The Adaptive Priority Queue with Elimination and Combining

TL;DR

This paper introduces a novel adaptive priority queue design that combines elimination and combining techniques to enhance scalability and concurrency in parallel environments, addressing limitations of existing methods.

Contribution

It presents a new priority queue architecture that leverages parallel insertions, batched removals, and elimination algorithms, improving scalability and reducing contention.

Findings

Enhanced scalability with parallel insertions and batched removals.

Increased concurrency through a new elimination algorithm.

Effective performance across various implementation techniques.

Abstract

Priority queues are fundamental abstract data structures, often used to manage limited resources in parallel programming. Several proposed parallel priority queue implementations are based on skiplists, harnessing the potential for parallelism of the add() operations. In addition, methods such as Flat Combining have been proposed to reduce contention by batching together multiple operations to be executed by a single thread. While this technique can decrease lock-switching overhead and the number of pointer changes required by the removeMin() operations in the priority queue, it can also create a sequential bottleneck and limit parallelism, especially for non-conflicting add() operations. In this paper, we describe a novel priority queue design, harnessing the scalability of parallel insertions in conjunction with the efficiency of batched removals. Moreover, we present a new elimination algorithm suitable for a priority queue, which further increases concurrency on balanced workloads with similar numbers of add() and removeMin() operations. We implement and evaluate our design using a variety of techniques including locking, atomic operations, hardware transactional memory, as well as employing adaptive heuristics given the workload.