Last April, Toshiba announced a new, data-crunching technique: the Simulated Bifurcation Algorithm. It uses combinatorial optimization to get the best solutions out of large swaths of data patterns. The algorithm surpasses the speed of a supercomputer without the need for a massive machine or intricate cooling processes, according to a new Bloomberg report. Toshiba hopes to find commercial uses for the algorithm in finance, drug development, and shipping route configurations.
The race to quantum computing — While the likes of Google and IBM are creating supercomputers the size of two basketball courts, Toshiba has focused on supercharging algorithms. Google’s quantum computer performed a 10,000-year task in a few minutes last October, but the qubits used in this kind of computing need to be kept at absolute zero. And even when these machines work well, they’re not immune to errors.
The speedy Simulated Bifurcation Algorithm doesn't quite match the overall capabilities of a supercomputer, but to get this close running on a desktop PC is an impressive feat. Toshiba is getting around the issues of quantum computing and already seeking to commercialize its algorithm, particularly in finance. In October, the company was able to make profitable currency trades that take advantage of price differences in 30 microseconds, piquing the interest of some in the financial sector. Others don’t believe that the level of speed is necessary based on how the markets work.
“Right now, what you can do with [quantum computing] is still hypothetical,” said Kazuyuki Takeda, a general manager at Mizuho-DL Financial Technology Co. to Bloomberg. “It will take quite a bit of time before we have practical uses of quantum computing. At least 10 years or so.”