8 papers
MX-SAFE: Versatile Inference- and Training-Proof Microscaling Format with On-the-Fly Exponent and Mantissa Bit Allocation
Dahoon Park, Jahyun Koo, Sangwoo Hwang +1
As the demand for deep learning grows, cost reduction through quantization has become essential for both training and inference. In 2022, the Open Compute Project (OCP) consortium…
Dissecting and Re-architecting 3D NAND Flash PIM Arrays for Efficient Single-Batch Token Generation in LLMs
Yongjoo Jang, Sangwoo Hwang, Hojin Lee +4
The advancement of large language models has led to models with billions of parameters, significantly increasing memory and compute demands. Serving such models on conventional har…
Flexible In-NAND Cryptographic Processing for Secure Flash Storage
Seock-Hwan Noh, Hoyeon Lee, Junkyum Kim +6
We present FlashVault, an in-NAND self-encryption architecture that embeds a reconfigurable cryptographic engine into the unused silicon area of a state-of-the-art 4D V-NAND struct…
Jack Unit: An Area- and Energy-Efficient Multiply-Accumulate (MAC) Unit Supporting Diverse Data Formats
Seock-Hwan Noh, Sungju Kim, Seohyun Kim +3
In this work, we introduce an area- and energy-efficient multiply-accumulate (MAC) unit, named Jack unit, that is a jack-of-all-trades, supporting various data formats such as inte…
FlexNeRFer: A Multi-Dataflow, Adaptive Sparsity-Aware Accelerator for On-Device NeRF Rendering
Seock-Hwan Noh, Banseok Shin, Jeik Choi +3
Neural Radiance Fields (NeRF), an AI-driven approach for 3D view reconstruction, has demonstrated impressive performance, sparking active research across fields. As a result, a ran…
OPAL: Outlier-Preserved Microscaling Quantization Accelerator for Generative Large Language Models
Jahyun Koo, Dahoon Park, Sangwoo Jung +1
To overcome the burden on the memory size and bandwidth due to ever-increasing size of large language models (LLMs), aggressive weight quantization has been recently studied, while…
One-Spike SNN: Single-Spike Phase Coding with Base Manipulation for ANN-to-SNN Conversion Loss Minimization
Sangwoo Hwang, Jaeha Kung
As spiking neural networks (SNNs) are event-driven, energy efficiency is higher than conventional artificial neural networks (ANNs). Since SNN delivers data through discrete spikes…
All-rounder: A Flexible AI Accelerator with Diverse Data Format Support and Morphable Structure for Multi-DNN Processing
Seock-Hwan Noh, Seungpyo Lee, Banseok Shin +3
Recognizing the explosive increase in the use of AI-based applications, several industrial companies developed custom ASICs (e.g., Google TPU, IBM RaPiD, Intel NNP-I/NNP-T) and con…