Quantum computers might someday make it possible to run simulations that are far too complex for conventional computers, enabling them for example to precisely model chemical reactions or the movement of electrons in materials, yielding better products from drugs to fertilizers to solar cells. Yet at the current pace of development, quantum computers powerful enough for these simulations may still be many years away.
With new funding from the National Science Foundation, an academic-industry collaboration led by The University of Texas at Austin aims to close the gap between current quantum hardware and these ambitious simulations. They are developing a new suite of quantum algorithm methods and software tools that more efficiently use qubits — the basic units of information in a quantum computer, analogous to bits in a conventional computer. They plan to test their approach on trapped ion quantum computing systems being developed at Honeywell Quantum Solutions.
"The most exciting thing is that we get to not just develop theoretical ideas in a vacuum, but we get to try them on real hardware and work with the people building it," said Drew Potter, principal investigator and assistant professor of physics at UT Austin.
The NSF grant is approximately $1 Million and is made through the Convergence Accelerator program, a new initiative designed to accelerate the transition of basic research and discovery into practical applications that address wide-scale societal challenges. The NSF has selected 29 teams addressing two transformative research areas: artificial intelligence and quantum technology.
Each team begins a nine-month phase one project developing their concept, participating in innovation curriculum and developing an initial prototype. At the end of phase one, each team participates in a pitch competition and a proposal evaluation. Selected teams from phase one will proceed to phase two, with potential funding up to $5 Million for two years.
In addition to Potter, the four co-PIs are Garnet Chan at the California Institute of Technology; Michael Zaletel at the University of California at Berkeley; and David Hayes and Michael Foss-Feig at Honeywell Quantum Solutions.
The researchers will also use Lonestar5 — a petascale, high performance computing system operated by the Texas Advanced Computing Center for use by academic researchers in Austin and across Texas — to model how the new algorithms will behave on Honeywell's quantum computer.
Their approach, called mid-circuit measurement and reset (MCMR), measures qubits in the middle of a model run and then resets them and reuses them, effectively expanding the number of qubits available in the system.
"It's getting around having a very limited quantum memory to work with," Potter said. "It's a way to simulate systems that are too big to fit directly on your quantum memory all at once."
Comments