In situ visualization and analysis, where analysis is performed while simulation data is still resident in system memory, is increasingly necessary for large-scale simulations that maintain more state than can be reasonably written to the file system. An effective in situ implementation should preserve simulation performance and accuracy while providing sufficient resources to an appropriate visualization algorithm. It is rare, however, for simulation and visualization experts to both be part of an implementation team: a simulation scientist alone might choose a simple visualization that does not express all available features; a visualization scientist alone might choose a simple integration path that breaks simulation performance and scalability. To address this problem, Intel, IXPUG and TACC assembled a collection of simulation and visualization experts for a three-day hackathon to implement in situ methods in a set of representative codes, including COSMOS GR-CHOMBO, UinTah and LAMMPS. This talk presents the findings and best practices distilled from the event, including both practical implementation advice and lessons learned that generalize to other implementation efforts.
Jim Jeffers is the Director, Principal Engineer of Software Defined Visualization Engineering for Intel’s Enterprise and Government Group. Jim joined Intel in 2008 participating in the development of many-core parallel computing and the Intel® Xeon Phi™ product family. Jim’s experience includes software design and technical leadership in high performance computing, graphics, digital television, and data communications. Jim has coauthored 4 leading parallel programming books including Intel® Xeon Phi™ Coprocessor High Performance Programming (Morgan Kaufmann, 2013); High Performance Parallelism Pearls (Morgan Kaufman, Volume 1, 2014; Volume 2, 2015) and Intel® Xeon Phi™ Processor High Performance Programming, Knights Landing Edition (Morgan Kaufmann, July 2016). Jim’s notable prior work includes development for the virtual ‘First Down Line’ technology seen on live American football TV broadcasts. Jim received an Intel Achievement Award for his role in launching the first Intel Xeon Phi products and has three granted US patents.
Paul A. Navrátil is an expert in high-performance visualization technologies, accelerator-based computing and advanced rendering techniques at the Texas Advanced Computing Center (TACC) at The University of Texas at Austin. His research interests include efficient algorithms for large-scale parallel visualization and data analysis (VDA) and innovative design for large-scale VDA systems. Dr. Navrátil’s recent work includes algorithms for large-scale distributed-memory ray tracing. This work enables photo-realistic rendering of the largest datasets produced on supercomputers today, such as cosmologic simulations of the Universe and computational fluid dynamics simulations at unprecedented levels of detail. He manages the Scalable Visualization Technologies group at TACC and TACC's remote visualization environments, including Maverick, the world's largest supercomputer dedicated to VDA. Dr. Navrátil's work has been featured in numerous venues, both nationally and internationally, including the New York Times, Discover, and PBS News Hour. He holds BS,MS and Ph.D. degrees in Computer Science and a BA in Plan II Honors from the University of Texas at Austin.