Filters








4,877 Hits in 7.5 sec

Asymptotic Behavior and Performance Constraints of Replication Policies

Davide Cerotti, Marco Gribaudo, Pietro Piazzolla, Giuseppe Serazzi
2015 Electronical Notes in Theoretical Computer Science  
In this paper we propose a simple replication policy to match thresholds on the system response times.  ...  Spreading the workload among a pool of replicated servers is a technique typically adopted to reduce the response time and increase the throughput in complex systems.  ...  In general, the response time of a job depends on the scheduling policy implemented by the task coordinator.  ... 
doi:10.1016/j.entcs.2014.12.012 fatcat:5q6veq6a5zaojldrxiartsoj7u

Stochastic bounds in Fork–Join queueing systems under full and partial mapping

Amr Rizk, Felix Poloczek, Florin Ciucu
2016 Queueing systems  
In a Fork-Join (FJ) queueing system an upstream fork station splits incoming jobs into N tasks to be further processed by N parallel servers, each with its own queue; the response time of one job is determined  ...  , at a downstream join station, by the maximum of the corresponding tasks' response times.  ...  Similarly, the response times of jobs, i.e., the times until the last corresponding tasks have finished their executions, are defined as r 1 = max n x n,1 for the first job, and for j ≥ 2 as r j = max  ... 
doi:10.1007/s11134-016-9486-x fatcat:mi4pdqmhanbxvmwu2j6i3toztm

Using Straggler Replication to Reduce Latency in Large-scale Parallel Computing

Da Wang, Gauri Joshi, Gregory Wornell
2015 Performance Evaluation Review  
One way to combat the variability in machine response time is to add replicas of straggling tasks and wait for one copy to finish.  ...  In cloud computing jobs consisting of many tasks run in parallel, the tasks on the slowest machines (straggling tasks) become the bottleneck in the completion of the job.  ...  Although we focus on single-fork policies in this paper, the analysis can be generalized to multi-fork policies, where replicas of straggling tasks can be launched multiple times during the job execution  ... 
doi:10.1145/2847220.2847223 fatcat:woslto5bnrd3xa7mmrfz5yh7ja

A Parallel Algorithm for EDF-Schedulability Analysis of Multi-modal Real-Time Systems

Masud Ahmed, Nathan Fisher, Daniel Grosu
2012 2012 IEEE International Conference on Embedded and Real-Time Computing Systems and Applications  
Recently, researchers have proposed serial schedulability analysis algorithms for multi-modal systems that account for mode changes at both software level (e.g., changing the set of executing tasks) and  ...  Experimental results show that the runtime of our parallel algorithm is very low even for systems with large number of modes, making it a tractable choice for design-space exploration of real-time multi-modal  ...  has been supported in part by an NSF-CAREER Grant (CNS-0953585), an NSF-CSR Grant (CNS-1116787), an NSF-IGERT Grant (DGE-0654014), and funding from Wayne State University's Office of the Vice President for  ... 
doi:10.1109/rtcsa.2012.49 dblp:conf/rtcsa/AhmedFG12 fatcat:yyy2v2khhfcipmqqpnbxjcukuu

Computable Bounds in Fork-Join Queueing Systems

Amr Rizk, Felix Poloczek, Florin Ciucu
2015 Performance Evaluation Review  
In a Fork-Join (FJ) queueing system an upstream fork station splits incoming jobs into N tasks to be further processed by N parallel servers, each with its own queue; the response time of one job is determined  ...  , at a downstream join station, by the maximum of the corresponding tasks' response times.  ...  Similarly, the response times of jobs, i.e., the times until the last corresponding tasks have finished their executions, are defined as r1 = maxn xn,1 for the first job, and for j ≥ 2 as rj = max 0≤k≤  ... 
doi:10.1145/2796314.2745859 fatcat:5jswvwenm5f3zc4eq2u4gixfcu

