Liquid Silicon: A Nonvolatile Fully Programmable Processing-In-Memory Processor with Monolithically Integrated ReRAM for Big Data/Machine Learning Applications (invited)

Liquid Silicon: A Nonvolatile Fully Programmable Processing-In-Memory Processor with Monolithically Integrated ReRAM for Big Data/Machine Learning Applications (invited). ZhaS, Y., Nowak, E., & Li, J. IEEE Journal of Solid-State Circuits (JSSC), 55(4):908–919, 2020.

Paper doi abstract bibtex 23 downloads

The slowdown of the CMOS technology scaling, and the trade-off between efficiency and flexibility have fueled the exploration into novel architectures with emerging post-CMOS technology [e.g., resistive-RAM (RRAM)]. In this article, a nonvolatile fully programmable processing-in-memory (PIM) processor named Liquid Silicon is demonstrated, which combines the superior programmability of general-purpose computing devices [e.g., field-programmable gate array (FPGA)] and the high efficiency of domain-specific accelerators. Besides the general computing applications, Liquid Silicon is particularly well suited for artificial intelligence (AI)/machine learning and big data applications, which not only poses high computational/memory demand but also evolves rapidly. To fabricate the Liquid Silicon chip, the HfO 2 RRAM is monolithically integrated on top of the commercial 130 nm CMOS. Our measurement confirms that Liquid Silicon chip can operate reliably at a low voltage of 650 mV. It achieves 60.9 TOPS/W in performing neural network (NN) inferences, and 480 GOPS/W in performing content-based similarity search (a key big data application) at a nominal voltage supply of 1.2 V, showing 3x and 100x improvement over the state-of-the-art domain-specific CMOS-/RRAM-based accelerators. In addition, it outperforms the latest nonvolatile FPGA in energy efficiency by 3x in general computing applications.

@ARTICLE{zha2020jssc, 
author={Zha<sup>S</sup>, Yue and Nowak, Etienne and Li, Jing}, 
journal={IEEE Journal of Solid-State Circuits (<strong>JSSC</strong>)}, 
title={{Liquid Silicon}: A Nonvolatile Fully Programmable Processing-In-Memory Processor with Monolithically Integrated {ReRAM} for {Big Data/Machine Learning} Applications (<strong>invited</strong>)}, 
abstract = {The slowdown of the CMOS technology scaling, and the trade-off between efficiency and flexibility have fueled the exploration into novel architectures with emerging post-CMOS technology [e.g., resistive-RAM (RRAM)]. In this article, a nonvolatile fully programmable processing-in-memory (PIM) processor named Liquid Silicon is demonstrated, which combines the superior programmability of general-purpose computing devices [e.g., field-programmable gate array (FPGA)] and the high efficiency of domain-specific accelerators. Besides the general computing applications, Liquid Silicon is particularly well suited for artificial intelligence (AI)/machine learning and big data applications, which not only poses high computational/memory demand but also evolves rapidly. To fabricate the Liquid Silicon chip, the HfO 2 RRAM is monolithically integrated on top of the commercial 130 nm CMOS. Our measurement confirms that Liquid Silicon chip can operate reliably at a low voltage of 650 mV. It achieves 60.9 TOPS/W in performing neural network (NN) inferences, and 480 GOPS/W in performing content-based similarity search (a key big data application) at a nominal voltage supply of 1.2 V, showing 3x and 100x improvement over the state-of-the-art domain-specific CMOS-/RRAM-based accelerators. In addition, it outperforms the latest nonvolatile FPGA in energy efficiency by 3x in general computing applications.},
 year = {2020},
 volume = {55},
 number = {4},
 pages = {908--919},
 url = {https://doi.org/10.1109/JSSC.2019.2963005},
 doi = {10.1109/JSSC.2019.2963005}, 
 keywords = {journal, sj, Liquid Silicon}
}

Downloads: 23