|
|
|
|
|
Condor Cycle ScavengerCondor is a software framework for coarse-grained distributed parallelization of computationally intensive tasks. It was developed at the University of Wisconsin-Madison. Condor et al. 1988 uses a cluster of Unix or Windows workstations on a network. The NASA Advanced Supercomputing (NAS) Division Condor pool consists of approximately 350 SGI and Sun workstations purchased and used for software development, visualization, email, document preparation, etc. Each workstation runs a daemon that watches user I/O and CPU load. When a workstation has been idle for two hours, a job from the batch queue is assigned to the workstation and will run until the daemon detects a keystroke, mouse motion, or high non-Condor CPU usage. At that point, the job will be removed from the workstation and placed back on the batch queue. Since Condor runs on collections of workstations continuously connected to a network, I/O to and from each job can run in the hundreds of megabytes or even a few gigabytes for long-running jobs. However, communication between jobs is difficult since jobs are always starting and stopping. Condor has a mechanism to set up master-slave messaging relationships via PVM et al. 2000, but JavaGenes does not use it. MPI is not used since MPI 1 does not allow processes to be added or subtracted from an application. External links
|
 |
|
| Copyright 2005-2009 OnPedia.com. All Rights Reserved |
|
|