Filters








1,821 Hits in 7.4 sec

On characterizing the data movement complexity of computational DAGs for parallel execution

Venmugil Elango, Fabrice Rastello, Louis-Noël Pouchet, J. Ramanujam, P. Sadayappan
2014 Proceedings of the 26th ACM symposium on Parallelism in algorithms and architectures - SPAA '14  
In this paper, we develop an extension of the well-known redblue pebble game to develop lower bounds on the data movement complexity for the parallel execution of computational directed acyclic graphs  ...  We demonstrate the use of the methodology by analyzing the CDAGs of several numerical algorithms, to develop lower bounds on data movement for their parallel execution.  ...  Department of Energy through award DE-SC0008844, and by the U.S. Army through contract W911NF-10-1-000.  ... 
doi:10.1145/2612669.2612694 dblp:conf/spaa/ElangoRPRS14 fatcat:s2fumsvmyba2rmqg3gb43ta5nm

On Characterizing the Data Movement Complexity of Computational DAGs for Parallel Execution [article]

Venmugil Elango, Fabrice Rastello (INRIA Grenoble Rhône-Alpes), Louis-Noël Pouchet, P. Sadayappan
2014 arXiv   pre-print
In this paper, we develop an extension of the well-known red-blue pebble game to develop lower bounds on the data movement complexity for the parallel execution of computational directed acyclic graphs  ...  We demonstrate the use of the methodology by analyzing the CDAGs of several numerical algorithms, to develop lower bounds on data movement for their parallel execution.  ...  the Data Movement Complexity of Computational DAGs for Parallel Execution11 Figure 1 : Distributed-memory system R1 (Input) A level-L pebble, R i L can be placed on any vertex that has a blue pebble;  ... 
arXiv:1404.4767v1 fatcat:wdl56go5znb2jk3c5jr4o6i6za

A taxonomy of scientific workflow systems for grid computing

Jia Yu, Rajkumar Buyya
2005 SIGMOD record  
In this paper, we propose a taxonomy that characterizes and classifies various approaches for building and executing workflows on Grids.  ...  Therefore, many efforts have been made towards the development of workflow management systems for Grid computing.  ...  TAXONOMY The taxonomy characterizes and classifies approaches of scientific workflow systems in the context of Grid computing.  ... 
doi:10.1145/1084805.1084814 fatcat:fs6bc7qtljay3p5dbjqlpgnfki

Data-Intensive Workflow Optimization Based on Application Task Graph Partitioning in Heterogeneous Computing Systems

Saima Gulzar Ahmad, Chee Sun Liew, M. Mustafa Rafique, Ehsan Ullah Munir, Samee U. Khan
2014 2014 IEEE Fourth International Conference on Big Data and Cloud Computing  
Data-intensive applications involve movement of huge amount of data between execution nodes that incurs large costs. Data-streaming model improves the execution performance of such applications.  ...  Optimization of these performance metrics in heterogeneous computing environment becomes more challenging due to the difference in the computing capacity of execution nodes and variations in the data transfer  ...  ACKNOWLEDGMENT The work presented in this paper is supported by the Ministry of Education Malaysia (FRGS FP051-2013A and UMRG RP001F-13ICT).  ... 
doi:10.1109/bdcloud.2014.63 dblp:conf/bdcloud/AhmadLRMK14 fatcat:q4vegly4tbg5joh6po3miz3j2e

PathTracing: Raising the Level of Understanding of Processing Latency in Heterogeneous MPSoCs

Claudion Rubattu, Francesca Palumbo, Shuvra Bhattacharyya, Maxime Pelcat
2021 Proceedings of the 2021 Drone Systems Engineering and Rapid Simulation and Performance Evaluation: Methods and Tools Proceedings  
However, the complexity of heterogeneous Multiprocessor Systems-on-Chips (MPSoCs) makes it difficult to provide early evaluation of system execution latency when executing parallel applications.  ...  For that purpose, we propose the concept of Jaccard Gantt similarity score and demonstrate that the straightforward method consisting in scheduling a Directed Acyclic Graph (DAG) of tasks, each with a  ...  Palumbo is grateful to the University of Sassari supporting her research activity through the "fondo di Ateneo per la ricerca 2019".  ... 
doi:10.1145/3444950.3447282 fatcat:a3rhkuehcvdlrirbd76f4rljj4

