16,877 Hits in 5.8 sec

Adaptively scheduling parallel loops in distributed shared-memory systems

Yong Yan, Canming Jin, Xiaodong Zhang
1997 IEEE Transactions on Parallel and Distributed Systems  
Using runtime information of load distributions and processor affinity, we propose an adaptive scheduling algorithm and its variations from different control mechanisms.  ...  The kernel application programs we used for performance evaluation were carefully selected for different classes of parallel loops.  ...  Variable s i is initially set to zero and is increased by one each time processor i finishes the execution of an iteration.  ... 
doi:10.1109/71.569656 fatcat:pmfuc7jl5ff5harpbfnafr42fi

An Improved Adaptive Space-Sharing Scheduling Policy for Non-dedicated Heterogeneous Cluster Systems

Amit Chhabra, Gurvinder Singh
2012 International Journal of Computer Applications Technology and Research  
Moreover existing adaptive policies use First-come-first-serve (FCFS) which is known to be sensitive of variance in service demand, as a job-selection policy for processor allocation.  ...  FCFS allocation of processors to jobs results in a situation where small jobs could be blocked by an earlier arrived large job.  ...  Moreover existing adaptive policies focus on using FCFS as a job selection algorithm to allocate processors to parallel jobs.  ... 
doi:10.7753/ijcat0102.1004 fatcat:mrpmxfpf3nc6hheii5vir7p6wy

An Efficient Adaptive Space-Sharing Policy for Non-dedicated Heterogeneous Cluster Systems

Amit Chhabra, Gurvinder Singh
2012 International Journal of Computer Applications  
This paper extends an existing non-workconserving adaptive space-sharing policy for dedicated heterogeneous systems to non-dedicated heterogeneous cluster systems.  ...  However community-owned clusters are naturally non-dedicated and tend to be heterogeneous over the time as cluster hardware is usually upgraded and new fast machines are also added to improve cluster performance  ...  Acknowledgements A shorter version of this paper appeared in International Journal of Computer Applications Technology and Research, sept.-oct., 2012.  ... 
doi:10.5120/9439-3882 fatcat:73z6miwzhvfotdzhlxpdvd4dgm

Kernel-level scheduling for the nano-threads programming model

Eleftherios D. Polychronopoulos, Xavier Martorell, Dimitrios S. Nikolopoulos, Jesus Labarta, Theodore S. Papatheodorou, Nacho Navarro
1998 Proceedings of the 12th international conference on Supercomputing - ICS '98  
Our approach relies on support for user level threads which are used to exploit parallelism within applications, and a two-level scheduling policy which coordinates the number of resources allocated by  ...  With such diverse workloads, system utilization and throughpuf as well as execution time become important performance metrics.  ...  We use this workload to compare our mechanisms with the complete IRIX parallelization environment.  ... 
doi:10.1145/277830.277911 dblp:conf/ics/PolychronopoulosMNLPN98 fatcat:kia2et2cgjfmrkyb2qjcwk4p6m

Integrated scheduling: the best of both worlds

Jon B. Weissman, Lakshman Rao Abburi, Darin England
2003 Journal of Parallel and Distributed Computing  
It dynamically controls resource allocation among a set of competing applications, but unlike a traditional job scheduler, it can interact directly with an application during execution to optimize resource  ...  The results indicate that iScheduling can improve both waiting time and overall turnaround time substantially for these workload classes, outperforming standard policies such as backfilling and moldable  ...  For example, an application that has 2 resources allocated with its ideal resources = 4 will be much greater impacted if a resource was taken away than an application with 8 resources allocated with its  ... 
doi:10.1016/s0743-7315(03)00013-3 fatcat:dibrw7ua6bbgdewyus3coy4vre

Performance-driven processor allocation

J. Corbalan, X. Martorell, J. Labarta
2005 IEEE Transactions on Parallel and Distributed Systems  
With this aim, we present the SelfAnalyzer, an approach to dynamically analyzing the performance of applications (speedup, efficiency and execution time), and the Performance-Driven Processor Allocation  ...  We perform the processor allocation taking into account the characteristics of the application measured at run-time.  ...  , resulting in an increase in the execution time.  ... 
doi:10.1109/tpds.2005.85 fatcat:jcfurhui7zbkxltwc6ewgoofoy

A Tool to Schedule Parallel Applications on Multiprocessors: The NANOS CPU Manager [chapter]

Xavier Martorell, Julita Corbalán, Dimitrios S. Nikolopoulos, Nacho Navarro, Eleftherios D. Polychronopoulos, Theodore S., JesÚs Labarta
2000 Lecture Notes in Computer Science  
For instance, we show through several experiments that benefits in execution time range from 15% to 100%, depending on the policy used and the characteristics of the workload.  ...  Scheduling parallel applications on shared-memory multiprocessors is a difficult task that requires a lot of tuning from application programmers, as well as operating system developers and system managers  ...  Synchronization and Processor Preemptions Each time an application needs to do some operation which depends on the number of running processors, it uses the number of processors allocated pro-vided by  ... 
doi:10.1007/3-540-39997-6_7 fatcat:rwrcljlwdraqvcj5m7ucassc5u

Dynamic versus adaptive processor allocation policies for message passing parallel computers: An empirical comparison [chapter]

