Search giant’s new software could help future powerful computers unlock chemistry’s secrets
Nov 10, 2017 | Atlanta, GA
Quantum computing has the potential to reboot everything that scientists know about present-day computing. The use of atoms and molecules to crunch data will mean faster, cheaper, and more powerful computers than ever before.
Unfortunately, practical quantum computers aren’t a reality yet. But when they are, chemistry may be the first discipline to take advantage of their power. And it’s a good bet that software from a team of Georgia Tech chemistry researchers will help make that happen.
Google is choosing Psi4 as a plug-in for OpenFermion, Google’s recently launched and free open-source chemistry package for quantum computers. Psi4 is a suite of quantum chemistry programs written by a team led by David Sherrill, a computational chemist and professor in the School of Chemistry and Biochemistry. The Google product takes the quantum chemistry information in Psi4 and translates it to run on a quantum computer.
“It’s always nice to have a product that people appreciate,” Sherrill says. “It gives you validation.” Sherrill cites the GitHub page for Open Fermion as an example. GitHub is a popular software developer’s platform, and it lists both Psi4 and a competing software program. “The description next to ours says in parentheses, ‘recommended,’” he adds with a laugh.
The promise of quantum chemistry
Google, Microsoft, IBM, and Intel are working on quantum computing projects because they recognize its potential, Sherrill says. Google reached out to Sherrill’s team in October 2016 and asked its members to modify Psi4 so they could plug it into OpenFermion.
Microsoft, which is making a competing product, is also using Psi4. “We’re players on either side,” Sherrill says.
Several research teams at Tech are already applying quantum computing methods to cybersecurity and data analysis. But chemistry could be, as Chemical and Engineering News recently put it, “quantum computing’s killer app.”
“There is good mathematical evidence that a quantum computer with a few hundred qubits would be able to solve chemical and materials science questions that are beyond the reach of current supercomputers,” says Kenneth Brown, associate professor in the School of Chemistry and Biochemistry. Brown is former chairman of the Division of Quantum Information of the American Physical Society. (Brown has accepted a position at Duke University and will be leaving Georgia Tech in January 2018.)
Sherrill says applications include rational drug design, which is based on how a drug interacts with its target; crystal engineering; energy conversion and energy storage materials; and organic electronics.
A brief history of quantum computing
The computers used by researchers like Sherrill and Brown are some of the fastest machines available. Today’s microprocessors can run mathematical operations at a billion times per second. Yet computers still rely on transistors, silicon-based microprocessors, and bits of data labeled as 1’s and 0’s.
In quantum computing – based on theories first explored by physicists Paul Benioff and Richard Feynman in the early 1980s –units of data function on the subatomic level. In this mode, they can develop an identity crisis. That’s good, because each data unit, now called a qubit, can effectively be both 1 and 0 at the same time, until a measurement is made. That means qubits can do calculations much faster, enabling more accurate simulations of larger, more complex molecules than ever before.
“You could do so many calculations. You could explore different kinds of molecules and see what their properties are,” Sherrill says. “The calculations are so expensive right now, but on a quantum computer, the calculations would be so cheap.”
Georgia Tech is carving out a special place in the quantum computing research realm. “Tech has a history of excellent work in experimental quantum computing,” Brown says. He’s had several collaborative grants with the Quantum Systems group at the Georgia Tech Research Institute, which is working on quantum computer architecture. He also recently organized a conference for the Center for Research in Novel Computer Hierarchies.
A chemist who codes
As an undergraduate at the Massachusetts Institute of Technology (MIT) in the late 1980s, Sherrill tried various subdisciplines of chemistry and found none that excited him.
Yet when he realized that “you could have a job as a chemist writing software, I thought this was the greatest thing I had ever discovered,” he says. “I had a knack for computer programming. I loved that.”
Sherrill started working on the precursor of the Psi4 software when he was a University of Georgia graduate student in the 1990s. He continued writing it when he joined Tech, where he has a joint appointment in the School of Computational Science and Engineering in the College of Computing.
His Psi4 focus for the past decade has been making the software package easier to use, while adding features like databases. “That’s why I think it became so popular,” he says. “When we added automation for work flows, a lot of power users got very excited about that.”
Many of those users work at major pharmaceutical companies and at biotech startups. Sherrill says OpenEye Scientific Software in New Mexico uses Psi4 to improve the efficiency of drug discovery techniques. His team recently published a paper on drug–protein bindings in collaboration with Bristol Myers-Squibb.
Many academics and undergraduates use Psi4 for both teaching and research because it’s free, he says. The power of cheap computing is key to the promise of quantum computing.
"We recently published a paper that involved more than a million quantum chemistry computations.” Sherrill says. “My group and I had to have some serious talks about how we were going to run so many computations; they took many months. On a quantum computer, these computations might take only a few days."