SYSTEMS

Researchers at the University of Houston and the University of Rennes are leveraging several SiCortex high-productivity computing (HPC) systems at the Houston site to drive a research project aimed at protecting the ecology of prairial systems.

Led by Professor Marc Garbey of University of Houston and Assistant Professor Cendrine Mony of University of Rennes, the Virtual Prairie Project (ViP) is studying agricultural management practices and their effects on plant competition and genetic prairie structures. The researchers’ respective teams are working together to evaluate and compare the unique make-up of naturally occurring ecological systems and of farm-fabricated environments. Researchers at both university project sites will use the information obtained from these fellows to collaboratively design new grassland systems that achieve the perfect balance of water purification, biodiversity conservation and carbon storage needed for true environmental sustainability.

“Collaborations like this project demonstrate the value and impact of applying high- productivity computing power to solve complex environmental issues — ecological sustainability in this case,” said Christopher Stone, president and CEO of SiCortex, maker of the most energy-efficient computers. “It is appropriate that the University of Houston chose a system that achieves a high level of compute power at the smallest carbon footprint to conduct environmental research.”

The ViP project has already far exceeded the size of any standard ecological simulation today. In phase one of the project, researchers modeled individual plant space colonization through clonal reproduction to make initial hypotheses about the complex relationships that exist among plant types. More than 22 million simulations later, the data revealed that not one single, but a few different strategies should be leveraged to achieve the best growth. The second phase of the project will run thousands of combinations of plant types in a single master simulation over an entire season, aiming to identify the optimal combinations for long-term sustainability for a prairial system under a variety of conditions.

The compact, energy-efficient design of the SiCortex system, powered by 648 64-bit processors, makes it ideally suited to run this complex simulation. Equally advantageous to the researchers is the fact that the SiCortex machine, a complete 2.1 TFlops computing system packaged in a single cabinet, is a joint capital venture between the University of Houston department of computer science and the Research Computing Center (RCC), a data center facility dedicated to serving researchers who have intensive computational tasks. The RCC team, together with Garbey and the lab led by internationally-renowned multicore and software optimization expert Barbara Chapman, are leveraging the SiCortex system in an overall effort to promote green parallel computing.

This SiCortex computer is the third such system for the University of Houston.

“It is our ultimate goal to research and design new computational frameworks, algorithms and methods that solve challenging problems in life sciences and ultimately promote environmental sustainability, and the SiCortex systems are integral to developing the applications that will drive this initiative,” said Garbey, chair of the University of Houston’s department of computer science. “Our work will have lasting impact — contributing to the design of prairies with high ecological value.”