Hardware

  • No Comments

MEGA.mini: An Energy-Efficient NPU Leveraging a Novel Big/Little Core With Hybrid Input Activation for Generative AI Acceleration

MEGA.mini: An Energy-Efficient NPU Leveraging a Novel Big/Little Core With Hybrid Input Activation for Generative AI Acceleration 150 150
Author(s): Donghyeon Han, Anantha P. Chandrakasan

Abstract:

This article presents a processor for the acceleration of generative AI (GenAI) based on a novel heterogeneous core architecture called MEGA.mini. The processor introduces three algorithmic features: 1) fixed-point (FXP) and floating-point (FP) hybrid input activation (IA) representation; 2) a delayed-statistics-based normalization (NORM); and 3) conditional polynomial-based nonlinear activation (NLA) approximation. These …

Published in: IEEE Journal of Solid-State Circuits
Page(s): 1 – 14
Year of Publication: 2025
Electronic ISSN: 1558-173X
DOI: 10.1109/JSSC.2025.3626894
Publisher: IEEE

View on IEEE Xplore

  • No Comments

A Calibration-Free Pipelined-SAR ADC With Cross-Stage Gain-Mismatch Error Shaping and Inherent Noise Shaping

A Calibration-Free Pipelined-SAR ADC With Cross-Stage Gain-Mismatch Error Shaping and Inherent Noise Shaping 150 150
Author(s): Jihang Gao, Yaohui Luan, Siyuan Ye, Jiajia Cui, Xinhang Xu, Linxiao Shen

Abstract:

This article presents a calibration-free pipelined-successive-approximation-register (SAR) analog-to-digital converter (ADC) based on the proposed cross-stage gain-mismatch-error shaping (CS-GMES) mechanism. The CS-GMES is realized by including the entire 2nd stage into MES operation to unify the gain error and the 2nd-stage mismatch error. A feedback capacitor provides cross-stage connection and mismatch …

Published in: IEEE Journal of Solid-State Circuits
Page(s): 1 – 13
Year of Publication: 2025
Electronic ISSN: 1558-173X
DOI: 10.1109/JSSC.2025.3624412
Publisher: IEEE

View on IEEE Xplore

  • No Comments

Space-Mate: A 303.5-mW Real-Time Sparse Mixture-of-Experts-Based NeRF-SLAM Processor for Mobile Spatial Computing

Space-Mate: A 303.5-mW Real-Time Sparse Mixture-of-Experts-Based NeRF-SLAM Processor for Mobile Spatial Computing 150 150
Author(s): Gwangtae Park, Seokchan Song, Haoyang Sang, Dongseok Im, Donghyeon Han, Sangyeob Kim, Hongseok Lee, Hoi-Jun Yoo

Abstract:

Simultaneous localization and mapping (SLAM) provides crucial ego-pose information and 3-D maps of the user environment, which are fundamental to emerging mobile spatial computing devices. Dense 3-D mapping and accurate pose estimation are particularly necessary for applications like augmented reality (AR) and autonomous navigation. However, existing SLAM processors are typically …

Published in: IEEE Journal of Solid-State Circuits
Page(s): 1 – 13
Year of Publication: 2025
Electronic ISSN: 1558-173X
DOI: 10.1109/JSSC.2025.3610931
Publisher: IEEE

View on IEEE Xplore

  • No Comments

A Microscaling Multi-Mode Gain-Cell Computing-in-Memory Macro for Advanced AI Edge Device

A Microscaling Multi-Mode Gain-Cell Computing-in-Memory Macro for Advanced AI Edge Device 150 150
Author(s): Jen-Chun Tien, Ping-Chun Wu, Win-San Khwa, Ashwin Sanjay Lele, Jian-Wei Su, Chiao-Yen Cheng, Jun-Ming Hsu, Yu-Chen Chen, Le-Jung Hsieh, Jyun-Cheng Bai, Yu-Sheng Kao, Tsung-Han Lou, Jui-Jen Wu, Chung-Chuan Lo, Ren-Shuo Liu, Chih-Cheng Hsieh, Kea-Tiong Tang, Meng-Fan Chang

Abstract:

The microscaling (MX) format is an emerging data representation that quantizes high-bitwidth floating-point (FP) values into low-bitwidth FP-like values with a shared-scale (SS) exponent. When implemented with computing-in-memory (CIM), MX allows an attractive tradeoff between accuracy and hardware efficiency for specific neural network (NN) workloads. This work presents the first …

Published in: IEEE Journal of Solid-State Circuits
Page(s): 1 – 14
Year of Publication: 2025
Electronic ISSN: 1558-173X
DOI: 10.1109/JSSC.2025.3617944
Publisher: IEEE

View on IEEE Xplore

  • No Comments

1.58-b FeFET-Based Ternary Neural Networks: Achieving Robust Compute-In-Memory With Weight-Input Transformations

1.58-b FeFET-Based Ternary Neural Networks: Achieving Robust Compute-In-Memory With Weight-Input Transformations 150 150
Author(s): Imtiaz Ahmed, Akul Malhotra, Revanth Koduru, Sumeet Kumar Gupta

