HPC & Visualization Visualization and High-Performance Computing Visualization is a critical step in gaining in-depth insight into research problems, empowering understanding that is not possible with raw numbers alone. Through visualization, researchers can explore their data intuitively, with relationships and interesting anomalies made clear through the visualization process. Through high-performance computing, researchers are able to address more complex problems, while creating datasets of unprecedented size. The research community is facing the challenge of how to manage, analyze, and visualize such data in a meaningful way. With expertise in high-performance, distributed visualization and data processing, Kitware is tackling these challenges by further developing ParaView, the Visualization Toolkit (VTK), and other solutions. Our cutting-edge toolkits and custom solutions enable real-time investigation of datasets, and provide options for both end-user applications and software development tools in support of data processing, visualization, and analysis.
Scientific Visualization of Large Data Data analysis is often the most critical bottleneck in the scientific discovery process. The exponential growth of datasets in size, scale, complexity, and richness makes it difficult for researchers and scientists to analyze data and obtain insight. As data approaches Exabyte scale, this problem becomes even more pressing. Moreover, the data is becoming increasingly complex not only because of enhanced resolution, but also through the integration of experimental observations and associated metadata within the dataset. Kitware is addressing these large data analysis issues by providing expertise in high-performance computing, distributed visualization, and data processing. With the computational infrastructure and tools to power large data analysis and visualization, scientists are able to tackle today s most pressing research challenges. Kitware s scientific visualization capabilities and toolkits, in particular VTK and ParaView, are being leveraged in many domains, as highlighted in the Visualization in Action section on the right. Large Data Informatics Informatics is the fusion of information science and information systems engineering for effectively presenting and visualizing meaningful data. By creating advanced informatics applications to analyze and visualize big data, we are empowering researchers to see the patterns, synergies, and unique outliers in their data. At Kitware, we are developing open-source solutions for analyzing and visualizing a variety of information data, from text documents to metabolic pathways. Text analysis and visualization techniques condense millions of diverse documents and social media content into summary diagrams that may be interactively traversed to find interesting new facts and relationships. Network analysis tools discover outliers in voluminous network traffic data in order to detect and avoid cyber attacks. In bioinformatics, we are designing and incorporating a new cloud-based architecture for genetics algorithms, integrating new visualization tools, and developing the ability to link algorithms to existing databases. With our tools, clinical researchers can analyze coupled data from experiments, environments and organisms; interface with large evolutionary trees for evolutionary data mining; and integrate anatomical imaging, pathology, and omics data for rich analysis.
Visualization in Action Mobile & Web Visualization As datasets grow and teams and resources become geographically dispersed, tools that enable collaborative, long-distance research will play a critical role in scientific innovation. Such frameworks are necessary to readily facilitate the sharing, analysis, and visualization of large data over the web and on mobile devices. Kitware is developing applications and libraries to meet these needs in several fields, bringing our robust visualization and analysis capabilities to developers and end-users in entirely new ways. Kitware has developed VES, a mobile / embedded system rendering library that integrates VTK with OpenGL ES to deliver scientific and medical visualization capabilities to mobile application developers. VES has been used by developers worldwide to create applications on ios, Android, QNX, and MeeGO platforms. VES also powers KiwiViewer, an application framework that delivers easy-to-use building blocks for embedding advanced visualizations. Kitware s ParaViewWeb tackles the challenge of web visualization by leveraging the latest HTML 5.0 technologies to enable client-server communication with a remote visualization node or cluster using a lightweight JavaScript API. With ParaViewWeb, web applications can easily embed interactive 3D visualization components and researchers can readily collaborate through web-based scientific workflows. ParaView in an Immersive Environment The open-source ParaView Immersive delivers sophisticated highperformance computing and visualization capabilities to the scientific community within immersive environments. Used in labs such as the Visualization and Collaboration Theater at Indiana University, ParaView Immersive supports various stereo rendering techniques and new ways of interacting with data by utilizing device inputs via VRPN or VRUI. Climate and Geovisualization The Climate Data Analysis Tools project is using ParaView in conjunction with other open-source tools to create a robust application for analyzing and visualizing climate data. These tools will enable analysts to track, monitor and predict climate changes and develop solutions to climate challenges.
