Scientific Visualization 210: Introduction to In Situ Visualization

Why In Situ? Stop Waiting for Your Data!

Are you tired of waiting hours for a simulation to finish, only to find out it crashed or the output isn't what you expected? In Situ Visualization is the game-changer. Instead of writing massive files to disk and processing them later (“Post-Hoc”), In Situ allows you to “peer inside” the memory of a running simulation. It’s like live-streaming your science. By processing data while it’s still “hot” in memory, you can:

• Save 90%+ on storage costs by only saving the images or analysis you need.
• Identify errors instantly—no more wasting compute time on failed runs.
• Achieve Exascale performance—run analysis at the same scale as your simulation.

What We Will Do

This is an all-action, hands-on session. We aren't just talking about theory; we are running the code. Using a containerized Gray-Scott Reaction-Diffusion miniapp, we will spend the majority of our time in the terminal and ParaView, connecting live simulations to four powerhouse backends: ADIOS2, Catalyst, Ascent, and Kombyne.

Agenda