Applications ported to edges
This presentation is the property of its rightful owner.
Sponsored Links
1 / 24

Applications Ported to EDGeS PowerPoint PPT Presentation


  • 70 Views
  • Uploaded on
  • Presentation posted in: General

Applications Ported to EDGeS. Tamas Kiss – University of Westminster Leader of the EDGeS Application Support Service E-mail: [email protected] The EDGeS Application Support Service. Aims and objectives: facilitate the porting of applications to the EDGeS infrastructure by targeting

Download Presentation

Applications Ported to EDGeS

An Image/Link below is provided (as is) to download presentation

Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author.While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server.


- - - - - - - - - - - - - - - - - - - - - - - - - - E N D - - - - - - - - - - - - - - - - - - - - - - - - - -

Presentation Transcript


Applications ported to edges

Applications Ported to EDGeS

Tamas Kiss – University of WestminsterLeader of the EDGeS Application Support Service E-mail: [email protected]

The EDGeS project receives Community research funding


The edges application support service

The EDGeS Application Support Service

  • Aims and objectives:

    • facilitate the porting of applications to the EDGeS infrastructure by targeting

      • user communities already associated with EGEE or DGs

      • new user communities

  • Activities:

    • Identify user communities that require the power of the EDGeS infrastructure

    • develop a generic methodology for application porting

    • provide a service in order to aid the migration to and running of applications on the EDGeS infrastructure

Tamas Kiss – [email protected]

Joint EDGeS EGEE Summer School, 29th June – 4th July 2009, Budapest


Applications ported to edges

The EDGeS user community

  • Academic user community:

    • 18 identified communities

      • BioInfoGrid, e-Health, Fusion, Computational Chemists, etc.

      • 4 new academic user partners

    • Establishing the EDGeS user forum

  • Industrial user community

    • 5 identified companies

    • 3 new industrial user partners

    • Establishing the EDGeS industry forum

Tamas Kiss – [email protected]

Joint EDGeS EGEE Summer School, 29th June – 4th July 2009, Budapest


Applications ported to edges

EDGeS Application Areas

Tamas Kiss – [email protected]

Joint EDGeS EGEE Summer School, 29th June – 4th July 2009, Budapest


An overview of edges applications

An Overview of EDGeS Applications

√: actual

√: plan

Tamas Kiss – [email protected]

Joint EDGeS EGEE Summer School, 29th June – 4th July 2009, Budapest


An overview of edges applications1

An Overview of EDGeS Applications

√: actual

√: plan

Tamas Kiss – [email protected]

Joint EDGeS EGEE Summer School, 29th June – 4th July 2009, Budapest


An overview of edges applications2

An Overview of EDGeS Applications

√: actual

√: plan

Tamas Kiss – [email protected]

Joint EDGeS EGEE Summer School, 29th June – 4th July 2009, Budapest


An overview of edges applications3

An Overview of EDGeS Applications

√: actual

√: plan

Tamas Kiss – [email protected]

Joint EDGeS EGEE Summer School, 29th June – 4th July 2009, Budapest


An overview of edges applications4

An Overview of EDGeS Applications

√: actual

√: plan

Tamas Kiss – [email protected]

Joint EDGeS EGEE Summer School, 29th June – 4th July 2009, Budapest


Applications ported to edges

Implementation environmentThe University of Westminster Local DG (WLDG)

6

1

2

5

4

3

  • New Cavendish Street576 nodes

  • Marylebone Campus559 nodes

  • Regent Street395 nodes

  • Wells Street31 nodes

  • Little Tichfield Street66 nodes

  • Harrow Campus254 nodes

Tamas Kiss – [email protected]

Joint EDGeS EGEE Summer School, 29th June – 4th July 2009, Budapest


Applications ported to edges

Implementation environmentThe University of Westminster Local DG (WLDG)

  • DG Server on private University network

  • Over 1500 Windows PCs from 6 different campuses

    • All machines running windows

    • Most of the machines have dual core processors

  • Running the Local SZTAKI Desktop Grid package

  • BOINC clients are installed automatically and maintained by specifically developed Novell ZENworks objects

  • All workers are registered under the same user account.

  • The workers are only available for DG computation when they are switched on but not used by students

Tamas Kiss – [email protected]

Joint EDGeS EGEE Summer School, 29th June – 4th July 2009, Budapest


Applications ported to edges

Implementation environmentThe University of Westminster Local DG (WLDG)

Tamas Kiss – [email protected]

Joint EDGeS EGEE Summer School, 29th June – 4th July 2009, Budapest


Applications ported to edges

Application case study Protein Molecule Simulation

  • The user community:

    • Department of Molecular & Applied Biosciences University of Westminster

  • The biological perspective:

    • Understanding how sugars interact with their protein partners may lead to development of new treatment methods for many diseases.

