Filters








9,409 Hits in 7.3 sec

Preparing Nuclear Astrophysics for Exascale [article]

Max P. Katz, Ann Almgren, Maria Barrios Sazo, Kiran Eiden, Kevin Gott, Alice Harpole, Jean M. Sexton, Don E. Willcox, Weiqun Zhang, Michael Zingale
2020 arXiv   pre-print
Examining these nuclear burning processes using high resolution simulations is critical for understanding how these astrophysical explosions occur.  ...  Astrophysical explosions such as supernovae are fascinating events that require sophisticated algorithms and substantial computational power to model.  ...  A system like Summit has > 95% of its peak theoretical computational throughput in GPUs, so one might think that attempting to use the CPUs effectively is fruitless, and this is probably true for a problem  ... 
arXiv:2007.05218v1 fatcat:x32sereeenf5niid65px6i5cci

Research of Data Storage and Querying Methods Based on Ring Distributed Hash

Ye Chen
2015 Open Automation and Control Systems Journal  
Second, the research of data storage and search method based on circular distributed hash. This thesis adopts distributed hash table and Distributed Hash ring to distributed file system.  ...  According to the saved metadata, we take use of binary search method to find the location of data. Third, the research of data migration based on circular distributed hash.  ...  These two methods solve the locality dependent problem of the current throughput-improving methods, and can process the nontraditional backup loads effectively.  ... 
doi:10.2174/1874444301507011203 fatcat:zgierrp72nebnd4lc3mx7vph6a

CNNLab: a Novel Parallel Framework for Neural Networks using GPU and FPGA-a Practical Study with Trade-off Analysis [article]

Maohua Zhu, Liu Liu, Chao Wang, Yuan Xie
2016 arXiv   pre-print
Moreover, we analyze the detailed quantitative performance, throughput, power, energy, and performance density for both approaches.  ...  Existing high-level parallel abstractions like MapReduce are insufficiently expressive while low-level tools like MPI and Pthreads leave ML experts repeatedly solving the same design challenges.  ...  In general, deep learning uses a multi-layer neural network model to extract high-level features into a combination of low-level abstractions to find the distributed data features, to solve complex problems  ... 
arXiv:1606.06234v1 fatcat:en7acoahonb7beqrnxv553g46e

Online Credit Card Fraud Detection: A Hybrid Framework with Big Data Technologies

You Dai, Jin Yan, Xiaoxin Tang, Han Zhao, Minyi Guo
2016 2016 IEEE Trustcom/BigDataSE/ISPA  
We further implement the workflow with a new framework which consists of four layers: distributed storage layer, batch training layer, key-value sharing layer and streaming detection layer.  ...  models to improve accuracy; 2) the ability to process large amount of data and 3) the ability to do the detection in real time.  ...  The authors would like to thank the anonymous reviewers for their valuable comments. This work is partially sponsored by the National Basic Research 973 Program of China (No.  ... 
doi:10.1109/trustcom.2016.0253 dblp:conf/trustcom/DaiYTZG16 fatcat:hv6kxgnperehblz4vlvx7dzfzq

Evaluation Model Queuing Task Scheduling Based on Hybrid Architecture Cloud Systems

Zeyu Sun, Yaping Li, Yangjie Cao, Yuanbo Li
2016 International Journal of Grid and Distributed Computing  
In addition, consider single backup task status, for the failure of more than one processor at the same time, present the minimum cost of backup scheduling algorithm, the algorithm to solve the problem  ...  The algorithm solves the problem that meeting customer service satisfaction and load balancing at the same time.  ...  Experimental results show that the proposed algorithm can solve the problem of large-scale computing in the cloud.  ... 
doi:10.14257/ijgdc.2016.9.6.17 fatcat:l7xd2f7jufglxaeace3ynph35a

A Hierarchical Multilayer Service Composition Model for Global Virtual Organizations

Abiud Wakhanu Mulongo, Elisha T. Opiyo Omulo, William Okello Odongo
2015 Computer Science and Information Technology  
MCDM global planning methods on the other hand suffer exponential state space explosion making them severely limited for large problems of industrial relevance.  ...  These web services can be differentiated on a high dimensionality of quality of service attributes.  ...  At a high level, in HMSCM, we map the composite service selection problem to NUM [32] problem as follows. 1.  ... 
doi:10.13189/csit.2015.030401 fatcat:6gwjljwshfgilpw34qyol5u5me

Large-Scale Parallel Computing on Grids

Henri Bal, Kees Verstoep
2008 Electronical Notes in Theoretical Computer Science  
This paper argues that computational grids can be used for far more types of applications than just trivially parallel ones.  ...  Algorithmic optimizations like latency-hiding and exploiting locality can be used effectively to obtain high performance on grids, despite the relatively slow wide-area networks that connect the grid resources  ...  The various DAS systems have been co-funded by the Netherlands Organization for Scientific Research (N.W.O.), the Netherlands National Computing Facilities foundation (N.C.F.), the Virtual Laboratory for  ... 
doi:10.1016/j.entcs.2008.11.010 fatcat:esb4axh6zberfdlf2m5y6w3rpe

Cutting Throughput on the Edge:App-Aware Placement in Fog Computing [article]

