High Performance Computing
 Featured Research ...

pasquinelli 1

Molecular Self-Assembly in Bicomponent Polymer Filaments         Dr. Melissa Pasquinelli and collaborators, funded by the Nonwovens Cooperative Research Center, have designed innovative new nnovative new polymer filaments that exhibit thermally-responsive shape memory (patent pending) by exploiting both the elasticity afforded by supramolecular networks formed by thermoplastic elastomers and also the interfacial characteristics created during the processing of bicomponent filaments. As a key aspect of this technology development, dissipative particle dynamics (DPD) simulations were used to predict the polymer behavior in their melt phase and to provide details about the interfacial interactions and molecular network formation that occurs during the self-assembly (via microphase segregation) of the polymer molecules during filament processing. The results from these DPD simulations ascertain the strenth, dynamics and extent of intermixing in these systems as a result of processing conditions and chemical compositions. The predicted trends from the simulations have been pivotal in choosing polymers for scalable fabrication of thermally-responsive polymer filaments. Applications are heat-sensitive textiles and form-fitting materials.[read more]

Future Patterns of Urban Growth

collazo 1 collazo 1

Jaime Collazo and the Biodiversity and Spatial Information Center are part of a collaborative effort funded by the USGS and NCSU Southeast Climate Science Center to simulate future patterns of urban growth across a nine state region over the next 50 years. Here the upper image shows 2010 urbanization, the lower a projection for 2060. Dr. Adam Terando, Curtis Belyea, and Jim White use the NCSU HPC blade center as part of their refinement of the SLEUTH-3r urbanization model. This is a cellular automata model that requires large amounts of computer memory and multi-processing capabilities in order to simultaneously estimate and update urbanization probabilities for 300 million points on the model grid. Primary funding is provided by the US Geological Survey through the Southeast Regional Assessment Project and the Department of Interior Southeast Climate Science Center located on the NCSU campus.

Predictive Models for High-Throughput Genetic Data

Alison Motsinger-Reif As members of the
Bioinformatics Research Center
and the Department of Statistics,
Dr. Alison Motsinger-Reif and her group develop methods for finding predictive models in high-throughput genomic data. She aims to develop analytic strategies to identify complexes of genetic and environmental factors. In particular, Dr. Motsinger's computational methods aim to detect genetic risk factors of common complex disease and drug response outcomes in human, canine, and porcine populations. Distributed and shared memory computations on the blade center have been key to her research, both in methodological development and real data applications.

Spring 2011: Maize Diversity
Jim Holland Dr. James Holland with the USDA-ARS maize breeding and genetics program at NCSU aims to understand and exploit the vast diversity of maize found world-wide. By studying the genetic variation of maize and how it relates to phenotypic variation for important agricultural traits, this research aims to enhance the genetic base of the U.S. corn crop.

In collaboration with other NCSU faculty, Major Goodman, Peter Balint-Kurti, Gary Payne and others, Dr. Holland is identifying new sources of improved productivity, stress resistance, and disease resistance in tropical maize. To understand the genetic basis of complex traits in maize, Holland and other USDA collaborators created the maize Nested Association Mapping population, a collection of 5000 lines representing much of the variability in maize. They conducted the single largest genetic mapping experiment ever in plants, involving evaluating these 5000 in 11 environments for diverse quantitative traits. Analyzing this huge experiment requires significant high performance computing resources. Dr. Holland uses the HPC blade center to perform statistical analysis of this trait data, which has enabled high resolution genetic mapping of important complex traits in maize. The HPC staff provided support to enable the use of specialized software on the blade center.

Summer 2010
Ocean Observing and Modeling Group -- Modeling Gulf Currents Ruoying He
The Ocean Observing and Modeling Group is headed by Dr. Ruoying He , with the team currently including Dr. Kyung Hoon Hyun, Dr. Zuo Xue, Ke Chen, Yizhen Li, Joseph Zambon, Hui Qian, Zhigan Yao, Yanlin Gong, Jill Nelson, Andrew Stieneke, and Yao Zhao. The group performs research in coastal and estuarine circulation, bio-physical interactions, numerical modeling, data simulation, and design and implementation of a coastal ocean observing system. Their research is supported by NSF, NOAA, ONR, NASA, USGS, and DOE, and uses the NCSU Blade Center to model ocean circulation. The graphic is generated using the South Atlantic Bight and Gulf of Mexico (so-called NC State SABGOM ) model. This model (along with weather prediction) is run daily on a myrinet equipped subcluster of the NCSU Blade Center, predicting present and future (84 hour) "ocean weather".

NOAA emergency response division has been using the SABGOM ocean current nowcast and forecast (along with 3 other ocean models) to generate an official oil trajectory prediction, used to guide responses of the local, state, and federal governments. See DeepwaterHorizon and today's oil trajectory prediction.

