AME: An Efficient Heterogeneous Agentic Memory Engine for Smartphones

TL;DR

This paper introduces AME, a specialized memory engine for smartphones that efficiently manages evolving vector data for on-device AI, addressing hardware and workload constraints to improve performance.

Contribution

AME is a novel on-device agentic memory engine co-designed with smartphone hardware, featuring a hardware-aware pipeline and scheduling scheme for efficient vector data management.

Findings

Up to 1.4x query throughput improvement at matched recall

Up to 7x faster index construction

Up to 6x higher insertion throughput under concurrent workloads

Abstract

On-device agents on smartphones increasingly require continuously evolving memory to support personalized, context-aware, and long-term behaviors. To meet both privacy and responsiveness demands, user data is embedded as vectors and stored in a vector database for fast similarity search. However, most existing vector databases target server-class environments. When ported directly to smartphones, two gaps emerge: (G1) a mismatch between mobile SoC constraints and vector-database assumptions, including tight bandwidth budgets, limited on-chip memory, and stricter data type and layout constraints; and (G2) a workload mismatch, because on-device usage resembles a continuously learning memory, in which queries must coexist with frequent inserts, deletions, and ongoing index maintenance. To address these challenges, we propose AME, an on-device Agentic Memory Engine co-designed with modern smartphone SoCs. AME introduces two key techniques: (1) a hardware-aware, high-efficiency matrix pipeline that maximizes compute-unit utilization and exploits multi-level on-chip storage to sustain high throughput; and (2) a hardware- and workload-aware scheduling scheme that coordinates querying, insertion, and index rebuilding to minimize latency. We implement AME on Snapdragon 8-series SoCs and evaluate it on HotpotQA. In our experiments, AME improves query throughput by up to 1.4x at matched recall, achieves up to 7x faster index construction, and delivers up to 6x higher insertion throughput under concurrent query workloads.