Computable Bounds in Fork-Join Queueing Systems

Amr Rizk, Felix Poloczek, Florin Ciucu
2015 Proceedings of the 2015 ACM SIGMETRICS International Conference on Measurement and Modeling of Computer Systems - SIGMETRICS '15  
In a Fork-Join (FJ) queueing system an upstream fork station splits incoming jobs into N tasks to be further processed by N parallel servers, each with its own queue; the response time of one job is determined  ...  , at a downstream join station, by the maximum of the corresponding tasks' response times.  ...  Similarly, the response times of jobs, i.e., the times until the last corresponding tasks have finished their executions, are defined as r1 = maxn xn,1 for the first job, and for j ≥ 2 as rj = max 0≤k≤  ... 
doi:10.1145/2745844.2745859 dblp:conf/sigmetrics/RizkPC15 fatcat:oauaewv4mjhnlmfeaghe4wj7gu

Fine-grained multiprocessor real-time locking with improved blocking

Bryan C. Ward, James H. Anderson
2013 Proceedings of the 21st International conference on Real-Time Networks and Systems - RTNS '13  
These sources of blocking stem from system overheads, varying critical section lengths, and a lack of support for replicated resources.  ...  Existing multiprocessor real-time locking protocols that support nesting are subject to adverse blocking that can be avoided when additional resource-usage-pattern information is known.  ...  However, even if more pessimistic blocking analysis is applied, the increased parallelism afforded by fine-grained lock nesting allows for improved response times in practice.  ... 
doi:10.1145/2516821.2516843 dblp:conf/rtns/WardA13 fatcat:sbxh25drrrdbngnhthlna66hoa

SCHEDULING ON LARGE SCALE DISTRIBUTED PLATFORMS: FROM MODELS TO IMPLEMENTATIONS

PIERRE-FRANÇOIS DUTOT, LIONEL EYRAUD, GRÉGORY MOUNIÉ, DENIS TRYSTRAM
2005 International Journal of Foundations of Computer Science  
For this kind of jobs, the theory of asymptotic behavior shows that optimal solutions can be computed in polynomial time.  ...  Minimizing the number of late tasks may produce starvation by delaying forever one particular task, thus it should be used with great care (for instance in a multi-criteria analysis, together with another  ... 
doi:10.1142/s0129054105002954 fatcat:pwjhqxesy5bgdo37mr4zk6adwq

Multiprocessor Real-Time Locking Protocols: A Systematic Review [article]

Björn B. Brandenburg
2019 arXiv   pre-print
A special focus is placed on the suspension-oblivious and suspension-aware analysis approaches for semaphore protocols, their respective notions of priority inversion, optimality criteria, lower bounds  ...  on maximum priority-inversion blocking, and matching asymptotically optimal locking protocols.  ...  response-time analysis equation for the PCP and SRP.  ... 
arXiv:1909.09600v1 fatcat:tmqcpiuxfbbd5jrcecgvoeanpm

On the Analysis of Parallel Real-Time Tasks with Spin Locks [article]

Xu Jiang, Nan Guan, He Du, Weichen Liu, Wang Yi
2020 arXiv   pre-print
Although the design and analysis of locking protocols have been intensively studied for sequential real-time tasks, there has been little work on this topic for parallel real-time tasks.  ...  In this paper, we study the analysis of parallel real-time tasks using spin locks to protect accesses to shared resources in three commonly used request serving orders (unordered, FIFO-order and priority-order  ...  The response time R(J) of a job J is R(J) = f (J)−r(J), and the worst-case response time (WCRT) R i of task τ i is the maximum R(J) among all its released jobs J.  ... 
arXiv:2003.08233v1 fatcat:dsccshqegjcazhzug67cuxu7iy

