DIRC-RAG: Accelerating Edge RAG with Robust High-Density and High-Loading-Bandwidth Digital In-ReRAM Computation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Shao, Kunming, Liao, Zhipeng, Yu, Jiangnan, Zhao, Liang, Li, Qiwei, Huang, Xijie, He, Jingyu, Tian, Fengshi, Zou, Yi, Wang, Xiaomeng, Cheng, Tim Kwang-Ting, Tsui, Chi-Ying
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909875853852672
author Shao, Kunming
Liao, Zhipeng
Yu, Jiangnan
Zhao, Liang
Li, Qiwei
Huang, Xijie
He, Jingyu
Tian, Fengshi
Zou, Yi
Wang, Xiaomeng
Cheng, Tim Kwang-Ting
Tsui, Chi-Ying
author_facet Shao, Kunming
Liao, Zhipeng
Yu, Jiangnan
Zhao, Liang
Li, Qiwei
Huang, Xijie
He, Jingyu
Tian, Fengshi
Zou, Yi
Wang, Xiaomeng
Cheng, Tim Kwang-Ting
Tsui, Chi-Ying
contents Retrieval-Augmented Generation (RAG) enhances large language models (LLMs) by integrating external knowledge retrieval but faces challenges on edge devices due to high storage, energy, and latency demands. Computing-in-Memory (CIM) offers a promising solution by storing document embeddings in CIM macros and enabling in-situ parallel retrievals but is constrained by either low memory density or limited computational accuracy. To address these challenges, we present DIRCRAG, a novel edge RAG acceleration architecture leveraging Digital In-ReRAM Computation (DIRC). DIRC integrates a high-density multi-level ReRAM subarray with an SRAM cell, utilizing SRAM and differential sensing for robust ReRAM readout and digital multiply-accumulate (MAC) operations. By storing all document embeddings within the CIM macro, DIRC achieves ultra-low-power, single-cycle data loading, substantially reducing both energy consumption and latency compared to offchip DRAM. A query-stationary (QS) dataflow is supported for RAG tasks, minimizing on-chip data movement and reducing SRAM buffer requirements. We introduce error optimization for the DIRC ReRAM-SRAM cell by extracting the bit-wise spatial error distribution of the ReRAM subarray and applying targeted bit-wise data remapping. An error detection circuit is also implemented to enhance readout resilience against deviceand circuit-level variations. Simulation results demonstrate that DIRC-RAG under TSMC40nm process achieves an on-chip non-volatile memory density of 5.18Mb/mm2 and a throughput of 131 TOPS. It delivers a 4MB retrieval latency of 5.6μs/query and an energy consumption of 0.956μJ/query, while maintaining the retrieval precision.
format Preprint
id arxiv_https___arxiv_org_abs_2510_25278
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle DIRC-RAG: Accelerating Edge RAG with Robust High-Density and High-Loading-Bandwidth Digital In-ReRAM Computation
Shao, Kunming
Liao, Zhipeng
Yu, Jiangnan
Zhao, Liang
Li, Qiwei
Huang, Xijie
He, Jingyu
Tian, Fengshi
Zou, Yi
Wang, Xiaomeng
Cheng, Tim Kwang-Ting
Tsui, Chi-Ying
Hardware Architecture
Retrieval-Augmented Generation (RAG) enhances large language models (LLMs) by integrating external knowledge retrieval but faces challenges on edge devices due to high storage, energy, and latency demands. Computing-in-Memory (CIM) offers a promising solution by storing document embeddings in CIM macros and enabling in-situ parallel retrievals but is constrained by either low memory density or limited computational accuracy. To address these challenges, we present DIRCRAG, a novel edge RAG acceleration architecture leveraging Digital In-ReRAM Computation (DIRC). DIRC integrates a high-density multi-level ReRAM subarray with an SRAM cell, utilizing SRAM and differential sensing for robust ReRAM readout and digital multiply-accumulate (MAC) operations. By storing all document embeddings within the CIM macro, DIRC achieves ultra-low-power, single-cycle data loading, substantially reducing both energy consumption and latency compared to offchip DRAM. A query-stationary (QS) dataflow is supported for RAG tasks, minimizing on-chip data movement and reducing SRAM buffer requirements. We introduce error optimization for the DIRC ReRAM-SRAM cell by extracting the bit-wise spatial error distribution of the ReRAM subarray and applying targeted bit-wise data remapping. An error detection circuit is also implemented to enhance readout resilience against deviceand circuit-level variations. Simulation results demonstrate that DIRC-RAG under TSMC40nm process achieves an on-chip non-volatile memory density of 5.18Mb/mm2 and a throughput of 131 TOPS. It delivers a 4MB retrieval latency of 5.6μs/query and an energy consumption of 0.956μJ/query, while maintaining the retrieval precision.
title DIRC-RAG: Accelerating Edge RAG with Robust High-Density and High-Loading-Bandwidth Digital In-ReRAM Computation
topic Hardware Architecture
url https://arxiv.org/abs/2510.25278