Article
Optics
Yulong Dong, Lin Lin
Summary: The study proposes a quantum version of the UNPACK benchmark to measure the whole machine performance of quantum computers, which is considered as the minimal requirement for quantum computers to solve useful tasks. The RACBEM model is efficient to be implemented on a quantum computer and can be designed to adapt to different quantum architectures for solving linear algebra problems.
Article
Computer Science, Theory & Methods
Guangming Tan, Chaoyang Shui, Yinshan Wang, Xianzhi Yu, Yujin Yan
Summary: This article presents a fine-grained pipelining algorithm on large-scale CPU-GPU heterogeneous cluster systems to achieve higher floating-point efficiency. The research results showcase some trade-offs between algorithm design, programming support, and architecture design.
IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS
(2021)
Article
Mathematics
Vladimir O. Rybintsev
Summary: Based on the analysis of TOP500 results, a functional dependence of the performance of clusters without accelerators on their parameters was identified. The comparison between calculated and tested results showed an estimation error of no more than 2% for processors from different generations and manufacturers. This level of accuracy allows for successful prediction of cluster performance.
Article
Computer Science, Artificial Intelligence
Giray Havur, Cristina Cabanillas, Axel Polleres
Summary: Declarative logic programming is suitable for modeling optimization and configuration problems, and this paper introduces a benchmark for ASP systems to evaluate their performance in complex resource allocation. The benchmark includes a declarative encoding, a configurable benchmark tool, and an evaluation of state-of-the-art ASP systems. It serves as a valuable resource for the ASP and BPM communities.
EXPERT SYSTEMS WITH APPLICATIONS
(2022)
Article
Computer Science, Information Systems
Enrico Calore, Sebastiano Fabio Schifano
Summary: This article introduces a performance evaluation tool FER based on FPGA accelerators, as well as the test results on Xilinx Alveo accelerator cards. FER can be used to evaluate the computing performance and memory bandwidth of FPGA accelerators, form Roofline plots for performance comparison and estimate performance upper bounds.
Article
Computer Science, Interdisciplinary Applications
Arthur Pratti Dadalto, Fabio Luiz Usberti, Mario Cesar San Felice
Summary: This paper addresses the Minimum Subgraph Diameter Problem (MSDP), which is an NP-hard problem with applications to network design. The paper proposes a Mixed-Integer Linear Programming (MILP) model for the MSDP and presents two families of valid inequalities as the basis for the first exact approaches to solve the problem. The paper also introduces an approach to apply lazy constraints on the MILP and a heuristic to generate upper bounds from fractional solutions. The authors conduct a comprehensive benchmark and computational study to evaluate the effectiveness of their contributions in reducing optimality gaps and computational times.
COMPUTERS & OPERATIONS RESEARCH
(2023)
Article
Computer Science, Artificial Intelligence
Qihao Liu, Xinyu Li, Liang Gao
Summary: Process planning is a crucial part of manufacturing systems that require complex models to accurately describe its characteristics. By introducing a new MILP mathematical model and a hybrid evolutionary algorithm, the efficiency and superiority of both methods have been successfully validated through numerical experiments.
JOURNAL OF INTELLIGENT MANUFACTURING
(2021)
Article
Computer Science, Software Engineering
Achiya Elyasaf, Tom Yaacov, Gera Weiss
Summary: This article identifies and demonstrates a weakness of Petri Nets (PN) in specifying composite behavior of reactive systems. To address this problem, the article proposes to model the system behavior using behavioral programming (BP), which allows for avoiding over-specification and increasing the agility of the specification.
IEEE TRANSACTIONS ON SOFTWARE ENGINEERING
(2023)
Article
Computer Science, Theory & Methods
Christian R. Trott, Damien Lebrun-Grandie, Daniel Arndt, Jan Ciesko, Vinh Dang, Nathan Ellingwood, Rahulkumar Gayatri, Evan Harvey, Daisy S. Hollman, Dan Ibanez, Nevin Liber, Jonathan Madsen, Jeff Miles, David Poliakoff, Amy Powell, Sivasankaran Rajamanickam, Mikael Simberg, Dan Sunderland, Bruno Turcksin, Jeremiah Wilke
Summary: This article introduces the Kokkos Performance Portable Programming Model, which allows developers to write single source applications for diverse high-performance computing architectures. Kokkos provides key abstractions for the compute and memory hierarchy of modern hardware and demonstrates the performance of new features for exascale-era architectures.
IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS
(2022)
Article
Green & Sustainable Science & Technology
Maria-Luisa Gomez-Moreno
Summary: This study examines the relationship between employment policies and the evolution of the productive system, finding that economic agents' responses are more influenced by changes in the international environment. It highlights the significant impact of corporate culture on the success of employment policies and suggests rethinking innovative theoretical and practical approaches in addressing endemic unemployment in semi-peripheral areas.
Article
Computer Science, Hardware & Architecture
Yichang Hu, Lu Lu
Summary: The paper proposes a simulation model of HPL that can simulate the execution process of HPL in heterogeneous GPU-enhanced clusters, helping researchers quickly evaluate floating-point performance and providing a reference for hardware investment.
JOURNAL OF SUPERCOMPUTING
(2021)
Article
Computer Science, Hardware & Architecture
Adriano Marques Garcia, Dalvan Griebler, Claudio Schepke, Luiz Gustavo Fernandes
Summary: This study extends and improves the SPBench benchmarking framework to support dynamic micro-batching and data stream frequency management. It proposes a set of algorithms for benchmarking stream processing and creates various test scenarios. The results demonstrate that micro-batching does not benefit test cases on multi-cores, and Intel TBB provides the lowest latency while FastFlow ensures higher throughput in shorter pipelines for different data stream frequency configurations.
JOURNAL OF SUPERCOMPUTING
(2023)
Article
Computer Science, Information Systems
Jovan Dukic, Marko Misic
Summary: This paper explores and evaluates the applicability of the directive-based programming model OpenACC for parallelizing algorithms on GPUs. The results show that parallelized mini-applications using OpenACC can significantly improve performance compared to sequential implementations, and in some cases, even outperform CUDA implementations. Furthermore, OpenACC requires less programming effort and code restructuring compared to low-level models like CUDA and OpenCL.
Article
Thermodynamics
Yannick Wack, Sylvain Serra, Martine Baelmans, Jean-Michel Reneaume, Maarten Blommaert
Summary: This paper compares two nonlinear topology optimization methods for District Heating Networks in terms of computational cost and optimality gap. The benchmark demonstrates that the density-based approach has subquadratic scaling in computational cost, making it suitable for large-scale problems, while the combinatorial approach has exponential scaling. The density-based method optimized a network for 600 streets in only 35 minutes, compared to 29 hours required by the combinatorial approach. Resolving the integer constraint on pipe placement does not necessarily lead to a superior design, but makes optimization of large-scale problems intractable. Further study highlights the importance of initialization strategies when solving the nonlinear topology optimization problem.
Article
Multidisciplinary Sciences
Xingzheng Lyu, Li Cheng, Sanyuan Zhang
Summary: Topological and geometrical analysis of retinal blood vessels is an economical way to detect common diseases. In this study, a novel benchmark database RETA was constructed to facilitate retinal vessel analysis, with improved annotation quality compared to existing datasets. The publicly available annotation software enables researchers to develop vessel segmentation algorithms and study cross-modality tubular structure segmentation and analysis.