processing.

The SENSEI project uses CMake 3.0 or later. The CMake build options allow you to choose which of the mini-apps to build as well as which frameworks to enable. It is fine to enable multiple infrastructures, however note that Catalyst and Libsim are currently mutually exclusive options due to their respective use of different versions of VTK.

$ mkdir build
$ cd build
$ ccmake .. # set one or more -D options as needed
$ make
$ make install

cmake -DENABLE_ASCENT=ON -DVTKM_DIR=[your path] -DVTKH_DIR=[your path] \
    -DCONDUIT_DIR=[your path] -DAscent_DIR=[your path] -DVTK_DIR=[your path] \
    ..

Note that the VTK build needs to explicitly disable use of VTK-m as this will conflict with the version required by Ascent. We used the instructions for building Ascent and its dependencies (VTK-m, VTK-h, Conduit, etc) manually as described in the Ascent documentation.

cmake -DENABLE_SENSEI=ON -DENABLE_LIBSIM=ON -DVTK_DIR=[your path] -DLIBSIM_DIR=[your path] ..

VTK_DIR should point to the VTK used by Libsim.

cmake -DENABLE_SENSEI=ON -DENABLE_CATALYST=ON -DParaView_DIR=[your path] ..

Optionally, -DENABLE_CATALYST_PYTHON=ON will enable Catalyst Python scripts. Note that a development version of ParaView is required when building with both ENABLE_CATALYST and ENABLE_VTKM are enabled as released versions of ParaView (5.5.2 and earlier) do not include a modern-enough version of vtk-m.

cmake -DENABLE_SENSEI=ON -DENABLE_ADIOS1=ON -DVTK_DIR=[your path] -DADIOS_DIR=[your path] ..

Can be used with either ParaView_DIR when configuring in conjunction with Catalyst, or VTK_DIR otherwise.

In essence this is as simple as adding -DENABLE_PYTHON=ON -DSENSEI_PYTHON_VERSION=3 However, VTK (or ParaView when used with Catalyst) needs to be built with Python enabled and the SENSEI build needs to use the same version; and NumPy, mpi4py, and SWIG are required. Note that there are some caveats when used with Catalyst and Libsim. These are described in more detail in the Newton mini app README.

cmake -DENABLE_SENSEI=ON -DENABLE_VTK_IO=ON  -DVTK_DIR=[your path] ..

Can be used with either ParaView_DIR or VTK_DIR.

cmake -DENABLE_SENSEI=ON -DENABLE_VTKM=ON -DVTK_DIR=[your path] ..

Note that a development version of VTK is required when building with both ENABLE_SENSEI and ENABLE_VTKM are enabled as released versions of VTK (8.1.1 and earlier) do not include a modern-enough version of vtk-m.

To use SENSEI from your CMake based project include the SENSEI CMake config in your CMakeLists.txt.

find_package(SENSEI REQUIRED)

add_executable(myexec ...)
target_link_libraries(myexec sensei ...)

Additionally, your source code may need to include senseiConfig.h to capture compile time configuration.

The SENSEI framework includes the following software:

• DIY2, Copyright (c) 2015, The Regents of the University of California, through Lawrence Berkeley National Laboratory (subject to receipt of any required approvals from the U.S. Dept. of Energy).
• pugixml, Copyright (c) 2006-2016 Arseny Kapoulkine.

The SENSEI framework makes use of (links to) the following software:

• ADIOS 1, Copyright (c) 2008 - 2009. UT-BATTELLE, LLC. Copyright (c) 2008 - 2009. Georgia Institute of Technology.
• ParaView/Catalyst, Copyright (c) 2005-2008 Sandia Corporation, Kitware Inc. Sensei requires ParaView v5.5.1 or later when ENABLE_CATALYST is on and a development version (v5.6.0 or later) when both ENABLE_CATALYST and ENABLE_VTKM are on.
• VisIt/libsim, Copyright (c) 2000 - 2016, Lawrence Livermore National Security, LLC.
• VTK, Copyright (c) 1993-2015 Ken Martin, Will Schroeder, Bill Lorensen. Sensei can use VTK provided separately or the VTK included with VisIt/libSim (VTK v6.1 when ENABLE_LIBSIM is on) or ParaView/Catalyst (VTK v8.1 when ENABLE_CATALYST is on). If VTK is provided separately and both ENABLE_VTK and ENABLE_VTKM are on, Sensei requires a development version (VTK v9.0 or later).
• VTKm, Copyright (c) 2014-2018 NTESS, SNL, LANL, UT-Battelle, Kitware, UC Davis. A development version is currently required as packaging infrastructure has recently changed.

For full license information regarding included and used software please refer to the file THIRDPARTY_SOFTWARE_LICENSES.