
Speaker:
Alexander Khitun
Institution:
UC Riverside
Speaker Link:

Date:
Thursday, September 24, 2015
Time:
4:00 pm
Location:
NS2 1201
ABSTRACT:
There is growing interest in novel computational devices able to overcome the limits of the current complimentary-metal-oxide-semiconductor (CMOS) technology and provide further increase of the computational throughput. Spin wave (magnonic) logic devices are one of the alternative approaches aimed to take the advantages of the wave interference at nanometer scale and utilize phase in addition to amplitude for building logic units for parallel data processing. In this talk, I will start with a brief introduction to the different types of spin wave logic devices including Boolean and non-Boolean types of circuits. The main part of the talk is based on the most recent results on Magnonic Holographic Memory (MHM). MHM evolves the general idea of optical approach to applications in the magnetic domain aimed to combine the advantages of magnetic data storage with the unique capabilities for read-in and read-out provided by spin waves. I will present experimental data on the 2-bit MHM built on the double cross YIG structure. This first prototype has demonstrated the possibility of parallel magnetic bits read-out using spin waves. Next, there will be results on the 8-terminal MHM device for pattern recognition. I will also present estimates on the spin wave holographic devices performance, including power consumption and functional throughput. According to the estimates, magnonic holographic devices may provide data processing rates higher than 1×1018 bits/cm2/s while consuming only 0.15mW. Potentially, magnonic holographic devices may provide a higher storage density compare to the optical counterparts due to a shorter wavelength. Technological challenges and fundamental physical limits of this approach will be also discussed.
Host:
Ilya Krivorotov