Francescomaria Faticanti, Francesco De Pellegrini, Domenico Siracusa, Daniele Santoro, Silvio Cretti
2018 arXiv   pre-print
Due to the complexity of the original problem, we resort to a simplified version, which is further solved using a greedy algorithm.  ...  By displacing workloads from the central cloud to the edge devices, fog computing overcomes communication bottlenecks avoiding raw data transfer to the central cloud, thus paving the way for the next generation  ...  INTRODUCTION Fog computing adopts cloud technology to move computation to the edge. It promises to solve the core problem of data explosion in the IoT domain [1] .  ... 
arXiv:1810.04442v1 fatcat:ryncrgsakncjpozxo4o4soknfq

Matrix-free approaches for GPU acceleration of a high-order finite element hydrodynamics application using MFEM, Umpire, and RAJA [article]

Arturo Vargas, Thomas M. Stitt, Kenneth Weiss, Vladimir Z. Tomov, Jean-Sylvain Camier, Tzanio Kolev, Robert N. Rieben
2021 arXiv   pre-print
With the introduction of advanced heterogeneous computing architectures based on GPU accelerators, large-scale production codes have had to rethink their numerical algorithms and incorporate new programming  ...  models and memory management strategies in order to run efficiently on the latest supercomputers.  ...  We would like to thank Veselin Dobrev for support on the integration of MFEM GPU capabilities into MARBL, and the RAJA & Umpire teams for their feedback in software integration and extensions.  ... 
arXiv:2112.07075v1 fatcat:itdrddooobfqjnuw23vdftpwz4

Dynamic Access and Power Control Scheme for Interference Mitigation in Femtocell Networks

2015 KSII Transactions on Internet and Information Systems  
To solve this problem, we propose a joint access and power control scheme that requires limited information exchange between the femto and macro networks.  ...  Through extensive simulations, we show that the proposed scheme outperforms earlier works in terms of the throughput and outage probability.  ...  A binary integer program used to solve the Knapsack problem has computational complexity of order 1 (2 )  m O , where m is the number of neighboring FBSs close to the MUE and 1 is added to count for the  ... 
doi:10.3837/tiis.2015.11.004 fatcat:aopcvsu32ra4thulawyitb2gqy

Performance Evaluation of Distributed Deep Learning Frameworks in Cloud Environment

Shuen-Tai Wang, Fang-An Kuo, Chau-Yi Chou, Yu-Bin Fang
2019 Zenodo  
2016 has become the year of the Artificial Intelligence explosion.  ...  At the present time, deep learning frameworks have been widely deployed on servers for deep learning applications in both academia and industry.  ...  In the meanwhile, TWCC not only supports large numbers of nodes for high-speed parallel computing across nodes, but also employs the latest container virtualization techniques [17] for GPGPU (General  ... 
doi:10.5281/zenodo.3455640 fatcat:ssepbbdbzjbltpph2tjiafn5xy

Parallel DFA Architecture for Ultra High Throughput DFA-Based Pattern Matching

Yi TANG, Junchen JIANG, Xiaofei WANG, Chengchen HU, Bin LIU, Zhijia CHEN
2010 IEICE transactions on information and systems  
To this end, Deterministic Finite Automaton (DFA) is widely used for multi-regex matching, but existing DFA-based researches have claimed high throughput at an expense of extremely high memory cost, so  ...  fail to be employed in devices such as high-speed routers and embedded systems where the available memory is quite limited.  ...  We regard the above mentioned two problems to two aspects respectively, where the transition assignment problem as well as the state allocation is decided in pre-computing phase while scheduling directly  ... 
doi:10.1587/transinf.e93.d.3232 fatcat:h6bnaoq5xvdcthiecvpnzi3a44

Replica Scheduling Strategy for Streaming Data Mining

Shufan Li, Siyuan Yu, Fang Xiao
2022 International Journal of Advanced Computer Science and Applications  
In a distributed storage and computing framework, traditional streaming data mining techniques are inefficient when processing massive amounts of data.  ...  saved by about 40-50% during parallel mining of streaming data, and the throughput rate is increased by 20% to 30%.  ...  We aim to extend the RepEM strategy to the use of heterogeneous clusters by taking both the data copy and memory resource as resources and use Lattice point method to solve the problem.  ... 
doi:10.14569/ijacsa.2022.0130503 fatcat:oxoljs5mnfcexdzeplu7ipyqdy

Steady-state performance evaluation of continuous mono-T-semiflow Petri nets

Jorge Júlvez, Laura Recalde, Manuel Silva
2005 Automatica  
A way to face this state explosion problem consists of relaxing the system model, for example by converting it to a continuous one.  ...  The main contribution of the paper lies in the computation of throughput bounds for continuous Petri net systems with a single T-semiflow. For that purpose, a branch and bound algorithm is designed.  ...  Conclusions Continuous Petri nets were introduced in order to overcome the state explosion problem of high traffic or highly populated discrete systems.  ... 
doi:10.1016/j.automatica.2004.11.007 fatcat:kbfcj7bhcnd3jofjmm2dqyoy6m

An Parallel FPGA SAT Solver Based on Multi‐Thread and Pipeline

LI Tiejun, MA Kefan, ZHANG Jianmin
2021 Chinese journal of electronics  
The Boolean Satisfiability (SAT) problem is the key problem in computer theory and application. A parallel multi-thread SAT solver named pprobSAT+ on a configurable hardware is proposed.  ...  In order to improve the working frequency and throughput of the SAT solver, the deep pipeline strategy is adopted.  ...  This method has high performance in solving the difficult small-scale problems. In recent years, there have been many solvers based on Compute unified device architecture (CUDA) [19−23] .  ... 
doi:10.1049/cje.2021.08.001 fatcat:qb7otyspfzhipkpamrmgwp2ep4
« Previous Showing results 1 — 15 out of 9,409 results