Ultrafast supercomputers that operate at speeds 100 times faster than current systems may now be one step closer to reality, thanks to researchers at IBM working on a DARPA-funded program. They transmitted massive amounts of data with unprecedented low power consumption, increasing the speed by 66 percent and shattering the previous power efficiency record by one-half. Scientists predict that the supercomputers of the future – exascale computers – will enable them to model the global climate, run molecular-level simulations of entire cells, design nanostructures and much more. “We envision machines reaching the exascale mark around 2020, but a great deal of research must be done to make this possible,” said Jonathan E. Proesel, a research staff member at the IBM T.J. Watson Research Center. Optical link test chips, including transmitter circuits, laser diodes, photodiodes and receiver circuits. First they must find a way to quickly move large amounts of data within the supercomputer while simultaneously keeping power consumption in check. Proesel and colleagues created an optical communication link operating at 25 Gb/s using 24 mW of total wall-plug power, or 1 pJ per bit. The link combines circuits in IBM’s 32-nm silicon-on-insulator CMOS technology with advanced VCSELs and photodetectors fabricated by Sumitomo Electric Device Innovations USA (formerly Emcore). “Compared to our previous work, we have increased the speed by 66 percent while cutting the power in half,” Proesel said. “We’re continuing the push for lower power and higher speed in optical communications. There will always be demand to move more data with less energy, and that’s what we’re working toward.” Proesel presented the work, titled “35-Gb/s VCSEL-Based Optical Link using 32-nm SOI CMOS Circuits,” at OFC/NFOEC 2013.