Abstract:

Ternary weight neural networks (TWNs), with weights quantized to three states (−1, 0, and 1), have emerged as promising solutions for resource-constrained edge artificial intelligence (AI) platforms due to their high energy efficiency with acceptable inference accuracy. Further energy savings can be achieved with TWN accelerators utilizing techniques such as compute-in-memory (CiM) and …

Published in: IEEE Journal on Exploratory Solid-State Computational Devices and Circuits
Page(s): 157 – 165
Year of Publication: 2025
Electronic ISSN: 2329-9231
DOI: 10.1109/JXCDC.2025.3621160
Publisher: IEEE

View on IEEE Xplore

  • No Comments

Opal: A 16-nm Coarse-Grained Reconfigurable Array SoC for Full Sparse Machine Learning Applications

Opal: A 16-nm Coarse-Grained Reconfigurable Array SoC for Full Sparse Machine Learning Applications 150 150
Author(s): Po-Han Chen, Bo Wun Cheng, Michael Oduoza, Zhouhua Xie, Rupert Lu, Sai Gautham Ravipati, Kalhan Koul, Alex Carsello, Yuchen Mei, Mark Horowitz, Priyanka Raina

Abstract:

Sparsity has recently attracted increased attention in the machine learning (ML) community due to its potential to improve performance and energy efficiency by eliminating ineffectual computations. As ML models evolve rapidly, reconfigurable architectures, such as coarse-grained reconfigurable arrays (CGRAs), are being explored to adapt to and accelerate emerging models. Previous …

Published in: IEEE Solid-State Circuits Letters
Page(s): 293 – 296
Year of Publication: 2025
Electronic ISSN: 2573-9603
DOI: 10.1109/LSSC.2025.3613245
Publisher: IEEE

View on IEEE Xplore

  • No Comments

Energy-Efficient Reconfigurable XGBoost Inference Accelerator With Modular Unit Trees via Selective Node Execution and Data Movement

Energy-Efficient Reconfigurable XGBoost Inference Accelerator With Modular Unit Trees via Selective Node Execution and Data Movement 150 150
Author(s): Chang Eun Song, Yidong Li, Amardeep Ramnani, Pulkit Agrawal, Purvi Agrawal, Sung-Joon Jang, Sang-Seol Lee, Tajana Rosing, Mingu Kang

Abstract:

The extreme gradient boosting (XGBoost) has emerged as a powerful AI algorithm, achieving high accuracy and winning multiple Kaggle competitions in various tasks including medical diagnosis, recommendation systems, and autonomous driving. It has great potential for running on edge devices due to its binary tree-based simple computing kernel, offering unique …

Published in: IEEE Journal of Solid-State Circuits
Page(s): 1 – 13
Year of Publication: 2025
Electronic ISSN: 1558-173X
DOI: 10.1109/JSSC.2025.3607896
Publisher: IEEE

View on IEEE Xplore

  • No Comments

ROZK: An Energy-Efficient DNN Accelerator Based on Reconfigurable NoC and Local Zero-Skipping

ROZK: An Energy-Efficient DNN Accelerator Based on Reconfigurable NoC and Local Zero-Skipping 150 150
Author(s): Heetak Kim, Yunpyo Hong, Byungsoo Kim, Jongsun Park

Abstract:

Zero-skipping is a famous technique to improve the energy efficiency of deep neural network (DNN) accelerators. When the zero-skipping is realized with encoded data using lossless compression, irregular and unpredictable size of data due to inconsistent compression rate incurs several design issues including: 1) load imbalance from irregularity of data stored …

Published in: IEEE Journal of Solid-State Circuits
Page(s): 1 – 12
Year of Publication: 2025
Electronic ISSN: 1558-173X
DOI: 10.1109/JSSC.2025.3604814
Publisher: IEEE

View on IEEE Xplore

  • No Comments

FIMA: A Scalable Ferroelectric Compute-in-Memory Annealer for Accelerating Boolean Satisfiability

FIMA: A Scalable Ferroelectric Compute-in-Memory Annealer for Accelerating Boolean Satisfiability 150 150
Author(s): Mohammad Khairul Bashar, T. H. Pantha, Z. Li, M. Farasat, S. Datta, V. Narayanan, S. Dutta, N. Shukla

Abstract:

In-memory compute kernels present a promising approach for addressing data-centric workloads. However, their scalability—particularly for computationally intensive tasks solving combinatorial optimization problems such as Boolean satisfiability (SAT), which are inherently difficult to decompose—remains a significant challenge. In this work, we propose a ferroelectric nonvolatile memory (NVM)-based compute-in-memory …

Published in: IEEE Journal on Exploratory Solid-State Computational Devices and Circuits
Page(s): 81 – 89
Year of Publication: 2025
Electronic ISSN: 2329-9231
DOI: 10.1109/JXCDC.2025.3603942
Publisher: IEEE

View on IEEE Xplore

Search this website