July 14-15 Meeting

Please edit this to add your own agenda items.
More information is available on the main meeting web page.
If editing here is more than you want to do, please
send items to Ben Allan.


Pre-Meeting Activities Wednesday, 13 July 2005

CCA Book Planning Session - Contact David Bernholdt
1-5pm (CDT) at the Argonne Guest House (Conference Room A)

Babel RMI Implementation Discussion - Contact Kenneth Chiu
10am-noon (CDT) at the Argonne Guest House (Conference Room A)


CCA Summer Meeting

Thursday, 14 July 2005
Begin 9:00a

Tutorial/Book Update
David Bernholdt (ORNL). Summary of Wednesday's meeting and related news.

SC05 Booth/Demos Discussion
David Bernholdt (ORNL). 'Tis the season, once again.

Scientific Workflow and Components
Steve Parker will give us an outline of some of the work he's been doing lately with the SDM ISIC on scientific workflow and what he sees as the relationship and opportunities in this area as we think about SciDAC2. (disclaimer: David Bernholdt put Steve up to this talk, and invented this description of it out of whole cloth,so blame David)

Time for a Coup? or Where is CCA 1.0?
At the recent DARPA HPCS tutorial, the question was raised as to whether we should have "gov." at the beginning of the CCA namespace. So should we get rid of it? This would obviously be a significant change — of the sort that one probably wants to reserve for a major new version of the spec. Which raises the question of where is (or what is) CCA 1.0 and when are we going to have it? (Rob: feel free to drop this item if the agenda is too full, or if you feel it is not timely. db) Update: CCA Users can vote on the CCA namespace question here.

SciDAC2 Proposal Discussion - Randy Bramley (IU): Randy had the idea that now was the time to get serious with what the structure of the CCA proposal will look like. See (and edit) the Wiki page

XCAT-C++: Current Features, Performance and Future Plans - Madhu Govindaraju
XCAT-C++ has been implemented with Proteus multi-protocol library as its communication substrate. The current performance is quite encouraging and we are actively looking to hear from application scientists on their requirements.

CBSE Projects at Tech-X Corp.
On DOE sponsored projects called "Framework for Modernization and Componentization of Fusion Modules", FSML and "Distributed Components for Integrating HPC Applications" by Osvei Volberg

The Nexus Between Parallel and Distributed Components - Randy Bramley, IU
This is a discussion of Randy's new framework that purports to sew together existing parallel applications.

Babel RMI changes your life! (In some ways you didn't expect) - Jim Leek (LLNL)
A short session to list the changes to Babel's user interface that are coming about with the addition of RMI. If the discussion gets long, we may extend into a BOF the next day.

Having Your Cake and Eating It Too: Native Language Bindings to CCA - Rob Armstrong (SNL), and Chris Rickett (LANL): This is a (slightly) more fleshed-out idea that was prompted by the CCAIN framework discussions last meeting. Having a native language CCA binding (e.g. C in the case of CCAIN) that can be made a CCA compliant through the magic of code generation. Craig has been working on this too. Rob has started a parser in Ruby.

F90-to-SIDL/components
Craig, Matt, and Chris have been doing with chasm for generating the componentized code and sidl files from legacy f90. The chasm tool set generates these files from the xml representation of the legacy code, using
Xalan and xsl stylesheets.

Friday 15 July 2005

9am-Noon

BOFS

7a PT/8a MT/9am CT/10am ET: Book — continue first draft of outline

8:30am PT/9:30am MT/10:30am CT/11:30am ET: Tutorial — first draft of outline for new lecture organization