1T1C DRAM Cell

1T1C DRAM cells, the fundamental building blocks of dynamic random-access memory, are being actively researched for their potential to accelerate various computationally intensive tasks directly within the memory array itself (processing-in-memory). Current research focuses on leveraging 1T1C cells for efficient deep neural network (DNN) inference and training, including exploring novel architectures like compositional embeddings and output-stationary matrix multiplication, as well as mitigating vulnerabilities like RowHammer attacks. This work aims to improve the speed and energy efficiency of AI applications, particularly on resource-constrained devices, by reducing data movement between memory and processing units.

Papers

October 25, 2024

Ripple: Accelerating LLM Inference on Smartphones with Correlation-Aware Neuron Management
Tuowei Wang, Ruwen Fan, Minxing Huang, Zixu Hao, Kun Li, Ting Cao, Youyou Lu, Yaoxue Zhang, Ju Ren
LLM Inference Smartphone Device Correlation Graph Neuron Coverage 1T1C DRAM Cell

October 17, 2024

Harnessing Your DRAM and SSD for Sustainable and Accessible LLM Inference with Mixed-Precision and Multi-level Caching
Jie Peng, Zhang Cao, Huaizhi Qu, Zhengyu Zhang, Chang Guo, Yanyong Zhang, Zhichao Zhang, Tianlong Chen
Scientific Inference LLM Inference Mixed Precision 1T1C DRAM Cell

February 6, 2024

ProactivePIM: Accelerating Weight-Sharing Embedding Layer with PIM for Scalable Recommendation System
Youngsuk Kim, Junghwan Lim, Hyuk-Jae Lee, Chae Eun Rhee
Low Rate ACCELERATION Hash Code Memory Capacity Near Memory Memory Hierarchy 1T1C DRAM Cell Heterogeneous Memory

December 12, 2023

LLM in a flash: Efficient Large Language Model Inference with Limited Memory
Keivan Alizadeh, Iman Mirzadeh, Dmitry Belenko, Karen Khatamifard, Minsik Cho, Carlo C Del Mundo, Mohammad Rastegari, Mehrdad Farajtabar
Large Language Model Medical LLM Large Language Model Inference Inference Speed Limited Memory 1T1C DRAM Cell

May 14, 2023

DNN-Defender: A Victim-Focused In-DRAM Defense Mechanism for Taming Adversarial Weight Attack on DNNs
Ranyang Zhou, Sabbir Ahmed, Adnan Siraj Rakin, Shaahin Angizi
Deep Neural Network DEfense Mechanism Bit Flip Attack 1T1C DRAM Cell

May 4, 2023

CAMEL: Co-Designing AI Models and Embedded DRAMs for Efficient On-Device Learning
Sai Qian Zhang, Thierry Tambe, Nestor Cuevas, Gu-Yeon Wei, David Brooks
Device Training Device Learning Electric SHEEP Human AI Co Creation Random Access Memory Chip Memory 1T1C DRAM Cell

April 8, 2023

EnforceSNN: Enabling Resilient and Energy-Efficient Spiking Neural Network Inference considering Approximate DRAMs for Embedded Systems
Rachmad Vidya Wicaksana Putra, Muhammad Abdullah Hanif, Muhammad Shafique
Deep Network Tiny Refinement Elicit Resilience Embedded System Energy Efficient Random Access Memory 1T1C DRAM Cell DRAM Access Energy

November 4, 2022

Learning to Rank Graph-based Application Objects on Heterogeneous Memories
Diego Moura, Vinicius Petrucci, Daniel Mosse
LeArning Abstract Memory Trace Memory Model Non Volatile Memory 1T1C DRAM Cell Heterogeneous Memory Inter Object Relationship Graph

July 16, 2022

MAC-DO: An Efficient Output-Stationary GEMM Accelerator for CNNs Using DRAM Technology
Minki Jeong, Wanyeong Jung
CNN Model Random Access Memory Multiply Accumulate 1T1C DRAM Cell DRAM Access Energy

February 19, 2022

HDAM: Heuristic Difference Attention Module for Convolutional Neural Networks
Yu Xue, Ziming Yuan
Convolutional Neural Network Attention Mechanism Receptive Field Receptive Field Size 1T1C DRAM Cell

1T1C DRAM Cell

Papers

Ripple: Accelerating LLM Inference on Smartphones with Correlation-Aware Neuron Management

Harnessing Your DRAM and SSD for Sustainable and Accessible LLM Inference with Mixed-Precision and Multi-level Caching

ProactivePIM: Accelerating Weight-Sharing Embedding Layer with PIM for Scalable Recommendation System

LLM in a flash: Efficient Large Language Model Inference with Limited Memory

DNN-Defender: A Victim-Focused In-DRAM Defense Mechanism for Taming Adversarial Weight Attack on DNNs

CAMEL: Co-Designing AI Models and Embedded DRAMs for Efficient On-Device Learning

EnforceSNN: Enabling Resilient and Energy-Efficient Spiking Neural Network Inference considering Approximate DRAMs for Embedded Systems

Learning to Rank Graph-based Application Objects on Heterogeneous Memories

MAC-DO: An Efficient Output-Stationary GEMM Accelerator for CNNs Using DRAM Technology

HDAM: Heuristic Difference Attention Module for Convolutional Neural Networks