VisIt

VisIt
Lawrence Livermore National Laboratory · Public domain · source
Name	VisIt
Developer	Lawrence Livermore National Laboratory; collaboration with Oak Ridge National Laboratory; contributions from Los Alamos National Laboratory; Sandia National Laboratories
Released	2002
Programming language	C++; Python; Qt
Operating system	Linux; Windows; macOS
Platform	x86; x86-64; ARM
Genre	Scientific visualization; data analysis; computational simulation post-processing
License	BSD-style

VisIt VisIt is an open-source scientific visualization and analysis application developed for large-scale simulation data. It provides interactive and batch visualization capabilities for researchers in computational science, offering rendering, plotting, and data interrogation suited to high-performance computing environments. The project emerged from national laboratory collaborations to address exascale-class datasets and supports community-driven extension through plugins and scripting.

History

VisIt originated in the early 2000s as a response to growing data volumes produced by simulation campaigns at Lawrence Livermore National Laboratory, Los Alamos National Laboratory, and Sandia National Laboratories. Funding and coordination involved programs at Department of Energy facilities and collaborations with Oak Ridge National Laboratory, reflecting priorities from initiatives like the Advanced Simulation and Computing Program. Early releases focused on visualization for shock physics and hydrodynamics codes used in projects tied to Stockpile Stewardship Program objectives. Over time, VisIt adopted community development practices influenced by open-source projects at institutions such as National Center for Supercomputing Applications and Argonne National Laboratory, while integrating ideas from visualization research at University of Utah and Lawrence Berkeley National Laboratory.

Features and Architecture

VisIt implements a client–server architecture enabling remote visualization of datasets stored on supercomputers such as Titan (supercomputer) and Summit (supercomputer). The architecture separates a lightweight graphical client from a parallel server process that performs data processing and rendering near storage systems like Lustre and GPFS. Its user interface is built on Qt (software) and supports Python scripting via interfaces inspired by NumPy and Matplotlib workflows. Rendering capabilities include ray tracing and hardware-accelerated OpenGL paths compatible with drivers from vendors like NVIDIA and Intel Corporation. The software design draws upon visualization patterns established by projects at University of Chicago and frameworks such as the Visualization Toolkit.

Supported Data Formats and File Types

VisIt supports native and community data formats used in computational science, including hierarchical and unstructured meshes common to codes developed at Los Alamos National Laboratory, Sandia National Laboratories, and Argonne National Laboratory. Supported file types encompass ADIOS, HDF5, NetCDF, CSV, and legacy formats from simulation packages used at Princeton Plasma Physics Laboratory and Fermilab. It also ingests output from CFD and finite element codes associated with ANSYS, OpenFOAM, and bespoke in-situ workflows from projects at Oak Ridge National Laboratory. Database readers leverage libraries maintained by HDF Group and community standards promoted by EuroHPC partners.

Visualization and Analysis Tools

VisIt provides a catalog of plots, operators, and queries for exploration of scalar, vector, tensor, and particle data frequently produced by experiments at Brookhaven National Laboratory and Lawrence Livermore National Laboratory. Plot types include contour, pseudocolor, streamlines, and volume rendering used in studies at NASA centers and National Renewable Energy Laboratory simulations. Operators such as thresholding, clipping, and resampling enable preconditioning workflows similar to those in ParaView and research from University of Utah visualization groups. Built-in analysis routines support time-series reductions, statistical queries, and derived field computations applied in climate modeling at National Center for Atmospheric Research and fusion research at Princeton Plasma Physics Laboratory.

Performance, Scalability, and Parallelism

VisIt targets HPC environments and parallel I/O strategies employed on systems like Frontera (supercomputer) and Sierra (supercomputer), using MPI for distributed-memory parallelism and multi-threading for node-level concurrency. Its server-side pipeline supports data streaming and out-of-core techniques informed by research from Oak Ridge National Laboratory and Argonne National Laboratory to handle terabyte- and petabyte-scale datasets. Performance tuning integrates with job schedulers common at national facilities—such as SLURM and PBS Professional—and takes advantage of GPU acceleration with toolchains from NVIDIA and standards like OpenCL where appropriate.

Development, Extensibility, and Plugins

The project adopts a plugin architecture enabling community contributions of file readers, plot types, and operators, with interfaces designed to be extended by researchers at University of California, Berkeley and other institutions. Development follows distributed version control and code review practices used in collaborations at GitHub and continuous integration strategies seen in large-scale scientific software from Cascadia National Labs. Python-based scripting and embeddable command interfaces permit automation and integration with workflows developed for Jupyter environments and workflow managers used at Los Alamos National Laboratory.

Adoption, Applications, and Use Cases

VisIt is widely used across national laboratories, universities, and research centers for visualization in domains such as astrophysics at Smithsonian Astrophysical Observatory, climate science at National Oceanic and Atmospheric Administration, nuclear engineering at Idaho National Laboratory, and computational fluid dynamics in aeronautics research at NASA Ames Research Center. Use cases include analysis of turbulence simulations, shock physics, magnetohydrodynamics, and material response studies associated with projects at Los Alamos National Laboratory, Lawrence Livermore National Laboratory, and academic groups at Massachusetts Institute of Technology and Stanford University. Its integration into pipelines supports reproducible post-processing and collaboration across teams at institutions like Oak Ridge National Laboratory and Argonne National Laboratory.

Category:Scientific visualization software