Competitive Dynamic Multiprocessor Allocation for Parallel Applications

1997 ◽  
Vol 07 (01) ◽  
pp. 89-100 ◽  
Author(s):  
Timothy Brecht ◽  
Xiaotie Deng ◽  
Nian Gu
Keyword(s):  
Execution Time ◽  
Completion Time ◽  
Constant Factor ◽  
Parallel Applications ◽  
Scheduling Policies ◽  
Scheduling Policy ◽  
Parallel Tasks ◽  
Parallel Jobs ◽  
Parallel Job ◽  
Two Parameter

We study dynamic multiprocessor allocation policies for parallel jobs, which allow the preemption and reallocation of processors to take place at any time. The objective is to minimize the completion time of the last job to finish executing (the makespan). We characterize a parallel job using two parameter. The job's parallelism, Pi, which is the number of tasks being executed in parallel by a job, and its execution time, li, when Pi processors are allocated to the job. The only information available to the scheduler is the parallelism of jobs. The job execution time is not known to the scheduler until the job's execution is completed. We apply the approach of competitive analysis to compare preemptive scheduling policies, and are interested in determining which policy achieves the best competitive ratio (i.e., is within the smallest constant factor of optimal). We devise an optimal competitive scheduling policy for scheduling two parallel jobs on P processors. Then, we apply the method to schedule N parallel jobs on P processors. Finally we extend our work to incorporate jobs for which the number of parallel tasks changes during execution (i.e., jobs with multiple phases of parallelism).

An Algorithm for Task Scheduling in Heterogeneous Distributed Systems Using Task Duplication

2011 ◽  
Vol 3 (1) ◽  
pp. 89-97 ◽  
Author(s):  
Amrit Agrawal ◽  
Pranay Chaudhuri
Keyword(s):  
Task Scheduling ◽  
Execution Time ◽  
High Speed ◽  
Critical Path ◽  
Task Assignment ◽  
Directed Acyclic Graphs ◽  
Parallel Tasks ◽  
Schedule Length ◽  
Task Duplication ◽  
Heterogeneous Distributed Systems

Task scheduling in heterogeneous parallel and distributed computing environment is a challenging problem. Applications identified by parallel tasks can be represented by directed-acyclic graphs (DAGs). Scheduling refers to the assignment of these parallel tasks on a set of bounded heterogeneous processors connected by high speed networks. Since task assignment is an NP-complete problem, instead of finding an exact solution, scheduling algorithms are developed based on heuristics, with the primary goal of minimizing the overall execution time of the application or schedule length. In this paper, the overall execution time (schedule length) of the tasks is reduced using task duplication on top of the Critical-Path-On-a-Processor (CPOP) algorithm.

The Sicilian Grid Infrastructure for High Performance Computing

2010 ◽  
Vol 1 (1) ◽  
pp. 40-54 ◽  
Author(s):  
Carmelo Marcello Iacono-Manno ◽  
Marco Fargetta ◽  
Roberto Barbera ◽  
Alberto Falzone ◽  
Giuseppe Andronico ◽  
...  
Keyword(s):  
High Performance Computing ◽  
High Performance ◽  
Parallel Applications ◽  
Grid Infrastructure ◽  
Scheduling Policy ◽  
Computing Paradigm ◽  
Regional Area ◽  
Computer Fluid Dynamics ◽  
Grid Infrastructures ◽  
Performance Computing

The conjugation of High Performance Computing (HPC) and Grid paradigm with applications based on commercial software is one among the major challenges of today e-Infrastructures. Several research communities from either industry or academia need to run high parallel applications based on licensed software over hundreds of CPU cores; a satisfactory fulfillment of such requests is one of the keys for the penetration of this computing paradigm into the industry world and sustainability of Grid infrastructures. This problem has been tackled in the context of the PI2S2 project that created a regional e-Infrastructure in Sicily, the first in Italy over a regional area. Present article will describe the features added in order to integrate an HPC facility into the PI2S2 Grid infrastructure, the adoption of the InifiniBand low-latency net connection, the gLite middleware extended to support MPI/MPI2 jobs, the newly developed license server and the specific scheduling policy adopted. Moreover, it will show the results of some relevant use cases belonging to Computer Fluid-Dynamics (Fluent, OpenFOAM), Chemistry (GAMESS), Astro-Physics (Flash) and Bio-Informatics (ClustalW)).

