Beyond a Single Queue: Multi-Level-Multi-Queue as an Effective Design for SSSP problems on GPUs

2026-02-10

AI Summary

Existing GPU-accelerated single-source shortest path (SSSP) algorithms suffer from high synchronization overhead, substantial memory latency, and poor adaptability to diverse graph structures due to their reliance on a single fixed queue. This work proposes a multi-level multi-queue (MLMQ) architecture that, for the first time, introduces a hierarchical queue design into SSSP computation. By aligning with the GPU’s memory hierarchy and parallelism characteristics, MLMQ achieves efficient queue scheduling through unified read-write primitives, a cache-aware cooperative queuing mechanism, and an input-aware dynamic configuration strategy. Evaluated across a variety of graph datasets, the proposed approach achieves average speedups ranging from 1.87× to 17.13× over the state-of-the-art methods, significantly expanding the optimization potential of SSSP on GPUs.

Technology Category

• Machine Learning: Hardware-aware ML
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines

Abstract

As one of the most fundamental problems in graph processing, the Single-Source Shortest Path (SSSP) problem plays a critical role in numerous application scenarios. However, existing GPU-based solutions remain inefficient, as they typically rely on a single, fixed queue design that incurs severe synchronization overhead, high memory latency, and poor adaptivity to diverse inputs. To address these inefficiencies, we propose MultiLevelMultiQueue (MLMQ), a novel data structure that distributes multiple queues across the GPU's multi-level parallelism and memory hierarchy. To realize MLMQ, we introduce a cache-like collaboration mechanism for efficient inter-queue coordination, and develop a modular queue design based on unified Read and Write primitives. Within this framework, we expand the optimization space by designing a set of GPU-friendly queues, composing them across multiple levels, and further providing an input-adaptive MLMQ configuration scheme. Our MLMQ design achieves average speedups of 1.87x to 17.13x over state-of-the-art implementations. Our code is open-sourced at https://github.com/Leo9660/MLMQ.git.

Problem

• SSSP
• GPU
• queue design
• synchronization overhead
• memory latency

Innovation

• MultiLevelMultiQueue
• GPU graph processing
• SSSP
• multi-queue design
• input-adaptive optimization