Meeting Notes for 9 April 2008

Present: Bob, Tom L., Tom J., Mihael, Liz;
Absent: Dan K, Mike W., Randy, Kris, Joao, Eric

Release of Refactored Code

take time to fix; users run the risk of the some of the analyses not running. Mihael is looking at other candidates for ssh code.

An alternative is running Globus on our local clusters but he does not want to do this. Her prefers using an open source library without licensing issues. There is a ossibility of using a licensed ssh product. Mihael is reluctant to use one.

For now Miheal recommends using SWIFT local.

Maybe we should turn off i2u2 cluster execution for now; www18 is more powerful than the current production using multiple machines. Bob wondered if we could beef up www18 more; Mihael says it is pretty powerful.

We wondered if we should buy a redundant machine. We could use one of the existing machine as a backup. It would be a lighter machine. (perhaps www11 or www14). Mihael does not think it is worth beefing it up and of the other machines. Jobs running on one could not be moved to another machine.

Deployment - We will have to wait until next week. Mihael traveling – he will be back in the office on Monday.

Supercomputing K12 education workshop

Tom L. reported that he learned more since the last telecon. The workshop is more a way for teachers to pull curriculum together. It would be good for Tom L. to be there and working with the teachers. We will not need to prepare a workshop. What a research project would look like. Tom might give a talk. It will probably be informal.

Tom L. described some ideas for research projects that students could work on together.

1) Detectable Forbush Decrease Searches – sun bursts at us; those particles would increase cosmic ray flux. You actually see a decrease.

2) Coordinated Shower Searches - How far apart do detectors have to be?

3) Issues of data blessing – serial or parallel effort.

To do either 1 or 2, we need to do 3.

Motivating people to bless data – perhaps having an online journal where students can publish.

Quarknet e-Lab fellows have come up with a low level approach to data blessing. How can we embed this method in the e-Lab? Tom thinks it will take a couple of days to convert performance study.

Jeff and Seattle folks providing a more sophisticated approach.

Low level: • Was the GPS on throughout it? • Lower number of orphans?

The time expired.

There was a long discussion that was continued by Tom J, Tom L and Bob about ways to implement blessing of data that took place after Liz and Mihael signed off.

Topics that we did not cover:

  • Bug report page

  • Problem with data server linking properly; how to monitor?

  • Saving plots for LIGO?

-- Main.LizQuigg - 14 Apr 2008
-- Main.EricMyers - 16 Apr 2008
Topic revision: r6 - 2008-04-16, EricMyers
This site is powered by FoswikiCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding Foswiki? Send feedback