Child pages
  • 2014-06-25 Meeting Notes

Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

...

  • Cathy: SSTAR replication in MAF - the goal here is to be sure the content produced by the MAF is sufficient to replace the SSTAR when we switch from MySQL/SSTAR to SQLite/MAF.  The visualization/presentation layer is still in development, so that is making it difficult to assess whether the numbers in the SummaryStats.dat file are complete "enough".  For example some tables of numbers need to have the bins identified in order to interpret the numbers listed.  Lynne will produce an improved summaryStats.dat file, and Cathy will do an assessment using that file on the content as a whole.
  • Kem: runs for Cadence Workshop - Francisco and Srini are off the grid while at SPIE this week, and there is some confusion as to how to update the codebase on opsimblitz1 to include the latest features in the Master as well as the additional changes Francisco made to Kem's machine.  Kem thinks he can run a series of tests, pulling code from git and reserving edited files, to see what combination is correct - will take a few days.  A duplicate process needs to happen on opsimblitz2.  Simon recommends using SourceTree (www.sourcetreeapp.com - for mac and windows) as a GUI to manage checkouts of the codebase.  Also, additional fixes are needed to the RAlimit that needs Francisco.
  • Iain: Cathy will check in with Iain about how the conversion of opsimcvs to a Virtual Machine is going
  • Mozilla: there may be an opportunity to work with additional software engineers to help create the presentation layer for the MAF (

    http://mozillascience.org/bringing-researchers-and-developers-together-a-call-for-proposals/) Lynne will talk to Andy to see if this program might be a benefit to us.

  • El Gato at UA (http://elgato.arizona.edu/):  would this be a program we could make use of in addition to SLAC HPC?  30% of time is for community use.  Cathy will look into it.  Lynne asks if speed is the bottleneck to progress, and it will be once we have a code base we want to optimally tune all the parameters for, but right now it is getting features implemented and tested.