SignalPU: A Programming Model for DSP Applications on Parallel and Heterogeneous Clusters

Farouk Mansouri, Sylvain Huet, Dominique Houzet
2014 2014 IEEE Intl Conf on High Performance Computing and Communications, 2014 IEEE 6th Intl Symp on Cyberspace Safety and Security, 2014 IEEE 11th Intl Conf on Embedded Software and Syst (HPCC,CSS,ICESS)  
Consequently, a high requirement of computing power characterize them.  ...  In the present work, we are characterizing DSP applications and propose a programming model based on their distinctiveness in order to implement them easily and efficiently on heterogeneous clusters.  ...  OmpSS [11] : Is another variant of OpenMP extended to support asynchrony, heterogeneity and data movement for task parallelism.  ... 
doi:10.1109/hpcc.2014.144 dblp:conf/hpcc/MansouriHH14 fatcat:pnsbyhpagvb2piyciurdr2c7tm

A Taxonomy of Workflow Management Systems for Grid Computing [article]

Jia Yu, Rajkumar Buyya
2005 arXiv   pre-print
In this paper, we propose a taxonomy that characterizes and classifies various approaches for building and executing workflows on Grids.  ...  With the advent of Grid and application technologies, scientists and engineers are building more and more complex applications to manage and process large data sets, and execute scientific experiments  ...  ACKNOWLEDGEMENTS We would like to acknowledge all developers of the workflow management systems described in the paper.  ... 
arXiv:cs/0503025v2 fatcat:uadtgvc3bvdpxblsxaepcbjkky

Cost models for geo-distributed massively parallel streaming analytics [article]

Anna-Valentini Michailidou, Anastasios Gounaris, Konstantinos Tsichlas
2021 arXiv   pre-print
in compute nodes, (2) geo-distribution, (3) massive parallelism, (4) complex DAGs and (5) streaming applications.  ...  This report is part of the DataflowOpt project on optimization of modern dataflows and aims to introduce a data quality-aware cost model that covers the following aspects in combination: (1) heterogeneity  ...  However, the operator execution is not parallelized, i.e., each DAG node is placed on exactly one compute node.  ... 
arXiv:2105.12507v1 fatcat:su5qm2arerf2ddadfm6ui6j5gm

Grid Workflow Modeling for Remote Sensing Retrieval Service with Tight Coupling [chapter]

Jianwen Ai, Yong Xue, Jie Guang, Yingjie Li, Ying Wang, Linyan Bai
2009 Lecture Notes in Computer Science  
In this paper we mainly discusses some application cases based on Grid computing for Geosciences and the application limit of Grid in remote sensing, and gives the method of Grid Workflow modeling for  ...  Therefore, we need a interactive graphical tool to present the executing relationships of algorithms and to generate automatically the corresponding submitted description files for grid platform.  ...  Aiming at largeness and instantaneousness of remote sensing data, we adopt tasks acting as data movement or computing code movement according to the schedule arithmetic of engine.  ... 
doi:10.1007/978-3-642-01973-9_39 fatcat:ibpdt2xg5fcwjibxutlvfzpska

Automating environmental computing applications with scientific workflows

Rafael Ferreira da Silva, Ewa Deelman, Rosa Filgueira, Karan Vahi, Mats Rynge, Rajiv Mayani, Benjamin Mayer
2016 2016 IEEE 12th International Conference on e-Science (e-Science)  
In order to cope with the needs of such applications, computational methods and technologies have emerged to support the execution of these applications on heterogeneous, distributed systems.  ...  Computational environmental science applications have evolved and become more complex over the last decade.  ...  ACKNOWLEDGEMENTS This work was funded by DOE under the contract number #DESC0012636, "Panorama-Predictive Modeling and Diagnostic Monitoring of Extreme Science Workflows"; and by the National Science Foundation  ... 
doi:10.1109/escience.2016.7870926 dblp:conf/eScience/SilvaDFVRMM16 fatcat:f7wumsedq5artf7u63wbv6g4pm

