Memory management

  • No Comments

Adelia: A 4-nm LLM Processing Unit With Streamlined Dataflow and Dual-Mode Parallelism for Maximizing Hardware Efficiency

Adelia: A 4-nm LLM Processing Unit With Streamlined Dataflow and Dual-Mode Parallelism for Maximizing Hardware Efficiency 150 150
Author(s): Sukbin Lim, Jung-Hoon Kim, Seungjae Moon, Junseo Cha, Dongjin Seo, Jongho Kim, Hunjong Lee, Jinwon Lee, Joo-Young Kim

Abstract:

The proliferation of large language models (LLMs) as cross-domain foundation models is fueled by aggressive scaling in both parameter counts and inference-time computation. The emergence of sophisticated reasoning models further accelerates this trend, demanding longer context windows and escalating the computational and memory burdens of inference. A fundamental challenge arises …

Published in: IEEE Journal of Solid-State Circuits
Page(s): 1513 – 1525
Year of Publication: 2026
Electronic ISSN: 1558-173X
DOI: 10.1109/JSSC.2026.3663603
Publisher: IEEE

View on IEEE Xplore

  • No Comments

A 7.5-μW 35-Keyword End-to-End Keyword Spotting System With Random Augmented On-Chip Training

A 7.5-μW 35-Keyword End-to-End Keyword Spotting System With Random Augmented On-Chip Training 150 150
Author(s): Hyuk-Jin Lee, Kyunghoon Pyo, Taekwang Jang, Mingoo Seok, SeongHwan Cho

Abstract:

Fully integrated keyword spotting (KWS) systems designed for low-power operation face two major challenges. First, increasing the number of supported keywords significantly raises system complexity and power consumption. Second, most existing systems are not personalized to individual users, as they are trained on data from native English speakers, leading to …

Published in: IEEE Journal of Solid-State Circuits
Page(s): 1 – 10
Year of Publication: 2026
Electronic ISSN: 1558-173X
DOI: 10.1109/JSSC.2026.3660634
Publisher: IEEE

View on IEEE Xplore

  • No Comments

An Electrophysiology-Optogenetics Closed-Loop Bi-Directional Neural Interface for Sleep Regulation With 0.2-μJ/class Multiplexer-Based Neural Network

An Electrophysiology-Optogenetics Closed-Loop Bi-Directional Neural Interface for Sleep Regulation With 0.2-μJ/class Multiplexer-Based Neural Network 150 150
Author(s): Chao Zhang, Fenyan Zhang, Dawid Sheng, Zhixiong Ma, Yongxiang Guo, Chao Sun, Yuwei Zhang, Wenxin Zhao, Xing Sheng, Tongfei A. Wang, Milin Zhang

Abstract:

This work proposed a multiplexer-based neural network (MUXnet), a multiplexer-based, multiplier-free neural network (NN) structure applicable to the implementation of all inner product-based NN layers. An on-chip MUXnet-based neural signal processing unit (NSPU) was designed, achieving a state-of-the-art accuracy of 82.4% on a public human sleep staging dataset, with the lowest …

Published in: IEEE Journal of Solid-State Circuits
Page(s): 1 – 13
Year of Publication: 2026
Electronic ISSN: 1558-173X
DOI: 10.1109/JSSC.2025.3639446
Publisher: IEEE

View on IEEE Xplore

  • No Comments

HUTAO: A Reconfigurable Homomorphic Processing UniT With Cache-Aware Operation Scheduling

HUTAO: A Reconfigurable Homomorphic Processing UniT With Cache-Aware Operation Scheduling 150 150
Author(s): Sijia Lu, Wenping Zhu, Bohan Yang, Jiajun Yang, Tongwei Dai, Chen Chen, Xiangdong Han, Jinjiang Yang, Hanning Wang, Min Zhu, Shaojun Wei, Aoyang Zhang, Leibo Liu

Abstract:

Fully homomorphic encryption (FHE) enables privacy-preserving machine learning (PPML) at the cost of intensive computational overhead, which necessitates the use of domain-specific accelerators. To achieve comprehensive support for leveled FHE, this article presents a reconfigurable multi-scheme FHE processor that supports both client-side encryption/decryption and server-side evaluation. First, a reconfigurable …

