In situ Data Analysis and Visualization with SENSEI
Big Data Analytics
Visualization & Virtual Reality
TimeSunday, June 16th2pm - 6pm
DescriptionThis tutorial presents the fundamentals of in situ data analysis and visualization leveraging the SENSEI generic in situ interface. We demonstrate the infrastructure to couple simulation codes to multiple tools: ParaView, VisIt, and custom Python-based methods. Attendees will learn the basics of in situ analysis and visualization, and will gain hands on experience instrumenting simple mini apps and real simulation codes. The SENSEI in situ framework gives simulations access to a diverse set of back end data consumers, such as ADIOS, Libsim, Catalyst, VTK-m, and Python, through a single data model and API. With SENSEI, the simulation selects the back-end at runtime through an XML configuration file allowing easy interchange of analysis tools. This tutorial is timely, because a key trend facing extreme-scale computational science is the widening gap between computational and I/O rates. The challenge is how to gain insight from simulation data best when it is increasingly impractical to save it to persistent storage for subsequent visualization and analysis. One approach to this challenge is the idea of in situ processing, where one performs visualization and analysis processing while data is still resident in memory.
Content Level 65% beginner, 35% intermediate, 5% advanced
Target AudienceThis tutorial focuses on in situ visualization and analytics of results computed by parallel codes on HPC platforms and is intended for a general HPC audience, particularly those who work with large-scale parallel codes.
PrerequisitesAt a minimum attendees will require a laptop with a web browser for the hands on exercises. It is also recommended that ssh be installed. Attendees will be given the option of installing an 8MB virtual machine with the run time environment or logging into training accounts that we will provide on the Cray system Cori at NERSC to access the software run time environment.