pHNSW: PCA-Based Filtering to Accelerate HNSW Approximate Nearest Neighbor Search

TL;DR

pHNSW introduces PCA filtering and custom hardware instructions to significantly accelerate HNSW approximate nearest neighbor searches, reducing energy consumption and increasing throughput on CPU and GPU platforms.

Contribution

The paper presents a novel algorithm-hardware co-optimized approach, pHNSW, that enhances HNSW search efficiency through PCA-based filtering and specialized processor design.

Findings

QPS increased by up to 21.37x on CPU

Energy consumption reduced by up to 57.4%

Significant throughput improvements on GPU

Abstract

Hierarchical Navigable Small World (HNSW) has demonstrated impressive accuracy and low latency for high-dimensional nearest neighbor searches. However, its high computational demands and irregular, large-volume data access patterns present significant challenges to search efficiency. To address these challenges, we introduce pHNSW, an algorithm-hardware co-optimized solution that accelerates HNSW through Principal Component Analysis (PCA) filtering. On the algorithm side, we apply PCA filtering to reduce the dimensionality of the dataset, thereby lowering the volume of neighbor access and decreasing the computational load for distance calculations. On the hardware side, we design the pHNSW processor with custom instructions to optimize search throughput and energy efficiency. In the experiments, we synthesized the pHNSW processor RTL design with a 65nm technology node and evaluated it using DDR4 and HBM1.0 DRAM standards. The results show that pHNSW boosts Queries per Second (QPS) by 14.47x-21.37x on a CPU and 5.37x-8.46x on a GPU, while reducing energy consumption by up to 57.4% compared to standard HNSW implementation.