Filters








18 Hits in 4.4 sec

Research trend of large-scale supercomputers and applications from the TOP500 and Gordon Bell Prize

Weimin Zheng
2020 Science China Information Sciences  
One year later, Haohuan FU et al. won the Gordon Bell Prize with their study again with a nonlinear earthquake simulation application running on Sunway TaihuLight.  ...  The second trend is that artificial intelligence applications are expected to become one of the main stream applications of supercomputing.  ...  scalable parallel systems.  ... 
doi:10.1007/s11432-020-2861-0 fatcat:73afvlh5wneq3oqkj2nlnkv3jy

An Algorithm Incarnating Deep Integration of Hardware-Software Energy Regulation Principles for Heterogeneous Green Scheduling

Shaohui Li, Hong Liu, Bin Gong, Jinglian Wang
2020 IEEE Access  
However, on the one hand, there is still considerable space beyond reach of the hardware energy regulation mode; on the other hand, as the core of green software methods, meta-heuristics algorithms are  ...  The experimental results show that compared with the other three metaheuristic scheduling algorithms, GHSA_di algorithm has obvious advantages in overall performance, energy saving and scalability, for  ...  A MULTI-LEVEL PARALLEL ALGORITHM-DESIGN Currently, there are two kinds of parallelism; one is the concurrency of the inherent evolutionary mechanism in the meta-heuristics algorithms, and the other is  ... 
doi:10.1109/access.2020.3003304 fatcat:ufbheyfuzfgxnkeqkhv2dbsc3m

Intelligent scheduling with deep fusion of hardware-software energy-saving principles for greening stochastic nonlinear heterogeneous super-systems

Jinglian Wang, Bin Gong, Hong Liu, Shaohui Li
2019 Applied intelligence (Boston)  
Focusing on deep fusion of hardware-software energy-saving principles, an energy-aware intelligent scheduling model and algorithm are proposed in this paper; throughout the stages of model preparation,  ...  Extensive simulator and simulation experiments highlight obvious superiorities in the proposed scheduler such as higher efficacy and better scalability, which fully considers nonlinear diversities of heterogeneous  ...  mathematical model.This paper focuses on highly intelligent driving-force of two-way integration of hardware and software, and then proposes a multi-objective optimization model and algorithm for adaptive  ... 
doi:10.1007/s10489-019-01424-5 fatcat:pygras6ojja7ppfto66z6vw4nu

A Systematic Survey of General Sparse Matrix-Matrix Multiplication [article]

Jianhua Gao, Weixing Ji, Zhaonian Tan, Yueyan Zhao
2020 arXiv   pre-print
The rationales of different algorithms in each category are analyzed, and a wide range of SpGEMM algorithms are summarized.  ...  Many optimization techniques have been developed for certain application fields and computing architecture over the decades.  ...  They also develop a hash table-based SpGEMM algorithm targeting shared-memory multi-core and many-core processors.  ... 
arXiv:2002.11273v1 fatcat:5ppccisodvaevdhvfhawvjam5q

Optimization of hybrid parallel application execution in heterogeneous high performance computing systems considering execution time and power consumption [article]

Paweł Rościszewski
2018 arXiv   pre-print
New programming models and algorithms that consider this criterion are one of the key areas where significant progress is necessary in order to achieve the goal.  ...  Utilizing full power of such systems requires programming parallel applications that are hybrid in two meanings: they can utilize parallelism on multiple levels at the same time and combine together programming  ...  The capacity of utilizing many cores in parallel depends on the algorithms used by the application.  ... 
arXiv:1809.07611v1 fatcat:f2vl3kmgznckroj6h3uwt2zwf4

Investigation of Parallel Data Processing Using Hybrid High Performance CPU

Paweł Czarnul
2020 Computing and informatics  
The paper investigates parallel data processing in a hybrid CPU + GPU(s) system using multiple CUDA streams for overlapping communication and computations.  ...  Furthermore, using standard memory allocation on a GPU and Unified Memory versions are compared, the latter including programmer added prefetching.  ...  Utilization of CUDA streams for parallel implementation of a genetic algorithm is presented in paper [39] .  ... 
doi:10.31577/cai_2020_3_510 fatcat:etcfpn7gevgi5mujrawp445al4

A Unified Framework For A Robust Conflict-Free Robot Navigation

S. Veera Ragavan, V. Ganapathy
2007 Zenodo  
Many environment specific methods and systems for Robot Navigation exist.  ...  However vast strides in the evolution of navigation technologies and system techniques create the need for a general unified framework that is scalable, modular and dynamic.  ...  models based on Bayesian Networks or uncertainty sets, possibility models based on Fuzzy Logic and Dempster-Shafer theory, or learning algorithms based on Neural Networks and Evolutionary Algorithms,  ... 
doi:10.5281/zenodo.1057317 fatcat:dmvdojgosvbgpk4ghkyx53nbnu

ETP4HPC's Strategic Research Agenda for High-Performance Computing in Europe 4 [article]

Michael Malms, Marcin Ostasz, Maike Gilliot, Pascale Bernier-Bruna, Laurent Cargemel, Estela Suarez, Herbert Cornelius, Marc Duranton, Benny Koren, Pascale Rosse-Laurent, María S. Pérez-Hernández, Manolis Marazakis (+11 others)
2020 Zenodo  
that leads to efficient and scalable programming mo- algorithms.  ...  Furthermore, the reliance of many machine learning algorithms on lower precision arith- metic is making exa-ops (i.e.  ... 
doi:10.5281/zenodo.4605343 fatcat:lcsgbea5dzgdfmj5dkw6pr7vni