  • The obstacle:

    • Wet laboratory experiments are expensive

    • Computer simulations run too long and programs are too complicated for biologists to handle

  • The solution:

    • Run simulations on the Grid

    • Execute simulations from user friendly portal interface

Tamas Kiss – [email protected]

Joint EDGeS EGEE Summer School, 29th June – 4th July 2009, Budapest


Applications ported to edges

Application case study Protein Molecule Simulation

  • AutoDock:

    • a suite of automated docking tools

    • designed to predict how small molecules, such as substrates or drug candidates, bind to a receptor of known 3D structure

    • one run of AutoDock finishes in a reasonable time on a single PC. However, thousands of scenarios have to be simulated and analysed to get stable and meaningful results.

      • AutoDock has to be run multiple times with the same input files but with random factors

      • Simulations runs are independent from each other – suitable for DG

Tamas Kiss – [email protected]

Joint EDGeS EGEE Summer School, 29th June – 4th July 2009, Budapest


Applications ported to edges

Application case study Protein Molecule Simulation – the parallelisation

Tamas Kiss – [email protected]

Joint EDGeS EGEE Summer School, 29th June – 4th July 2009, Budapest


Applications ported to edges

Application case study Protein Molecule Simulation – performance results

Tamas Kiss – [email protected]

Joint EDGeS EGEE Summer School, 29th June – 4th July 2009, Budapest


Applications ported to edges

BOINC to EGEE bridge

Application case study Protein Molecule Simulation – running through the bridge

gUSE WS P-GRADE portal

UoW Local Desktop Grid

G-USE DG Submitter submits jobs and retrieve results via 3G Bridge

BOINC Server

Pulls work units

Adds user credentials

Submits as job to EGEE

Define input parameters

and number of work units

WMS

Workers: Download executable and input files

Upload: result

BOINC workers

End User

EGEE EDGeS VO

Tamas Kiss – [email protected]

Joint EDGeS EGEE Summer School, 29th June – 4th July 2009, Budapest


Applications ported to edges

An Application case study Protein Molecule Simulation – the g-USE interface

UoW Local DG

WS-PGRADE portal

Tamas Kiss – [email protected]

Joint EDGeS EGEE Summer School, 29th June – 4th July 2009, Budapest


Applications ported to edges

Application ExamplesDigital Alias-free Signal Processing (DASP)

  • Users: Centre for Systems Analysis – University of Westminster

  • Traditional DSP -> Uniform sampling

    • Suffers from aliasing

  • Aim: digital alias-free signal processing (DASP)

    • One solution is Periodic Non-uniform Sampling (PNS)

  • The DASP application designs PNS sequences

  • Selection of optimal sampling sequence is a computationally expensive process

    • A linear equation has to be solved and a large number of solutions (in the magnitude of 1010) have to be compared.

  • The analysis of the solutions are independent from each other, suitable for DG parallelisation

Tamas Kiss – [email protected]

Joint EDGeS EGEE Summer School, 29th June – 4th July 2009, Budapest


Applications ported to edges

Application ExamplesDASP – the parallelization

Master

Workers

Master

Tamas Kiss – [email protected]

Joint EDGeS EGEE Summer School, 29th June – 4th July 2009, Budapest


Applications ported to edges

BOINC to EGEE bridge

DASP – Running through the BOINC->EGEE Bridge

UoW Local Desktop Grid

Master: Define input parameters (T, L,…)

Define number of work units

BOINC Server

DC-API: Creates work units, collects and processes results

Adds user credentials

Submits as job to EGEE

Pulls work units

WMS

Workers: Download: executable and solution vector

Upload: result (best local J value)

BOINC workers

EGEE EDGeS VO

Tamas Kiss – [email protected]

Joint EDGeS EGEE Summer School, 29th June – 4th July 2009, Budapest


Applications ported to edges

Application examples Patient Readmission Application

  • The users:

    • Health and Social Care Modeling Group at UoW

  • The aim:

    • profiling hospitals based on patient readmission profiles

  • The challenge:

    • The Hospital Episode Statistics dataset includes 80 million episodes in total for a seven financial year period.

    • Implementing a statistical model using the complete population dataset is unfeasible.

    • Hierarchical cluster analysis is performed where the data is divided into clusters and then sampled

    • The samples assigned a rank value independently

    • The process has to be repeated thousands of times to get statistically valid results

Tamas Kiss – [email protected]

Joint EDGeS EGEE Summer School, 29th June – 4th July 2009, Budapest


Applications ported to edges

Application examples Rendering Application

  • The challenge:

    • The rendering of complex 3D contents is a computationally very expensive process. In a single processor it takes weeks or months to render a longer animation.

    • Blender is one of the most popular Open Source 3D graphics application in the world.

    • In Blender each frames of an animation are rendered separately, that’s why the application is suitable for DG style of parallelisation.

Tamas Kiss – [email protected]

Joint EDGeS EGEE Summer School, 29th June – 4th July 2009, Budapest


Thank you for your attention

Any

questions?

Thank you for your attention …

For more information please visit the EDGeS Website:

Join the EDGeS User & Industry Forum here!

http://www.edges-grid.eu/

Please contact us if you need support in porting your application to EDGeS!

Email:[email protected]

Tamas Kiss – [email protected]

Joint EDGeS EGEE Summer School, 29th June – 4th July 2009, Budapest


  • Login