Supercomputing partnership yields life-changing science, strong bond between Colombian university and Purdue
Juan Carlos Vergara would go two weeks at a time without his personal computer because it was busy grinding numbers for his research modeling earthquakes. His hard drive broke — twice.
Then, Vergara found out about a resource capable of getting his work, which normally took months to finish, done in days or less, while also letting him expand the scale of his seismic engineering problems to cover an area 5 million times larger. All that thanks in part to a partnership between Purdue University and Colombia and, in particular, between Purdue’s research computing unit and Colombia’s Apolo Scientific Computing Center at the Universidad EAFIT in Medellin.
That partnership had yielded Apolo, EAFIT’s first research supercomputer, and a staff to run it for researchers, including Juan David Pineda, Apolo’s technology coordinator, and Mateo Gomez, a high-performance computing analyst.
“Sometimes we would be up until 1 a.m. helping me solve problems,” says Vergara, a doctoral student of applied mechanics at EAFIT. “I saw them as part of my team, fundamental to what I do every day.”
When Purdue partnered with EAFIT it had a lot to offer where accelerating discoveries in science and engineering with supercomputers is concerned. Purdue’s central information technology organization has built and operated nine high-performance computing systems for faculty researchers in as many years, most rated among the world’s top 500 supercomputers. They’ve given Purdue the best research computing resources for use on a single campus in the nation. Hardware from one of those machines, the retired Steele cluster, became the foundation of Apolo.
With Apolo’s installation in 2012, EAFIT’s research capacity ballooned. Once fully functional, the computer was almost constantly running at nearly 100 percent capacity. In 2015, Apolo helped 69 researchers from EAFIT, other universities and local industry complete their research. In 2016, the machine executed what’s equivalent to 129 years of computation.
Research enabled by Apolo has ranged from earthquake science and a groundbreaking examination of the tropical disease leishmaniasis to the most “green” way of cement processing and quantum mechanics, with the supercomputer speeding EAFIT researchers’ time to science.
In the fall of 2016, EAFIT retired Apolo and bought Apolo II. As it goes with technology, Apolo II is a third the size of the original with twice the power. Purdue will be adding to that power by selling EAFIT part of its retiring Carter research supercomputer.
Still, Juan Luis Mejía, rector at Universidad EAFIT, says that in retrospect the decision to buy the first Apolo was almost “irresponsible” considering obstacles like the lack of support in Colombia or the staff required on top of the hardware. There was a lot of trial and error involved in learning how to run its own supercomputing center.
During Apolo’s first years, the team had problems with air flow and cooling, for example. It got so hot one point they had to turn Apolo off immediately. With no “expert” on staff, they enlisted a professor of mechanical engineering, whose specialty was air flow, to design a tailored system to ensure constant cooling.
“The reality here in Colombia is that there’s incentive to invest in the machines but not for the human capital necessary to run them,” says Juan Guillermo Lalinde, director of the Apolo team and professor of informatics at EAFIT.
Decades of isolation due to the violent drug wars of the 1980s and ‘90s took a toll on Colombian universities’ ability to grow. Mejía says EAFIT had been searching for an international partner to help. What it found in Purdue was unexpected.
“Finding an alliance without a hidden agenda, with a true interest in sharing knowledge of technology that would allow us to progress,” Mejía said, “because of all this, I believe that the relationship between our university and Purdue is one of the most valuable and trusting.”
Gerry McCartney, Purdue’s vice president for information technology and chief information officer, says the credit really should go to EAFIT, which was willing to make a leap into high-performance research computing and recognized the avenues it could open.
“They had the academic environment, the infrastructure and the willingness to invest in people,” McCartney says. “We think of them as a partner now, and we expect to deepen that.”
While the hardware is important, the partnership is more about people. If watching the Apolo center at work is a lot like watching Purdue’s research computing operation at work, there’s a reason. From the start, Purdue emphasized the sensibility of its Community Cluster Program partnership with faculty and its centrally managed, shared “condo” model for operating research supercomputers, says Donna Cumberland, executive director of research computing.
“Anybody can buy a machine, but getting people to run it and getting faculty to use it, that’s what we wanted to impart,” Cumberland says.
Lalinde developed a similar dynamic in the Apolo team with a customer service focus and a limit on bureaucracy. The result is a cohesive group of system administrators and students who are curious, hungry and focused on enabling research.
“Every week, we save the world,” says Pineda, the Apolo technology coordinator. “After we fix something, the faculty are smiling because everything works, even if they don’t know why.”
Purdue research computing staff members have traveled to Colombia to help train and to work with EAFIT colleagues, and EAFIT students have participated in Purdue’s Summer Undergraduate Research Fellow (SURF) program working with a variety of supercomputing experts at Purdue.
EAFIT and Purdue have even sent joint teams to student supercomputing competitions in New Orleans and Frankfurt, Germany. Some of the Colombian students on the teams have become key staff members at the Apolo center, which, in turn, is training the next generation of Colombia’s high-performance computing experts.
Gomez, who participated in the SURF program and was on the student teams, brought back techniques such as using Torque, resource management software that helps administrators distribute jobs so as to most efficiently use the computer’s power and space.
“Learning that technique was key to Apolo’s growth,” says Gomez, who later became a high-performance computing analyst at the center. “The SURF internship was so important for that transference of knowledge. I feel privileged to be on this path. For my country, it’s a unique opportunity.”
Cumberland says a story like Gomez’s is the kind of result she hoped for from the SURF program and Purdue’s partnership with EAFIT.
Lalinde has made students an integral part of the Apolo team, much as Purdue research computing employs Purdue students.
“Having that opportunity to see and work with my professors in a way that’s not the normal relationship was enlightening,” Lalinde says, hearkening back to a graduate mathematics seminar he once took, alongside some of the very professors who had taught him. “In Apolo, we’ve been able to replicate the same thing.”
EAFIT is embarking on a journey that could bring an enhanced academic reputation and a significant increase in research funding. Colciencias, Colombia’s version of the National Science Foundation, is partnering with the ministries of education, industry and tourism to enhance the country’s academic profile. Funded by the World Bank, the program, Colombia Científica, is aimed at producing more robust research and more advanced-degree graduates.
Because EAFIT is one of the few universities in Colombia with a supercomputer and a strong partnership with an American Land Grant university, it is poised to receive a chunk of funding and to grow partnerships with local industries.
They’ve already attracted Grupo Nutresa, a Latin American food processing company headquartered in Medellín, and Pilar Cossio, a Colombian HIV researcher working for the Max Planck Institute in Germany.
When Cossio came home to Colombia after studying and working in Italy, the U.S. and Germany the biophysicist figured that one big task she was going to face would be building her own supercomputer and finding someone to run it. But she’s able to conduct her research at the Universidad de Antioquia in Medellín with help from the Apolo Scientific Computing Center at EAFIT.
Cossio’s research combines physics, computational biology and chemistry. She’s studying protein changes at the atomic level on the way to helping design drugs to cure HIV. That requires examining around two million different compounds to see which ones bind the best with a particular protein. High-performance and parallel computing power are vital for her research and she didn’t anticipate finding a ready-made solution in her home country. Then she found Apolo.
“There are only two supercomputers in Colombia for bioinformatics,” Cossio says. “Apolo is the only one that focuses on satisfying scientific needs. It’s important for us in the developing countries to have partnerships with universities that can help us access these crucial scientific tools.”
Last Updated: March 29, 2017