Fall 2009
Linear/Nonlinear Equations and Multilevel Methods Tim Kelley
Dr. Tim Kelley and his team of mathematicians (left to right, Anne Costolanski, Kai Fan, Corey Winton, Dave Mokrauer, Tim Kelley, Anna Meade, and Deena Hannoun) explore algorithms for solving systems of linear and nonlinear equations, multievel methods for integral equations, radiative transfer problems, optimal control, large scale optimization, optimization of noisy functions, and flow in porous media. Dr. Kelley has recently been selected as a Society of Industrial and Applied Mathematics Fellow for his work in porous flow.

Dr. Kelley supports Matlab software for Linear Equations, Nonlinear Equations, and Optimization. His group supports Matlab codes for optimization of noisy functions. They use the blade center to solve large distributed problems using the DOE ACTS software Trilinos, as well as smaller problems using many copies of matlab and the chemistry software Gaussian. Dr. Kelley's work is partly supported by the Army Research Office, which has purchased several hundred of the processors currently in use on the Blade Center.


RNA in Biopolymer Nanostructures Yara Yingling and RNA
Dr. Yara Yingling and her group in materials science explore biopolymeric and composite nanoassembly processes. The NCSU Blade Center enables Dr. Yingling to perform complex simulations to guide efforts in synthesis and to analyze properties of biopolymers.

For example, RNA based biopolymers can be used as nanoarrays and nanocircuits. Dr. Yingling proposes to use protein-free RNA nanoparticles as efficient minimally toxic drug delivery devices.

Dr. Yingling and her team designed an RNA hexagonal nanoring formed from six simple building blocks held together via loop-loop contacts. Drugs or therapeutic agents can be incorporated within or attached to the ends of building blocks. Moreover, the 3' and 5' sequence can be engineered to facilitate a self-assembly of these rings into a nanotube.

Winter 2008
Blonding and SuperNova
Dr. Blondin's work was featured in a recent RESULTS article . Astrophysics professor John Blondin recalls staring into the nighttime Wisconsin skies when he was growing up and marveling at the stars. Even as a kid, I can remember being fascinated by the fact that our world here on Earth is such a small piece of the universe, he says. Now, he exudes the same boyish enthusiasm as he marvels at some of the world's most powerful computers, which are helping him solve a vexing problem of the universe: How does a supernova explode?

Summer 2006 Featured Research

Tools for Tuning Dr. Frank Mueller and his research team in the Department of Computer Science are developing software tools to tune and scale applications in the high performance computing area. The tools detect and alleviate sources of scalability problems and use runtime/operating system synergy to exploit shared-memory multi-processors and simultaneous multithreading for shared memory computing. As multi-core architectures spread to the desktop (and the laptop), so does the applicability of their work in shared memory computation.

Spring 2006 Featured Research

Carbon lattice with oxygen
molecure Dr. Marco Buongiorno Nardelli in the Department of Physics and his research group are using first principles simulations to study the application of carbon nanotubes as a potential method to produce hydrogen by dissociation of water. Defects in the carbon lattice structure provide pathways that significantly reduce the energy required for the dissociation of water. An economical and environmentally friendly source of hydrogen could provide an important future energy source.

Fall 2005 Featured Research

Genomics Research Laboratory - picture of Partner's II Building - Sequence data Dr. Yang Zhang's Air Quality Forecasting Laboratory applies state-of-the-science numerical models to simulate human-induced air pollution and its impact on human health, climate, the environment, and society. By using university HPC resources, the AQF Lab is able to focus on constructing, improving, and implementing atmospheric modeling systems.

Spring 2005 Featured Research

Genomics Research Laboraotry -picture of Partner's II Building -
Sequence data The Genome Research Laboratory transitioned the computationally intensive processing of sequencer output from an internal cluster to the university Linux center. By utilizing the university resource for the computationally intensive work the GRL has been able to focus their resources on the sequence generation services the Laboratory provides.

Winter 2005 Featured Research

Dr. Blondin's, Numerical supernova simulation results
NC State researchers are simulating the death of a massive star leading to a supernova explosion. Of particular interest is the dynamics of the shock wave generated by the initial implosion of the star which ultimately destroys the star as a highly energetic supernova.

Dr. John Blondin and his research group work with models of supernova as part of the Department of Energy sponsored TeraScale Supernova Initiative. In addition to NC State HPC resources, the group also uses the computational resources at Oak Ridge National Laboratory and, for visualization of model results, their own 22-node Linux cluster.


Fall 2004 Featured Research

Dr. Edwards, NASA Scramjet in flight, Numerical simulation results

NC State researchers are assisting in the development of scramjet engines, such as the one that powered NASAs X-43A on its record breaking flight November 16, 2004. Dr. Jack Edwards and Jason Star of the NC State Aerospace and Computational Fluid Dynamics Laboratory are performing numerical simulations of the starting and stopping processes within scramjet engines. A typical transient run simulation may use up to 50 processors of the NC State cluster and may take up to 10 days to complete.

The work is sponsored by NASA Langley Research Center and the National Institute of Aerospace.