Name
Abstract | ||
|---|---|---|
Artificial neural networks have become ubiquitous in modern life, which has triggered the emergence of a new class of application specific integrated circuits for their acceleration. ReRAM-based accelerators have gained significant traction due to their ability to leverage in-memory computations. In a crossbar structure, they can perform multiply-and-accumulate operations more efficiently than standard CMOS logic. By virtue of being resistive switches, ReRAM switches can only reliably store one of two states. This is a severe limitation on the range of values in a computational kernel. This paper presents a novel scheme in alleviating the single-bit-per-device restriction by exploiting frequency dependence of v-i plane hysteresis, and assigning kernel information not only to the device conductance but also partially distributing it to the frequency of a time-varying input. We show this approach reduces average power consumption for a single crossbar convolution by up to a factor of ×16 for an unsigned 8-bit input image, where each convolutional process consumes a worst-case of 1.1mW, and reduces area by a factor of ×8, without reducing accuracy to the level of binarized neural networks. This presents a massive saving in computing cost when there are many simultaneous in-situ multiply-and-accumulate processes occurring across different crossbars. |
Year | DOI | Venue |
|---|---|---|
2019 | 10.1109/AICAS.2019.8771550 | 2019 IEEE International Conference on Artificial Intelligence Circuits and Systems (AICAS) |
Keywords | Field | DocType |
accelerator,analog,memristor,neural network,ReRAM | Kernel (linear algebra),Convolution,Computer science,Electronic engineering,Application-specific integrated circuit,CMOS,Acceleration,Artificial neural network,Crossbar switch,Resistive random-access memory | Conference |
ISBN | Citations | PageRank |
978-1-5386-7885-5 | 2 | 0.37 |
References | Authors | |
0 | 6 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Jason K. Eshraghian | 1 | 25 | 5.62 |
| Sung-Mo Steve Kang | 2 | 1198 | 213.14 |
| Seungbum Baek | 3 | 4 | 0.82 |
| Orchard, G. | 4 | 24 | 3.77 |
| Herbert H. C. Iu | 5 | 334 | 60.21 |
| Wen Lei | 6 | 5 | 0.78 |