The High Performance Computing Center at Stanford University was founded to provide high performance computing resources & services to enable computationally-intensive research within the School of Engineering.
Advanced Scientific Visualization resources including computing systems with high performance graphics hardware, large displays, display walls, and high-end pre/post processing facilities to enable large data analysis and promote discovery.
Impact on a Real World Problem
It's hard not to appreciate the immediate impact of improving the performance of a Formula 1 racecar. One of
the first projects loaded into the visualization cluster at
Stanford's Flow Physics and Computational Engineering
Group involved animating the turbulent flow of a Formula
1 racecar past a sophisticated model on the visualization
cluster (Figure 1).
The process turned out to be far more useful than animating it on a single-display (or even dual-display) worksta
tion. The superior resolution and larger display size of the
visualization cluster revealed salient features of the exterior and interior flow fields that were not visible on typical
visualization workstations. This contributed to a better
understanding of the mechanism of diffuser stall and its
impact on downforce. The large display area of the visualization cluster also enabled a group gathering and discussion of the observed flow, and a brain storming session on the follow-up research steps for this project. Visualization walls often lead to a collaborative effort by scientists, bringing the display and analysis process into a
conference-type setting, which breaks researchers free of
the single-computer, single-scientist approach to analysis.
The visualization cluster also proved extremely useful in the
analysis of the computational results in another project at
Stanford's Center for Integrated Turbulence Simulations:
integrated simulations of turbulent flow in an entire turbofan
jet engine (Figure 2). Jet engine is a complex engineering system
consisting of a fan, compressor, combustor, turbine and exit
nozzle. Visualizing the details of the flow in each of those
components, such as boundary layers on turbumachinery
blades, tip-gap vortices and unsteady wakes in the blade
passages, as well as the details of the combustion process,
including the fuel sprays and the flame front, was made much
easier using the large display of the visualization cluster. In
addition, the use of the visualization cluster helped with the
understanding of the interaction between the individual
components of the engine. Using single displays, even with
zooming, proved inadequate, because the entire ensemble
needed to be visualized at once, with a sufficiently large image.
For example, the fan blade wakes propagate through multiple
stages of the low pressure compressor and that phenomenon
could be analyzed in much greater detail on the large display
of the visualization cluster. Finally, tackling the flow in a jet
engine required a collaboration of a dozen research scientists -
and again, the visualization of the flow features on a large screen
proved to be an excellent vehicle for a fruitful scientific interaction.