Published in: IEEE Journal of Solid-State Circuits
Page(s): 1 – 15
Year of Publication: 2026
Electronic ISSN: 1558-173X
DOI: 10.1109/JSSC.2026.3657525
Publisher: IEEE

View on IEEE Xplore

  • No Comments

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
Author(s): Prerna Budhkar, Mirco Sciulli, Srivatsa Rangachar Srinivasa, Gauthaman Murali, Ragh Kuttappa, Paolo Aseron, Trang Nguyen, Vinayak Honkote, Tanay Karnik

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 …

Published in: IEEE Solid-State Circuits Letters
Page(s): 69 – 72
Year of Publication: 2026
Electronic ISSN: 2573-9603
DOI: 10.1109/LSSC.2026.3659575
Publisher: IEEE

View on IEEE Xplore

  • 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 29-Gb/mm2 1-Tb 3-b/Cell 3-D Flash Memory With CMOS Direct Bonded Array (CBA) Technology

A 29-Gb/mm2 1-Tb 3-b/Cell 3-D Flash Memory With CMOS Direct Bonded Array (CBA) Technology 150 150
Author(s): Kosuke Yanagidaira, Mario Sako, Yasuhiro Hirashima, Yumi Higashi, Yutaka Shimizu, Takeshi Nakano, Yusuke Ochi, Hiroaki Yamada, Nobushi Matsuura, Akihiro Imamoto, Kazuaki Kawaguchi, Koji Tabata, Hiroaki Hoshino, Takeshi Hioka, Shigehito Saigusa, Hiroki Date, Masaki Unno, Jumpei Sato, You Kamata, Takahiro Shimizu, Akio Sugahara, Taira Shibuya, Atsushi Okuyama, Junji Yamada, Takatoshi Minamoto, Hardwell Chibvongodze, Naoki Ojima, Hiroshi Sugawara, Masahiro Kano, Jang-Woo Lee, Hiroyuki Mizukoshi, Ryuji Yamashita, Kensaku Abe, Naohito Morozumi, In-Soo Yoon, Takuya Ariki, Jong Hak Yuh, Khin Htoo, Yosuke Kato, Yoshihiro Oyama, Yen-Lung Jason Li, Yoshihisa Watanabe, Toshiyuki Kouchi, Koji Hosono

Abstract:

This article reports a 1-Tb 3-b/cell 3-D flash memory fabricated with CMOS direct bonded array (CBA) technology. Compaction of circuits and wires achieves the highest bit density in the world over 29 Gb/mm2 with 332-word line (WL) layers. The bit density is improved by 71% from a previous generation despite …

Published in: IEEE Journal of Solid-State Circuits
Page(s): 237 – 249
Year of Publication: 2026
Electronic ISSN: 1558-173X
DOI: 10.1109/JSSC.2025.3617579
Publisher: IEEE

View on IEEE Xplore

  • No Comments

MINOTAUR: A Posit-Based 0.42–0.50-TOPS/W Edge Transformer Inference and Training Accelerator

MINOTAUR: A Posit-Based 0.42–0.50-TOPS/W Edge Transformer Inference and Training Accelerator 150 150
Author(s): Kartik Prabhu, Robert M. Radway, Jeffrey Yu, Kai Bartolone, Massimo Giordano, Fabian Peddinghaus, Yonatan Urman, Win-San Khwa, Yu-Der Chih, Meng-Fan Chang, Subhasish Mitra, Priyanka Raina

Abstract:

Transformer models have revolutionized natural language processing (NLP) and enabled many new applications, but are challenging to deploy on resource-constrained edge devices due to their high computation and memory demands. We present MINOTAUR, an edge system-on-chip (SoC) for inference and fine-tuning of Transformer models with all memory on the chip. …

Published in: IEEE Journal of Solid-State Circuits
Page(s): 1311 – 1323
Year of Publication: 2025
Electronic ISSN: 1558-173X
DOI: 10.1109/JSSC.2025.3545731
Publisher: IEEE

View on IEEE Xplore

Search this website