Some Extensions of Hilbert-Type Inequalities

2012 ◽  
Vol 542-543 ◽  
pp. 1403-1406
Author(s):  
Bao Ju Sun
Keyword(s):  
Constant Factor ◽  
Best Constant ◽  
Best Constant Factor ◽  
Two Parameter ◽  
Hilbert Type

In this paper, an extension of Hilbert-type inequalities with a best constant factor is given by introducing two parameter .

scholarly journals Space profiling for parallel functional programs

2010 ◽  
Vol 20 (5-6) ◽  
pp. 417-461 ◽  
Author(s):  
DANIEL SPOONHOWER ◽  
GUY E. BLELLOCH ◽  
ROBERT HARPER ◽  
PHILLIP B. GIBBONS
Keyword(s):  
Resource Use ◽  
Source Code ◽  
Semantic Space ◽  
Runtime System ◽  
Scheduling Policies ◽  
Use Patterns ◽  
Scheduling Policy ◽  
Standard Ml ◽  
Cost Semantics ◽  
The Impact

AbstractWe present a semantic space profiler for parallel functional programs. Building on previous work in sequential profiling, our tools help programmers to relate runtime resource use back to program source code. Unlike many profiling tools, our profiler is based on a cost semantics. This provides a means to reason about performance without requiring a detailed understanding of the compiler or runtime system. It also provides a specification for language implementers. This is critical in that it enables us to separate cleanly the performance of the application from that of the language implementation. Some aspects of the implementation can have significant effects on performance. Our cost semantics enables programmers to understand the impact of different scheduling policies while hiding many of the details of their implementations. We show applications where the choice of scheduling policy has asymptotic effects on space use. We explain these use patterns through a demonstration of our tools. We also validate our methodology by observing similar performance in our implementation of a parallel extension of Standard ML.

A Parallel Job Execution Time Estimation Approach Based on User Submission Patterns within Computational Grids

2013 ◽  
Vol 43 (3) ◽  
pp. 440-454 ◽  
Author(s):  
Feng Liang ◽  
Yunzhen Liu ◽  
Hai Liu ◽  
Shilong Ma ◽  
Bettina Schnor
Keyword(s):  
Execution Time ◽  
Time Estimation ◽  
Computational Grids ◽  
Parallel Job

The Sicilian Grid Infrastructure for High Performance Computing

2012 ◽  
pp. 215-227
Author(s):  
Carmelo Marcello Iacono-Manno ◽  
Marco Fargetta ◽  
Roberto Barbera ◽  
Alberto Falzone ◽  
Giuseppe Andronico ◽  
...  
Keyword(s):  
High Performance Computing ◽  
High Performance ◽  
Parallel Applications ◽  
Grid Infrastructure ◽  
Scheduling Policy ◽  
Computing Paradigm ◽  
Regional Area ◽  
Computer Fluid Dynamics ◽  
Grid Infrastructures ◽  
Performance Computing

The conjugation of High Performance Computing (HPC) and Grid paradigm with applications based on commercial software is one among the major challenges of today e-Infrastructures. Several research communities from either industry or academia need to run high parallel applications based on licensed software over hundreds of CPU cores; a satisfactory fulfillment of such requests is one of the keys for the penetration of this computing paradigm into the industry world and sustainability of Grid infrastructures. This problem has been tackled in the context of the PI2S2 project that created a regional e-Infrastructure in Sicily, the first in Italy over a regional area. Present paper will describe the features added in order to integrate an HPC facility into the PI2S2 Grid infrastructure, the adoption of the InifiniBand low-latency net connection, the gLite middleware extended to support MPI/MPI2 jobs, the newly developed license server and the specific scheduling policy adopted. Moreover, it will show the results of some relevant use cases belonging to Computer Fluid-Dynamics (Fluent, OpenFOAM), Chemistry (GAMESS), Astro-Physics (Flash) and Bio-Informatics (ClustalW)).

Sign in / Sign up

Export Citation Format

Share Document