@inproceedings{08629d7da4e441f89574b1a8cecfcc42,
title = "A DRAM-based Near-Memory Architecture for Accelerated and Energy-Efficient Execution of Transformers",
abstract = "Transformers-based language models have achieved remarkable accuracy in various NLP tasks, employing self-attention mechanisms primarily based on matrix multiplication. However, their significant size leads to data movement issues, causing latency and energy efficiency challenges in conventional Von-Neumann systems. To mitigate these issues, several in-memory and near-memory architectures have been proposed. This paper introduces PACT-3D, a near-memory architecture featuring novel computing units integrated with DRAM banks. PACT-3D significantly reduces latency by 1.7 {\texttimes} and improves energy efficiency by 18.7 {\texttimes} compared to state-of-the-art near-memory architectures.",
keywords = "DRAM, Energy Efficiency, In/Near-memory Processing, LLMs, Memory Wall, Transformers",
author = "Gian Singh and Sarma Vrudhula",
note = "Publisher Copyright: {\textcopyright} 2024 ACM.; 34th Great Lakes Symposium on VLSI 2024, GLSVLSI 2024 ; Conference date: 12-06-2024 Through 14-06-2024",
year = "2024",
month = jun,
day = "12",
doi = "10.1145/3649476.3658732",
language = "English (US)",
series = "Proceedings of the ACM Great Lakes Symposium on VLSI, GLSVLSI",
publisher = "Association for Computing Machinery",
pages = "57--62",
booktitle = "GLSVLSI 2024 - Proceedings of the Great Lakes Symposium on VLSI 2024",
}