TFBN: A Cost Effective High Performance Hierarchical Interconnection Network

M. M. Hafizur Rahman, Mohammed Al-Naeem, Mohammed N. M. Ali, and Abu Sufian
2020 Applied Sciences  
In order to fulfill the increasing demand for computation power to process a boundless data concurrently within a very short time or real-time in many areas such as IoT, AI, machine learning, smart grid  ...  Thus, to have this level of computation, we need a massively parallel computer (MPC) system that shall consist of millions of nodes; and, for the interconnection of these massive numbers of nodes, conventional  ...  This paper focuses on the more detailed study of static network performance for the upper-level TFBN along with the cost-effectiveness analysis.  ... 
doi:10.3390/app10228252 fatcat:z4eowfk7gbgqjgn7ngm45wsiu4

Toward a Power-Efficient Backbone Network: The State of Research

M. Nishan Dharmaweera, Rajendran Parthiban, Y. Ahmet Sekercioglu
2015 IEEE Communications Surveys and Tutorials  
In this paper, we provide a comprehensive survey of the most relevant research efforts on minimizing power consumption of backbone networks.  ...  In a related study [195] , Bonetto et al. developed three different algorithms, namely, Least Flow Algorithm (LFA), Genetic Algorithm (GA), and Energy Watermark Algorithm (EWA) by extending past efforts  ...  Optical labels that perform the task of a header can be sent parallel to the payload on a separate wavelength.  ... 
doi:10.1109/comst.2014.2344734 fatcat:bz4wlu5rfnbqhke4zitiyogud4

A modeler's guide to handle complexity in energy systems optimization [article]

Leander Kotzur, Lars Nolting, Maximilian Hoffmann, Theresa Groß, Andreas Smolenko, Jan Priesmann, Henrik Büsing, Robin Beer, Felix Kullmann, Bismark Singh, Aaron Praktiknjo, Detlef Stolten (+1 others)
2021 arXiv   pre-print
Based on this overview, we develop a guide for modelers who encounter computational limitations.  ...  Thus, we first analyze the determinants of complexity and note that many drivers of complexity could be avoided a priori with a tailored model design.  ...  Therefore, today's supercomputers with a huge amount of cores, e.g., JUWELS with 122,768 cores or Sunway TaihuLight with 10,649,600, are not capable of efficiently tackling this mathematical complexity  ... 
arXiv:2009.07216v3 fatcat:lpl322ounfglperzpnmqbvrfha

Intel�lig�ncia artificial i transpar�ncia algor�tmica : �It's complicated�

2018 BiD: Textos Universitaris de Biblioteconomia i Documentació  
for the Evaluation of Scientific and Technological Choices (OPECST), sponsored by my fellow MPs Claude de Ganay and Dominique Gillot; not to mention the CNIL's outstanding works on the ethics of algorithms  ...  Designing AI that Uses Less EnergyRecent progress in AI has largely been due to the increased use of GPUs, graphics processors, to carry out general-purpose and massively parallel computing.  ...  Most of them have been known for decades and many of the algorithms used today were developed in the '60s and '70s.  ... 
doi:10.1344/bid2018.41.11 fatcat:42khobtbzvcw7h7bzdt6xhrkxe

Exploring Scheduling for On-demand File Systems and Data Management within HPC Environments

Mehmet Soysal
2021
ADA-FS aims to improve I/O performance for highly parallel applications through distributed on-demand file systems.  ...  The traditional HPC applications are primarily compute-intensive, parallel and scalable simulations. It is common practice to reduce data operations as much as possible.  ...  Parallel applications: Some highly-scalable and tightly-coupled scientific applications run many parallel processes with synchronized execution steps.  ... 
doi:10.5445/ir/1000130537/v2 fatcat:q6mruzwe6fg5njkb667kgg57fe

Exploring Scheduling for On-demand File Systems and Data Management within HPC Environments

Mehmet Soysal
2021
ADA-FS aims to improve I/O performance for highly parallel applications through distributed on-demand file systems.  ...  The traditional HPC applications are primarily compute-intensive, parallel and scalable simulations. It is common practice to reduce data operations as much as possible.  ...  Parallel applications: Some highly-scalable and tightly-coupled scientific applications run many parallel processes with synchronized execution steps.  ... 
doi:10.5445/ir/1000130537 fatcat:ux2csrymobegzdrdr6ow2rvahu

Approaches to genome analysis through the application of graph theory

Alice M. Kaye
2021
This thesis presents an alternative perspective on how we can take advantage of new computational methods to enhance the reference genome in the era of widespread sequencing and big data.  ...  To circumvent the above limitations, recent progress has been achieved by using more CPUs simultaneously: multi-core processors, supercomputers and more efficient parallelisation of algorithms.  ...  Next Generation Sequencing The completion of the HGP sparked significant research interest into the development of more scalable sequencing techniques, leading to the emergence of massively parallel sequencing  ... 
doi:10.14288/1.0401888 fatcat:hy7riaj4ajbsbhkcjr3pkftb7a
« Previous Showing results 1 — 15 out of 18 results