1 / 28

Recommended Viewing:

Recommended Viewing:. Since this is a pictorial presentation, I’ve gone through the effort to type up what I normally say in the “notes” section. To enjoy the animations, I recommend printing out the “notes” then watching via the slide show viewer. --Gary Kumfert.

constancer
Download Presentation

Recommended Viewing:

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. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. Recommended Viewing: Since this is a pictorial presentation, I’ve gone through the effort to type up what I normally say in the “notes” section. To enjoy the animations, I recommend printing out the “notes” then watching via the slide show viewer. --Gary Kumfert

  2. A Pictorial Introduction to Components in Scientific Computing Gary Kumfert with Steve Smith, Scott Kohn, Tom Epperly, Tammy Dahlgren, & Bill Bosl

  3. Input Output Program Once upon a time...

  4. Bottle Neck As Scientific Computing grew...

  5. Tried to ease the bottle neck

  6. 1 1 2 2 1 3 3 4 4 2 4 3 SPMD was born.

  7. But it isn’t easy!!! But it isn’t easy!!! 1 1 2 2 1 3 3 4 4 2 4 3 SPMD worked.

  8. Input Input email client spreadsheet browser graphics Output editor Unicode Program multimedia database Meanwhile, corporate computing was growing in a different way

  9. email client spreadsheet browser graphics editor Unicode multimedia database This created a whole new set of problems... • Interoperability across multiple languages • Interoperability across multiple platforms • Incremental evolution of large legacy systems (esp. w/ multiple 3rd party software)

  10. Component Technology addresses these problems

  11. So what’s a component ??? Implementation : No Direct Access Interface Access : Generated by Tools Matching Connector : Assigned by Framework Hidden from User

  12. Language & Platform independent interfaces C Java F77 C++ Automatically generated bindings to working code Python 1. Interoperability across multiple languages

  13. What if the source to this one part is lost??? 2. Interoperability Across Multiple Platforms Imagine a company migrates to a new system, OS, etc.

  14. These wires are very, very smart! Transparent Distributed Computing internet internet

  15. v 1.0 v 3.0 v 2.0 3. Incremental Evolution With Multiple 3rd party software

  16. v 1.0 v 3.0 v 2.0 Now suppose you find this bug...

  17. 2.0 v 1.0 v 3.0 v 2.0 2.1 Good news: an upgrade available Bad news: there’s a dependency

  18. Great News: Solvable with Components 2.0 v 3.0 2.1

  19. v 1.0 2.0 2.1 Great News: Solvable with Components v 3.0

  20. SAMRAI DataFoundry JEEP Sapphire Scientific Viz Ardra Overture nonlinear solvers ALPS hypre linear solvers Why Components for Scientific Computing? • Interoperability across multiple languages • Interoperability across multiple platforms • Incremental evolution of large legacy systems (esp. w/ multiple 3rd party software)

  21. Science Industry ? The Model for Scientific Component Programming

  22. Parallel Distributed Component-Based Application

  23. Research Issues: #1. The “MxN Problem”

  24. Science Industry ? Research Issues:#2: Programming Model vector.getLocalSize() on remote objects?

  25. Science Industry Is This Still SPMD?

  26. Is This Still SPMD? • No • Each “component” may be an entire legacy SPMD code • Multiple components (possibly distributed) working together on a single problem • MPMD, MCMD, DPMD??? • But • Will look like SPMD to application developer • Business components look like serial code.

  27. SAMRAI DataFoundry JEEP Sapphire Scientific Viz Ardra Overture nonlinear solvers ALPS hypre linear solvers Why Components for Scientific Computing? • Interoperability across multiple languages • Interoperability across multiple platforms • Incremental evolution of large legacy systems (esp. w/ multiple 3rd party software)

  28. TheEnd

More Related