IEEE Solid-State Circuits Letters

EMO-CIM: An Input/Stationary-Data Similarity-Aware Computing-In-Memory Design for Variable Vector-Wise Computation in Edge Multi-Operator AI Acceleration

EMO-CIM: An Input/Stationary-Data Similarity-Aware Computing-In-Memory Design for Variable Vector-Wise Computation in Edge Multi-Operator AI Acceleration 150 150

Abstract:

We propose an edge multi-operator computing-in-memory design (EMO-CIM) that supports variable vector-wise multiply-and-accumulate (MAC) in CNN, Depthwise-Convolution (DW), and Attention operators. It features: (1) A single EMO-CIM bank excels in variable vector-wise MAC for multi-operators; (2) Merging local input-shared compute units with a decode-unit & adder-tree facilitates input/stationary-data similarity-aware computing to improve …

View on IEEE Xplore

A 0.19-PEF Bandwidth/Power Scalable Dynamic Amplifier

A 0.19-PEF Bandwidth/Power Scalable Dynamic Amplifier 150 150

Abstract:

This letter presents an energy-efficient dynamic amplifier. It utilizes source-coupled input boosting and time-domain differential sampling techniques to boost the effective input signal by 4× compared to its floating inverter amplifier (FIA) prototype without noise or power penalties. With discharge-based dynamic biasing, the bandwidth and power of the amplifier can be …

View on IEEE Xplore

1-1.7 GHz Single-Voltage-Controlled Tunable MMIC BPF With Over 45 dB Rejection to 30 GHz Using Tapped Inductor Technique

1-1.7 GHz Single-Voltage-Controlled Tunable MMIC BPF With Over 45 dB Rejection to 30 GHz Using Tapped Inductor Technique 150 150

Abstract:

This paper presents a novel fourth-order single-voltage-controlled (SVC) tunable monolithic microwave integrated circuit (MMIC) bandpass filter (BPF) for tunable filter bank applications. Utilizing the tapped inductor technique (TIT), the proposed filter achieves compact magnetic coupling with reduced loss, avoiding the use of bulky inductor-or transformer-based coupling structures. The filter demonstrates …

View on IEEE Xplore

A 7056-PPI Pixel Circuit with Low-Leakage Structure for Active-Matrix Monochrome Micro-LED Displays

A 7056-PPI Pixel Circuit with Low-Leakage Structure for Active-Matrix Monochrome Micro-LED Displays 150 150

Abstract:

This work presents a 5T2C pixel circuit for active-matrix (AM) micro-displays in near-eye display applications. The circuit supports monochrome micro light-emitting diode (micro-LED) displays with ultra-high resolution of 7056 pixels per inch (PPI). The circuit is designed and fabricated based on medium-voltage (MV) devices from the 55-nm high-voltage (HV) CMOS …

View on IEEE Xplore

Analysis and Design of Power Amplifier Using Parallel-Combined Multisegment Transformer

Analysis and Design of Power Amplifier Using Parallel-Combined Multisegment Transformer 150 150

Abstract:

This letter presents a highly efficient power amplifier (PA) using a parallel-combined vertical multisegment transformer for 5G new radio (NR) applications operating in bands n257 and n258, in a 65-nm bulk CMOS process. A multisegment transformer facilitates a lower provided input impedance than a conventional transformer, enabling the PA to …

View on IEEE Xplore

An Aging-Robust 32-MHz RC Frequency Reference With 0.4-ppm Allan Deviation and ±1550-ppm Inaccuracy From −40 °C to 125 °C After a 1-Point Trim

An Aging-Robust 32-MHz RC Frequency Reference With 0.4-ppm Allan Deviation and ±1550-ppm Inaccuracy From −40 °C to 125 °C After a 1-Point Trim 150 150

Abstract:

This letter presents an aging-robust 32-MHz RC frequency reference based on a frequency-locked-loop (FLL). With a temperature compensation scheme that combines BJTs and aging-robust diffusion resistors, the FLL achieves ±1550-ppm inaccuracy from $-40~^{\circ }$ C to $125~^{\circ }$ C after batch calibration and a low-cost 1-point trim, which increases to ±2350-ppm …

View on IEEE Xplore

A Folded-Differential Switched-Capacitor SRAM CIM Macro With Scalable MAC Sizes for TinyML Inference

A Folded-Differential Switched-Capacitor SRAM CIM Macro With Scalable MAC Sizes for TinyML Inference 150 150

Abstract:

This letter presents a switched-capacitor SRAM compute-in-memory macro optimized for TinyML inference. Key features include: 1) an area-efficient folded-differential multiply-and-accumulate (FD-MAC) scheme to double the signal margin; 2) a closed-loop floating-inverter amplifier (FIA)-based charge accumulation technique for signal-to-noise ratio enhancement and multiply-and-accumulate (MAC) voltage integration; and 3) a sparsity-aware multistep MAC method …

View on IEEE Xplore

A 70-GHz Bandwidth Amplifier With Integrated Differential Bridged T-coil Peaking and Uniform Group Delay

A 70-GHz Bandwidth Amplifier With Integrated Differential Bridged T-coil Peaking and Uniform Group Delay 150 150

Abstract:

A two-stage amplifier in 22-nm FD-SOI CMOS integrates a fully-differential bridged T-coil for the first time. Circuit performance is benchmarked against an identical amplifier topology designed with single-ended T-coils (pseudo-differential) and an unpeaked reference. It realizes 70-GHz bandwidth with $12~\pm ~2$ -ps group delay and >10-dB return loss across 90 GHz. Bandwidth …

View on IEEE Xplore

A 3-D HBI Compliant 1.536 TB/s/mm2 Bandwidth Scalable Attention Accelerator With 22.5-GOPS Throughput High Speed SoftMax for Quantized Transformers in Intel 3

A 3-D HBI Compliant 1.536 TB/s/mm2 Bandwidth Scalable Attention Accelerator With 22.5-GOPS Throughput High Speed SoftMax for Quantized Transformers in Intel 3 150 150

Abstract:

This letter presents a novel hardware accelerator compatible with <3- $\mu $ m pitch 3-D Cu-Cu hybrid bonding interconnect (HBI) technology, particularly designed to efficiently execute multihead attention (MHA) of encoder transformer models. We present an accelerator that addresses performance losses due to low precision models by incorporating specialized hardware optimizations …

View on IEEE Xplore