Abstract
We describe a scheme for subdividing long-running, variable-length analyses into short, fixed-length boinc workunits using phylogenetic analyses as an example. Fixed-length workunits decrease variance in analysis runtime, improve overall system throughput, and make boinc a more useful resource for analyses that require a relatively fast turnaround time, such as the phylogenetic analyses submitted by users of the garli web service at molecularevolution.org. Additionally, we explain why these changes will benefit volunteers who contribute their processing power to boinc projects, such as the Lattice boinc Project (http://boinc.umiacs.umd.edu). Our results, which demonstrate the advantages of relatively short workunits, should be of general interest to anyone who develops and deploys an application on the boinc platform.
Highlights
Computing resources volunteered by members of the general public can greatly benefit scientific research, as demonstrated by high-profile research projects in disparate areas such as radio astronomy (SETI@home; setiathome.berkeley.edu), climate modeling, protein folding (Rosetta@home; boinc.bakerlab.org/rosetta), and particle accelerator physics (LHC@home; lhcathomeclassic.cern.ch/sixtrack), to name just a few
The Lattice BOINC Project is an outstanding resource for running GARLI analyses: a significant proportion of volunteer computers have an appreciable amount of memory, which GARLI analyses often require; and GARLI automatically checkpoints its state when running on BOINC, which allows for efficient use of the BOINC platform
Far it has not been feasible to run GARLI web service analyses on BOINC because it has been difficult to guarantee complete results from BOINC in a timely manner. We address this problem by subdividing long-running GARLI analyses into short, fixed-length BOINC workunits
Summary
Computing resources volunteered by members of the general public can greatly benefit scientific research, as demonstrated by high-profile research projects in disparate areas such as radio astronomy (SETI@home; setiathome.berkeley.edu), climate modeling (climateprediction.net), protein folding (Rosetta@home; boinc.bakerlab.org/rosetta), and particle accelerator physics (LHC@home; lhcathomeclassic.cern.ch/sixtrack), to name just a few. Far it has not been feasible to run GARLI web service analyses on BOINC because it has been difficult to guarantee complete results from BOINC in a timely manner. We address this problem by subdividing long-running GARLI analyses into short, fixed-length BOINC workunits (the term used for a unit of work on the BOINC platform). This speeds up analysis completion by reducing the variance in workunit runtimes, making BOINC a more attractive resource for analyses that require a relatively fast turnaround time.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have