Curtis on DC-1

From GlueXWiki
Jump to: navigation, search
From: "Curtis A. Meyer" <cmeyer@cmu.edu>
Subject: Data Challenge Meeting
Date: Mon, 17 Dec 2012 12:16:18 -0500
Cc: Paul Mattione <pmatt@jlab.org>, "Meyer Curtis A." <curtis.meyer@cmu.edu>
To: "Ito Mark M." <marki@jlab.org>

Hi Mark -

One of the things I would like to be able to do now that we have
finished this is to write a short report that we will then have for
the software review next fall. Information that will need to go into
it are things like:

     * A clear description of the type of events that we generated.
     * No. of events generated at each site.
     * Total expected data footprint of these events
     * Estimate of how much running at 10**7 photons this represents.
     * Development of a common repository for all the events and a
       plan to move the data there.
     * Summary of issues encountered.
        - crashes
        - random number sees
        - are there obvious bottle necks we could improve upon?
     * Plans to address the issues
     * Moving to the next challenge
        - what did we strip out of this challenge to get it going?
        - plans to implement that in the next challenge ?
        - How well did our release mechanism work?
        - How well did things on the JLab Batch farms go? Were there
          bottlenecks?

Also, as I noted in the Physics meeting, I think that it is reasonable
to stop our challenge now and take stock of where we are. Our goal
should be to make the data available as quickly as we can for the
people in the physics group that want to use it.

Thanks for all the hard work to everyone, I think that this has been a
great success!

  Curtis