DeepCPU: Serving RNN-based Deep Learning Models 10x Faster

Minjia Zhang, Samyam Rajbhandari, Wenhan Wang, Yuxiong He
2018 USENIX Annual Technical Conference  
We build DeepCPU, a fast serving library on CPUs, to integrate these optimizations for efficient RNN computation.  ...  This work characterizes RNN performance and identifies low data reuse as a root cause.  ...  We thank our shepherd, Deniz Altinbuken, for helping improving the presentation quality of the work.  ... 
dblp:conf/usenix/ZhangRWH18 fatcat:3x7esz5fkvh3dc66yv2fsatc2m

A Visual Programming Model to Implement Coarse-Grained DSP Applications on Parallel and Heterogeneous Clusters [chapter]

Farouk Mansouri, Sylvain Huet, Dominique Houzet
2014 Lecture Notes in Computer Science  
The digital signal processing (DSP) applications are one of the biggest consumers of computing. They process a big data volume which is represented with a high accuracy.  ...  In the present work, we characterize DSP applications, and based on their distinctiveness, we propose a high level visual programming model and an execution model in order to drop down their implementations  ...  Also, according to the data kind and the algorithm of each kernel, it's interesting in most of cases to offload the execution of certain of kernels on a massively parallel computation unit (accelerators  ... 
doi:10.1007/978-3-319-14325-5_13 fatcat:ubbpa4r6jrdltgqczo2j4tfmqa

StencilFlow: Mapping Large Stencil Programs to Distributed Spatial Computing Systems [article]

Johannes de Fine Licht, Andreas Kuster, Tiziano De Matteis, Tal Ben-Nun, Dominic Hofer, Torsten Hoefler
2021 arXiv   pre-print
Spatial computing devices have been shown to significantly accelerate stencil computations, but have so far relied on unrolling the iterative dimension of a single stencil operation to increase temporal  ...  for stencil programs on FPGAs to date.  ...  ACKNOWLEDGEMENTS We authors wish to thank Tobias Kenter, Christian Plessl, and the Paderborn Center for Parallel Computing (PC 2 ) for generously providing support and compute hours on the Noctua FPGA  ... 
arXiv:2010.15218v2 fatcat:kev77kpks5abhjjgosqidjfdki

A Taxonomy of Workflow Management Systems for Grid Computing

Jia Yu, Rajkumar Buyya
2005 Journal of Grid Computing  
In this paper, we propose a taxonomy that characterizes and classifies various approaches for building and executing workflows on Grids.  ...  With the advent of Grid and application technologies, scientists and engineers are building more and more complex applications to manage and process large data sets, and execute scientific experiments  ...  ACKNOWLEDGEMENTS We would like to acknowledge all developers of the workflow management systems described in the paper.  ... 
doi:10.1007/s10723-005-9010-8 fatcat:b2ptnwxnhbeltlybivpxl6m654

In-situ feature-based objects tracking for data-intensive scientific and enterprise analytics workflows

Solomon Lasluisa, Fan Zhang, Tong Jin, Ivan Rodero, Hoang Bui, Manish Parashar
2014 Cluster Computing  
Emerging scientific simulations on leadership class systems are generating huge amounts of data and processing this data in an efficient and timely manner is critical for generating insights from the simulations  ...  In this paper, we investigate an alternate approach that aims to bring the analytics closer to the data using in-situ execution of data analysis operations.  ...  One attractive approach for reducing data movement is to perform runtime data analytics in-situ, directly on the compute node where data is generated, so as to maximize intranode data sharing.  ... 
doi:10.1007/s10586-014-0396-6 fatcat:rnlbq3djgvd7jivd6mhlv6q5by
« Previous Showing results 1 — 15 out of 1,821 results