未提供

本文提出了一种新的图论算法评估基准，旨在解决现有图基准的不足。该基准包括一个图生成器和一系列核心组件，每个核心组件代表一类特定的图算法。图生成器基于偏好依附方法快速合成规模自由图，核心组件则涵盖了广泛的图算法，包括搜索、组合优化、度量计算以及近年来更受欢迎的图挖掘方法。该基准设计用于准确模拟目标算法的运行时特性，同时限制计算和空间开销，以确保在合理时间内完成计算。此基准预期将成为评估不同架构和编程模型运行图算法的重要工具。

This paper proposes a novel graph algorithm evaluation benchmark aimed at addressing the shortcomings of existing graph benchmarks. This benchmark consists of a graph generator and a set of core components, each representing a specific category of graph algorithms. The graph generator quickly synthesizes scale-free graphs based on the preferential attachment model, while the core components cover a wide range of graph algorithms, including search, combinatorial optimization, metric computation, and the increasingly popular graph mining methods in recent years. This benchmark is designed to accurately simulate the runtime characteristics of target algorithms, while limiting computational and spatial overhead to ensure computations can be completed within a reasonable time frame. This benchmark is expected to become an important tool for evaluating graph algorithms across different architectures and programming models.