Erlangen 2018 – wissenschaftliches Programm
Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
Q: Fachverband Quantenoptik und Photonik
Q 42: Quantum Optics and Photonics II
Q 42.7: Vortrag
Mittwoch, 7. März 2018, 15:45–16:00, K 0.016
Calculating with light - an all-optical abacus using phase-change materials — •Johannes Feldmann1, Matthias Stegmaier1, Nico Gruhler1, Carlos Rios2, Harish Bhaskaran2, David Wright3, and Wolfram Pernice1 — 1Physikalisches Institut, WWU Münster, Germany — 2Department of Materials, University of Oxford, United Kingdom — 3Department of Engineering, University of Exeter, United Kingdom
Since conventional data processing based on the von Neumann architecture faces more and more difficulties in increasing the processing speeds due to limitations in electrical data transfer and heat dissipation, substantial research is dedicated to developing new and unconventional processing schemes as e.g. brain inspired computing using neural networks or accumulation based computing.
A recently emerging technology in this field is based on memristor devices that combine memory and resistor in one element and help to circumvent the need to shuffle data between processor and memory but instead calculate and store it in the same physical location. In our work we transfer these ideas to a nanophotonic platform using phase-change materials embedded on top of integrated waveguides to enable fast and low energy arithmetic in analogy to an abacus. Employing single picosecond optical pulses with energies below 20 pJ our on-chip devices are capable of addition, subtraction, multiplication and division directly in base ten in speeds approaching the GHz-range. Using a waveguide-crossing array we present a scalable nanophotonic framework providing first steps towards non-von Neumann computation.