Attendees outside the  Santa Tecla Palace on Sicily’s southeastern shore. Image courtesy Summer School

The Santa Tecla Palace on Sicily’s southeastern shore was recently a classroom for a summer school dedicated to fostering collaboration and innovation in computational science among graduate and postdoctoral scholars from Europe and the United States.

A joint effort of the EU’s DEISA and America’s TeraGrid, it provided a multicultural student community the opportunity to learn about high performance computing (HPC) resources, tools and methods.

“We hope to continue with such events every year — alternating between EU and US destinations,” said Hermann Lederer, who presented a DEISA infrastructure and service overview.

Sixty graduate and postdoctoral scholars from 20 nations were selected from more than 100 applications. Participant expenses were paid by DEISA and TeraGrid. “The number of applications received from US and European computational scholars reflects a broad interest in international collaboration and HPC-related training,” said John Towns, TeraGrid’s forum chair.  “We were especially happy with the broad range of disciplines represented by the student community,” he added. In addition to the vast range of research arenas and nationalities represented, about 20 percent were female.

The program opened with a keynote presentation by Barry Schneider from the Office of Cyberinfrastructure at the National Science Foundation, who said “Ever since Galilean times, through the early works of Stephen Hawking, scientific breakthroughs have been largely achieved by individuals in a laboratory. The grand challenge problems of today and tomorrow will be solved by globally distributed, multidisciplinary teams of researchers using powerful technology.” He added: “The five biggest challenges for the global research community are: computing technology; data, provenance, and visualization; software; organization for multidisciplinary computational science; and education. There is need for a global computational science/cyberinfrastructure ecosystem,” he added.

Katelyn White, a graduate student at the University of California at Santa Cruz, felt the experience was very worthwhile. “I enjoyed the conversations (in particular the birds-of-a-feather) where I talked to people I wouldn't have met if it were not for this opportunity. I made some great connections that I will undoubtedly leverage,” she said. “I will also be able to implement some of the tools I learned in my research.”

Katelyn White of UC/ Santa Cruz (second from right), was among the US attendees. White especially enjoyed networking with other attendees who were also interested in computational fluid dynamics. Image courtesy Summer School

HPC challenges in computational sciences

 “I was especially interested in the presentations about visualization by Sean Ahern of Oak Ridge National Laboratory and Uwe Woessner from the High Performance Computing Center of the University of Stuttgart,” said summer school participant Filippo Spiga, graduate of the University of Milano-Bicocca. “From a business and communication standpoint, I believe that if you want to effectively communicate your research to non-technical people, it helps to show them what the numbers mean,” he added. “Visualization helps to support pure numbers and promote a deeper understanding of the science.”  

“For me the biggest problem is that I have more ideas than I have time to implement!” said Jake Searcy, University of Oregon Ph.D. student. “I study high energy physics and work on the ATLAS experiment at the LHC. One thing that I learned at the Summer School is how to add our database files (ROOT files) to VisIt (a visualization tool developed by Lawrence Livermore Laboratory) so I can try different visualization methods for multivariate data. Given some free time I would really like to see if a GPU would speed up our event simulation, or see if we could do a better job indexing our large datasets. This experience gave me a lot to think about,” he added.

In addition to learning how to tackle a broad spectrum of HPC challenges in disciplines such as material sciences, quantum chromo dynamics, plasma physics, life sciences, astro sciences and climate research, students learned about HPC resources that are available through DEISA, PRACE and TeraGrid. Additional presentations addressed a range of related topics, such as parallel programming and models, algorithmic approaches and libraries, performance analysis and profiling, and data intensive computing and visualization. The Summer School closed with a lively discussion about ways to facilitate future trans-continental collaboration. 

—Elizabeth Leake, TeraGrid External Relations. More information is available at the Summer School website.