Jitendra Padhye, Lawrence Dowdy
1996 Lecture Notes in Computer Science  
An adaptive policy may consider the state of the system at arrival time but it does not allow preemption of any of the running jobs.  ...  The workload model is based on matrix computation applications commonly found on large systems used for scienti c programming. Results are reported for single and multiclass cases.  ...  Both policies are implemented on a parallel computer with a message passing architecture. An open system model of the workload ow is implemented.  ... 
doi:10.1007/bfb0022296 fatcat:5fwp7djijnc5vij4oiuq4vdng4

Dynamic Parallel Job Scheduling in Multi-cluster Computing Systems [chapter]

J. H. Abawajy
2004 Lecture Notes in Computer Science  
In this paper, we address an on-line parallel job scheduling problem in heterogeneous multi-cluster computing systems.  ...  Job scheduling is a complex problem, yet it is fundamental to sustaining and improving the performance of parallel processing systems.  ...  This is achieved by varying the number of processors allocated to an application during its execution.  ... 
doi:10.1007/978-3-540-24685-5_4 fatcat:53foztfkibalzd7skmvdyvsdja

Design considerations for network processor operating systems

Tilman Wolf, Ning Weng, Chia-Hui Tai
2005 Proceedings of the 2005 symposium on Architecture for networking and communications systems - ANCS '05  
In particular, we explore the performance impact of (1) application analysis for partitioning, (2) network traffic characterization, (3) workload mapping, and (4) run-time adaptation.  ...  To make full use of the capabilities of network processors, it is imperative to provide the ability to dynamically adapt to changing traffic patterns and to provide run-time support in the form of a network  ...  Heavily used applications typically need to be replicated multiple times to provide sufficient performance (e.g., multiple parallel IP forwarding applications).  ... 
doi:10.1145/1095890.1095901 dblp:conf/ancs/WolfWT05 fatcat:jjdsrip2lbcrzizrfxmmy6742a

Parallel Application Scheduling on Networks of Workstations

Stergios V. Anastasiadis, Kenneth C. Sevcik
1997 Journal of Parallel and Distributed Computing  
Parallel applications can be executed using the idle computing capacity of workstation clusters.  ...  We conclude that AEP (2) is suitable for use in scheduling parallel applications on networks of workstations.  ...  An important question that remains is how to allocate processors among parallel applications that execute in a multiprogrammed environment.  ... 
doi:10.1006/jpdc.1997.1335 fatcat:o3oi7qg57zd2fftwp3vz762t6i

Runtime Support for Multicore Packet Processing Systems

Tilman Wolf, Ning Weng, Chia-hui Tai
2007 IEEE Network  
In particular, we explore the performance impact of application analysis on partitioning, traffic characterization, workload mapping, and runtime adaptation.  ...  To make full use of the capabilities of network processors, it is imperative to provide the ability to dynamically adapt to changing traffic patterns in the form of a network processor runtime system.  ...  This will bring us closer to realizing network processors as easy-to-use components of network systems.  ... 
doi:10.1109/mnet.2007.386467 fatcat:qxbpwg4ajvfhbggqpyldm26hfq

Multiprocessor scheduling for high-variability service time distributions [chapter]

Eric W. Parsons, Kenneth C. Sevcik
1995 Lecture Notes in Computer Science  
Many disciplines have been proposed for scheduling and processor allocation in multiprogrammed multiprocessors for parallel processing.  ...  These have been, for the most part, designed and evaluated for workloads having relatively low variability in service demand.  ...  job to adapt to changes in processor allocation [ZM90, MVZ93] .  ... 
doi:10.1007/3-540-60153-8_26 fatcat:m5jyfvrh7zeuxnpoftecgxuxiy

Allocation strategies for utilization of space-shared resources in Bag of Tasks grids

César A.F. De Rose, Tiago Ferreto, Rodrigo N. Calheiros, Walfredo Cirne, Lauro B. Costa, Daniel Fireman
2008 Future generations computer systems  
This paper presents the Explicit Allocation Strategy, in which an adaptor automatically fits grid requests to the resource in order to decrease turn-around time of the application.  ...  Using space-shared resources in grids is not straightforward since they require jobs to a priori specify some parameters, such as allocation time and amount of processors. Current solutions (e.g.  ...  They are traces of real machines and were obtained from Parallel Workload archive [34] . We filtered out jobs with missing request time. The workloads used are described in Table 1 .  ... 
doi:10.1016/j.future.2007.05.005 fatcat:6akeiwlxbzhbnix3yysnz72fnq

Analysis of non-work-conserving processor partitioning policies [chapter]

E. Rosti, E. Smirni, G. Serazzi, L. W. Dowdy
1995 Lecture Notes in Computer Science  
Variability in the workload behavior including bursty arrivals, a high coe cient of variation in the workload execution time, unstable systems with processor failures are among the situations where non-work-conserving  ...  One policy class keeps a certain percentage of the processors free. The other policy class makes an allocation decision based on previously observed system behavior.  ...  However, multiprogramming complicates scheduling since an allocation policy is needed to determine the number of processors which should be allocated to each parallel program.  ... 
doi:10.1007/3-540-60153-8_28 fatcat:ctabzjqczzeuzjw4z2bpcpmdzi
« Previous Showing results 1 — 15 out of 16,877 results