First exposure to the Open Science Grid

BTejerinaDr. Baudilio Tejerina works as a scientific IT consultant for the Department of Earth and Planetary Sciences at Northwestern University. His first exposure to the Open Science Grid (OSG) was this past March at the 2015 OSG All-Hands meeting, held at the Kellogg School of Management’s Allen Center at Northwestern.

Tejerina attended the OSG Connect tutorial led by Dr. Emelie Harstad and came away impressed. He recently submitted his first proposal for a new OSG Connect project he’s calling MEMPhys (Modeling Earth Materials Physics). “It was fantastic,” says Tejerina. “I opened an account right away. I have already written a few scripts, and I’m learning how to use the Condor scheduler. (I have computing experience, so the learning curve isn’t so bad for me.)”

Image courtesy Baudilio Tejerina

Continues Tejerina, “The All-Hands meeting gave me the opportunity to exchange ideas with other attendees. Before then, I didn’t really know what the OSG was or what its purpose was. But I immediately saw potential applications to my work. Networking with others at the workshops brought many offers of help like working with the University of Chicago to help me get started.”

After an undergraduate degree in organic chemistry, a Master’s and PhD in physical chemistry, and a postdoc in physical chemistry and quantum mechanics at Ames Lab (Gordon Group) at Iowa State, Tejerina has been at Northwestern since 2004. He maintains some of Northwestern’s computational infrastructure, and also works with students in the areas of geophysics and geochemistry.

In geophysics, Tejerina develops algorithms to simulate the effects of water waves (tsunamis), and develops models that study the propagation of these waves and their effects. In geochemistry, he studies the composition, properties, and materials of the earth’s mantle (30 miles under the crust), running quantum mechanics simulations.

“In both cases,” says Tejerina, “we have to solve very complicated equations. So we must use computing—we cannot do this by hand. I will use the OSG first for the geochemistry work. We need to decide the parameters, what to study, and write the scripts. Then I’ll investigate how we can use the OSG to study tsunamis, but I won’t do that at first.”

To study the structure of materials 30 miles under the earth’s crust requires huge computational power for a single calculation. “We don’t study merely one single structure, but instead we study millions of structures,” says Tejerina. “We compare that with the information we have from seismology. Then, we need to calculate many structures many times. For that we need many CPUs to perform millions of calculations. Up until now, we have been using small clusters at Argonne National Laboratory. Northwestern also has large resources that we sometimes use.”

Tejerina is drawn to the potential of the OSG to see how they can apply their research into the development of new materials. “We want to explore crystalline structures and how atoms move inside a crystal to form another crystal,” says Tejerina. “We suspect this is what happens in the mantle at high temperature and pressure. In order to do that, we need to create millions of simulations. Each instance should be running in one CPU. The OSG lends itself very well for this kind of problem where each parameter needs its own computation. It’s a very challenging problem to explore the inner structure of a crystal.”

Tejerina uses the example of how carbon manifests as diamond and graphite. One is transformed into the other and vice versa. “We know the extremes of each but we don’t know how they transform—how the atoms move inside. We must simulate all this, so the OSG should prove very helpful.” Tejerina not only wishes to tackle this problem, but also learn about other materials—how to design or synthesize these materials, or design new materials with specific properties for particular applications.


Two known crystal structures of magnesium silicate show the same composition with different structures. Image courtesy Baudilio Tejerina.

Tejerina’s first OSG project proposal, MEMPhys, will focus on clarifying how polymorphic materials (materials that can exist in more than one form or crystal structure) transform into various phases. The project description states: “The potential energy surface (PES) of the material is defined by the crystal parameters and the positions of the atoms in the unit cell. The large number of degrees of freedom of the system makes the exploration and characterization of the PES a challenging computational task.”

It sounds like a great fit for the OSG.

Greg Moore