Abstract:Computing-in-memory (CiM) based on spintronic devices is considered to be one of the most promising solutions to address the von Neumann bottleneck. In this article, a novel CiM platform based on hybrid spintronic/CMOS memory is proposed, which can work in memory mode and computing mode. Instead of integrating complex logic units in memory, minor modifications are made to peripheral circuits to perform calculations. Profiting from the unique properties of hybrid memory and well-modified peripheral circuitry, three different computing schemes are introduced, and the computing results are obtained as typical memory readouts. Complex bulk bitwise logic functions including 16 Boolean logic functions and arithmetic operations can be implemented by combining the proposed schemes and dynamically configuring the input variables. All computing functions can be implemented without nonvolatile bit write operations, and there is no restriction on the location of operands. A 1-bit full adder and a multibit ripple carry adder are designed as instances of flexible utilization of the platform. The hybrid spintronic/CMOS simulations at the 40 nm technology node further demonstrate the functionality and performance of this work. The performance of nonvolatile memory write operations is significantly improved and requirements for device fabrication are reduced by exploiting the toggle spin torques mechanism. The resistance margins of our design are twice greater compared to the previous parallel-connected schemes.

An 8-Bit RRAM Based Multiplier for Hybrid Memory Computing

Nonlinear Quantization for In-Sram Multi-Bit MAC Design

A Low-Power In-Memory Multiplication and Accumulation Array with Modified Radix-4 Input and Canonical Signed Digit Weights

In-memory Multiplication Engine with SOT-MRAM Based Stochastic Computing

RRAM-Based Single Device for Vector Multiplication and Multibit Storage with Ultrahigh Area Efficiency

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

In-Memory Wallace Tree Multipliers Based on Majority Gates Within Voltage-Gated SOT-MRAM Crossbar Arrays

A 28-Nm Compute SRAM with Bit-Serial Logic/Arithmetic Operations for Programmable In-Memory Vector Computing

An RRAM-Based Digital Computing-in-Memory Macro with Dynamic Voltage Sense Amplifier and Sparse-Aware Approximate Adder Tree

A Novel Computing-in-Memory Platform Based on Hybrid Spintronic/CMOS Memory

A Computing In-Memory Multibit Multiplication Based on Decoupling and In-Array Storing

A Hybrid Computing-In-Memory Architecture by Monolithic 3D Integration of BEOL CNT/IGZO-based CFET Logic and Analog RRAM

A 28nm 8928Kb/mm 2 -Weight-Density Hybrid SRAM/ROM Compute-in-Memory Architecture Reducing >95% Weight Loading from DRAM.

Multifunctional Computing-in-memory SRAM Cells Based on Two-Surface-channel MoS2 Transistors

A 65nm 8b-Activation 8b-Weight SRAM-Based Charge-Domain Computing-in-Memory Macro Using A Fully-Parallel Analog Adder Network and A Single-ADC Interface

A 28 nm 81 Kb 5995.3 TOPS/W 4T2R ReRAM Computing-in-Memory Accelerator With Voltage-to-Time-to-Digital Based Output

An 8.8 TFLOPS/W Floating-Point RRAM-Based Compute-in-Memory Macro Using Low Latency Triangle-Style Mantissa Multiplication

A 28nm Hybrid 2T1R RRAM Computing-in-Memory Macro for Energy-efficient AI Edge Inference

Energy-Efficient Hybrid-RAM with Hybrid Bit-Serial Based VMM Support

C 2 IM: A Compact Computing-In-Memory Unit of 10 Transistors with Standard 6T SRAM

An Ultra-Low-Power Analog Multiplier–Divider Compatible with Digital Code for RRAM-Based Computing-in-Memory Macros