Abstract:Approximate Nearest Neighbor Search (ANNS), which enables efficient semantic similarity search in large datasets, has become a fundamental component of critical applications such as information retrieval and retrieval-augmented generation (RAG). However, ANNS is a well-known I/O-intensive algorithm with a low compute-to-I/O ratio, often requiring massive storage due to the large volume of high-dimensional data. This leads to I/O bottlenecks on CPUs and memory limitations on GPUs. DRAM-based Processing-in-Memory (DRAM-PIM) architecture, which offers high bandwidth, large-capacity memory, and the ability to perform efficient computation in or near the data, presents a promising solution for ANNS. In this work, we investigate the use of commercial DRAM-PIM for ANNS for the first time and propose DRIM-ANN, an optimized ANNS engine based on DRAM-PIMs from UPMEM. Notably, given that the target DRAM-PIM exhibits an even lower compute-to-I/O ratio than basic ANNS, we leverage lookup tables (LUTs) to replace more multiplications with I/O operations. We then systematically tune ANNS to search optimized configurations with lower computational load, aligning the compute-to-I/O ratio of ANNS with that of DRAM-PIMs while maintaining accuracy constraints. Building on this tuned ANNS algorithm, we further explore implementation optimizations to fully utilize the two thousand parallel processing units with private local memory in DRAM-PIMs. To address the load imbalance caused by ANNS requests distributed across different clusters of large datasets, we propose a load-balancing strategy that combines static data layout optimization with dynamic runtime request scheduling. Experimental results on representative datasets show that DRIM-ANN achieves an average performance speedup of 2.92x compared to a 32-thread CPU counterpart.

DAS: A DRAM-Based Annealing System for Solving Large-Scale Combinatorial Optimization Problems.

A Compute-in-Memory Annealing Processor With Interaction Coefficient Reuse and Sparse Energy Computation for Solving Combinatorial Optimization Problems

AIM: Annealing in Memory for Vision Applications.

DaDianNao: A Machine-Learning Supercomputer

A Variation-Tolerant Continuous-Time Ising Machine with Edram-Based Spin Interaction and Leaked Negative Feedback Annealing

30.5 A Variation-Tolerant In-eDRAM Continuous-Time Ising Machine Featuring 15-Level Coefficients and Leaked Negative-Feedback Annealing

Annealing Processing Architecture of 28-nm CMOS Chip for Ising Model With 512 Fully Connected Spins

DASM: Data-Streaming-Based Computing in Nonvolatile Memory Architecture for Embedded System

Scalable In-Memory Clustered Annealer With Temporal Noise of Charge Trap Transistor for Large Scale Travelling Salesman Problems

A Scalable Universal Ising Machine Based on Interaction-Centric Storage and Compute-in-memory

CMOS-compatible Ising and Potts Annealing Using Single Photon Avalanche Diodes

DRIM-ANN: An Approximate Nearest Neighbor Search Engine based on Commercial DRAM-PIMs

Energy-efficient superparamagnetic Ising machine and its application to traveling salesman problems

A Ferroelectric Compute-in-Memory Annealer for Combinatorial Optimization Problems

Parallel Ising Annealer via Gradient-based Hamiltonian Monte Carlo

ASiM: Improving Transparency of SRAM-based Analog Compute-in-Memory Research with an Open-Source Simulation Framework

Scalable Fully-Coupled Annealing Processing System Implementing 4096 Spins Using 22nm CMOS LSI

Area-and Power-Efficient Compute-in-Memory Macro Based on Oxide Semiconductor 2T0C Edram with Cell-System Co-Optimization

Multifunctional RRAM Chip with Configurability for Sparsity-Aware In-Memory ISNG Machine

ReDas: A Lightweight Architecture for Supporting Fine-Grained Reshaping and Multiple Dataflows on Systolic Array