report CSC (IFC meeting preparations)
Gregory Dubois-Felsmann 2008.6.10 CSC
Computing model / "CSC spreadsheet"
All sites have now agreed to this basic proposition:
During the "intense analysis period", SLAC and Tier-A sites will maintain the BaBar computing capacity that they currently have. The CSC will oversee the availability and use of these resources. This will be done without CF payments or rebates, i.e., there will be no financial mechanism.
This means the "CSC spreadsheet" common fund budget goes to $0 in 2009 and beyond.
2008.6.10
CSC - Dubois-Felsmann
2
Details will evolve with time
The baseline plan is to maintain the current association of sites with AWGs.
Adjustments can be made if there are large changes in the relative load from different AWGs.
Clearly some ramp-down of total need can be anticipated.
The details are strongly dependent on the evolution of BaBar physics analysis this year. (E.g., on what goes to ICHEP / gets published.) The CSC will watch over this. The CSC will work with all sites, beginning today, to determine a date by which it be would helpful for more precise estimates for 2009 needs to be available (sometime this fall, presumably), and will work with the BaBar PAC and with computing management to develop a model. Drops in need may allow some of the oldest hardware to be retired without replacement.
2008.6.10 CSC - Dubois-Felsmann 3
Inputs to OCF budget calculation for 2009
Tapes (revised!)
Skims: 1917 tapes (was 730)
Based on two R22c/e-like incremental cycles and half of one R22d-like full skim, scaled up to 620ifb, allowing 15% contingency / bad-data
Signal MC: 95 tapes (based on recent signal production rate) Astore: 97 tapes (based on last 12-month period of use) Total: 210...