ParaView Catalyst: In situ Analysis ParaView Catalyst is an open-source data analysis and visualization library designed to be directly embedded into parallel simulation codes to perform in situ analysis at run time. The in situ approach, commonly referred to as coprocessing, addresses the IO challenge of storing massive simulation output data by tightly coupling simulation, data analysis, and visualization codes. ParaView s co-processing tools were coupled with PHASTA, a massively parallel CFD code developed by Ken Jansen from the University of Colorado, to simulate a synthetic jet s impact on flow over an aircraft wing. Results were obtained while the simulation code was running on more than 160,000 cores, enabling designers to get immediate feedback and adjust the jet s flow control parameters during the simulation, rather than waiting for final results. Magnetic Reconnection Researchers from Los Alamos National Laboratory and the University of California at San Diego used ParaView to visualize the output of a simulation that was run on 196,608 cores of Jaguar at Oak Ridge National Laboratory. The goal of the project was to study the magnetic reconnection in the Earth s magnetosphere using fully kinetic particle-in-cell simulations. Due to the size of the dataset (~200 TB), analysis and visualization presented significant challenges. However, by harnessing the power of parallel processing through ParaView, the data was successfully analyzed on Jaguar. This image shows the development of a turbulent reconnection in a large-scale electron-positron plasma using an isosurface of particle density colored by current density. ParaFoam: Post-processing in OpenFoam The OpenFOAM computational fluid dynamics package and ParaView have experienced tremendous growth and adoption through the efforts of the open-source community. For example, ParaView s OpenFOAM format reader was developed by Terry Jordan, Takuya Oshima, and others from the shared developer community and then contributed back to ParaView. Since ParaView is conveniently packaged in OpenFOAM distributions, this enables any user to couple the scalable simulation capabilities of OpenFOAM with the scalable visualization and analysis capabilities of ParaView in order to confidently approach any industrial design problem.
Computational Chemistry Chemistry and in-silico studies are playing an increasingly significant role in many areas of research. Kitware has developed a state-of-the-art suite of tools for computational chemists, making the premier computational chemistry codes and databases accessible to practitioners. We have developed three open-source, cross-platform desktop applications addressing major areas: Avogadro 2 for 3D editing, pre- and post-processing of data, and advanced visualization; MongoChem to store, index, and retrieve chemical data; and MoleQueue to provide seamless integration of HPC resources on the desktop. The applications can be used in unison, or as separate components to integrate with existing systems. Preprocessing & Simulation Lifecycle Tools Kitware s efforts in the area of pre-processing include the development of toolkits and frameworks that can address the needs of defining numerical simulations. This includes describing the geometric domains and the specification of simulation information, such as boundary conditions and material properties, to these domains. One example of these efforts is the Computational Model Builder project being developed for the U.S. Army Engineering Research and Development Center (ERDC). The Computational Model Builder (CMB) application suite is based on ParaView s client/server architecture and targets the needs of hydrological simulations. The CMB suite includes tools for processing scatter point data, creating surface and volumetric domains, associating simulation information, and creating and manipulating of computational meshes.
Consulting & Collaboration The Kitware Advantage Develop custom visualization applications that scale from a tablet to a supercomputer with Kitware. Kitware provides a variety of advanced software development services for commercial organizations, academic institutions, and government agencies. We have experience both in leading large-scale collaborative teams to tackle some of today s toughest research challenges, and in developing custom, proprietary software applications for our commercial customers. Specific examples of what Kitware can do for you include: Create custom, proprietary extensions to VTK or ParaView to solve your visualization needs. Integrate with your simulation package to visualize results as they are generated. Develop a domain-specific visual data analysis application to meet your specifications. Extend and integrate our open-source software toolkits and applications into your workflow. Tailor ParaView to accommodate your simulation data for verification and visualization. Awards and Recognition Kitware is an enthusiastic proponent of open-source business models and community involvement. As an indication of our dedication to the work we do, we would like to highlight the following awards and recognition: HPCwire Readers and Editors Choice Awards for Best HPC Visualization Product or Technology. These awards, presented to VTK in 2011 and ParaView in 2010 and 2012, are determined through online polling of the global HPCwire audience, along with a rigorous selection process involving HPCwire editors and industry luminaries. Kitware was honored to accept the awards on behalf of the global VTK and ParaView communities, and to be recognized as an HPC-related company that is making a significant impact in the field. Inc. 5000 List Kitware has been on the Inc. 5000 list since 2008, marking us as one of the fastest-growing companies in the country. In 2012, we were ranked #1,245 overall based on 248% growth over the preceeding three years. Kitware, Inc., 28 Corporate Drive, Clifton Park, NY 12065 Phone: (518) 371-3971 Fax: (518) 371-4573 www.kitware.com