Coordination mechanisms for selfish multi-organization scheduling

Johanne Cohen, Daniel Cordeiro, Denis Trystram, Frederic Wagner
2011 2011 18th International Conference on High Performance Computing  
We conduct a game theoretic analysis on the problem of scheduling jobs on computing platforms composed of several independent and selfish organizations, known as the Multi-Organization Scheduling Problem  ...  We modeled MOSP as a non-cooperative game where each agent is responsible for assigning all jobs belonging to a particular organization to the available processors.  ...  We are considering workloads of sequential jobs (also called bags-of-tasks), which are one of the most common computational models for the target parallel platforms.  ... 
doi:10.1109/hipc.2011.6152720 dblp:conf/hipc/CohenCTW11 fatcat:t7b7ptynmjatpg2j6mgs4nlrrm

Modeling and analysis of dynamic coscheduling in parallel and distributed environments

Mark S. Squillante, Yanyong Zhang, Anand Sivasubramaniam, Natarajan Gautam, Hubertus Franke, Jose Moreira
2002 Proceedings of the 2002 ACM SIGMETRICS international conference on Measurement and modeling of computer systems - SIGMETRICS '02  
Our mathematical model and analysis is then used to explore several fundamental design and performance tradeoffs associated with the class of dynamic coscheduling policies across a broad spectrum of parallel  ...  We derive a matrix-analytic analysis based on a stochastic decomposition and a fixed-point iteration.  ...  In addition to mean job response time and maximum system throughput measures, our analysis provides detailed statis-tics (e.g., time spent in certain system/application states) that can be used to explain  ... 
doi:10.1145/511334.511341 dblp:conf/sigmetrics/SquillanteZSGFM02 fatcat:encd2nsj3bcm7lp7tcbcsr7v2a

Modeling and analysis of dynamic coscheduling in parallel and distributed environments

Mark S. Squillante, Yanyong Zhang, Anand Sivasubramaniam, Natarajan Gautam, Hubertus Franke, Jose Moreira
2002 Performance Evaluation Review  
Our mathematical model and analysis is then used to explore several fundamental design and performance tradeoffs associated with the class of dynamic coscheduling policies across a broad spectrum of parallel  ...  We derive a matrix-analytic analysis based on a stochastic decomposition and a fixed-point iteration.  ...  In addition to mean job response time and maximum system throughput measures, our analysis provides detailed statis-tics (e.g., time spent in certain system/application states) that can be used to explain  ... 
doi:10.1145/511399.511341 fatcat:legl4f5bsfhtvioyoycyxrwahe

Modeling and analysis of dynamic coscheduling in parallel and distributed environments

Mark S. Squillante, Yanyong Zhang, Anand Sivasubramaniam, Natarajan Gautam, Hubertus Franke, Jose Moreira
2002 Proceedings of the 2002 ACM SIGMETRICS international conference on Measurement and modeling of computer systems - SIGMETRICS '02  
Our mathematical model and analysis is then used to explore several fundamental design and performance tradeoffs associated with the class of dynamic coscheduling policies across a broad spectrum of parallel  ...  We derive a matrix-analytic analysis based on a stochastic decomposition and a fixed-point iteration.  ...  In addition to mean job response time and maximum system throughput measures, our analysis provides detailed statis-tics (e.g., time spent in certain system/application states) that can be used to explain  ... 
doi:10.1145/511339.511341 fatcat:j2xlb3r675hydjuai7r36a7epm

Efficient Straggler Replication in Large-scale Parallel Computing [article]

Da Wang, Gauri Joshi, Gregory Wornell
2017 arXiv   pre-print
In a cloud computing job with many parallel tasks, the tasks on the slowest machines (straggling tasks) become the bottleneck in the job completion.  ...  Our analysis reveals that for certain execution time distributions, a small amount of task replication can drastically reduce both latency as well as the cost of computing resources.  ...  Thus, a key challenge in executing a job that consists of a large number of parallel tasks is the latency in waiting for the slowest tasks, or the "stragglers" to finish.  ... 
arXiv:1503.03128v3 fatcat:36xmen2rfveudmac62r3ed5oke
« Previous Showing results 1 — 15 out of 4,877 results