iMARS: An In-Memory-Computing Architecture for Recommendation Systems

2022/02/18 by Mengyuan Li, Ann Franchesca Laguna, Li, Mengyuan +9 · 2 citations
Computer Science · Engineering · #Advanced Memory and Neural Computing #Caching and Content Delivery #FOS: Computer and information sciences #Ferroelectric and Negative Capacitance Devices #Hardware Architecture (cs.AR)

paper · pdf · doi:10.48550/arxiv.2202.09433

openalex publication_date 2022/02/18 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Recommendation systems (RecSys) suggest items to users by predicting their preferences based on historical data. Typical RecSys handle large embedding tables and many embedding table related operations. The memory size and bandwidth of the conventional computer architecture restrict the performance of RecSys. This work proposes an in-memory-computing (IMC) architecture (iMARS) for accelerating the filtering and ranking stages of deep neural network-based RecSys. iMARS leverages IMC-friendly embedding tables implemented inside a ferroelectric FET based IMC fabric. Circuit-level and system-level evaluation show that \fw achieves 16.8x (713x) end-to-end latency (energy) improvement compared to the GPU counterpart for the MovieLens dataset.

Cited by

• ReCross: Efficient Embedding Reduction Scheme for In-Memory Computing using ReRAM-Based Crossbar
• Content Addressable Memory Design with Reference Resistor for Improved Search Resolution

Related

• MP-Rec: Hardware-Software Co-Design to Enable Multi-Path Recommendation
• MicroRec: Efficient Recommendation Inference by Hardware and Data Structure Solutions
• RecNMP: Accelerating Personalized Recommendation with Near-Memory Processing
• ReCross: Efficient Embedding Reduction Scheme for In-Memory Computing using ReRAM-Based Crossbar
• Mem-Rec: Memory Efficient Recommendation System using Alternative Representation