Abstract:Artificial intelligence has made remarkable advancements in recent years, leading to the development of algorithms and models capable of handling ever-increasing amounts of data. The computational demands of these algorithms necessitate circuit and architecture designs that go beyond the von-Neumann paradigm. Content addressable memories (CAMs), which implement parallel associative search functionality within memory blocks to overcome the memory wall bottleneck, have proven to be effective for data-intensive tasks. While current CAM designs have achieved higher storage density and energy efficiency than their CMOS-based counterparts by leveraging emerging non-volatile memories (NVM), most of these implementations are limited to binary storage cells. In this work, we propose SEE-MCAM, scalable and compact multi-bit CAM (MCAM) designs that utilize the three-terminal ferroelectric FET (FeFET) as the proxy. By exploiting the multi-level-cell characteristics of FeFETs, our proposed SEE-MCAM designs enable multi-bit associative search functions and achieve better energy efficiency and performance than existing FeFET-based CAM designs. We validated the functionality of our proposed designs by achieving 3 bits per cell CAM functionality, resulting in 3x improvement in storage density. The area per bit of the proposed SEE-MCAM cell is 8% of the conventional CMOS CAM. We thoroughly investigated the scalability and robustness of the proposed design. Evaluation results suggest that the proposed 2FeFET-1T SEE-MCAM achieves 9.8x more energy efficiency and 1.6x less search latency compared to the CMOS CAM, respectively. When compared to existing MCAM designs, the proposed SEE-MCAM can achieve 8.7x and 4.9x more energy efficiency than ReRAM-based and FeFET-based MCAMs, respectively. Benchmarking results show that our approach provides up to 3 orders of magnitude improvement in speedup and energy efficiency over a GPU implementation in accelerating a novel quantized hyperdimensional computing (HDC) application.

Nonvolatile and Energy-Efficient FeFET-Based Multiplier for Energy-Harvesting Devices.

In-Memory Multi-Bit Multiplication and Accumulation (MAC) Using FeFET for Energy Efficient IoT

An Energy Efficient Floating Point Computing Infrastructure Embedding Ferroelectric Field Effect Transistor Based Ternary Content Addressable Memories

CSA-CiM: Enhancing Multi-Functional Computing-in-Memory with Configurable Sense Amplifiers

FeFET-Based In-Memory Hyperdimensional Encoding Design

LMM: A Fixed-Point Linear Mapping Based Approximate Multiplier for IoT

Energy Efficient Data Search Design and Optimization Based on a Compact Ferroelectric FET Content Addressable Memory

A FeFET-based Time-Domain Associative Memory for Multi-bit Similarity Computation

Nonvolatile processor optimization for ambient energy harvesting scenarios

Low Power and Temperature-Resilient Compute-In-Memory Based on Subthreshold-FeFET

A High-Performance and Energy-Efficient Ternary Multiplier Using CNTFETs

An 8-Bit Radix-4 Non-Volatile Parallel Multiplier

Design of low-power and high-speed multiplier

SEE-MCAM: Scalable Multi-bit FeFET Content Addressable Memories for Energy Efficient Associative Search

Exploiting Ferroelectric FETs for Low-Power Non-Volatile Logic-in-memory Circuits

Lowering Area Overheads for FeFET-Based Energy-Efficient Nonvolatile Flip-Flops

Energy Efficient Dual Designs of FeFET-Based Analog In-Memory Computing with Inherent Shift-Add Capability

Reconfigurable Frequency Multipliers Based on Complementary Ferroelectric Transistors

Energy efficient approximate multipliers compatible with error-tolerant application

Tri-state GNRFET-based fast and energy-efficient ternary multiplier

Low-Power Pass-Transistor Logic-Based Full Adder and 8-Bit Multiplier