Scientific visualization and analysis are important components of HPC simulation workflows. For decades, data was first saved from the simulation code to disk so that scientists could retrieve the data for analysis at a later time. As the growth rate between I/O and computing speeds continues to diverge, this approach is becoming increasingly time-consuming and costly. In-situ processing helps to overcome these I/O bottlenecks by performing simulation and visualization calculations with higher spatial and temporal resolution in-memory and avoiding the transfer of raw data to hard drives.
This workshop will introduce ParaView/Catalyst and Ascent, two open-source implementations that enable in-situ processing. Both packages use Conduit software to intuitively describe hierarchical scientific data in C++, C, Fortran and Python. You will learn how to describe simulation data with Conduit and how Ascent or Catalyst can be used to transform data, render images and export results.