Private:Weekly Group 10/20/2011

Igor
Outline of issues addressed/tinkered with for the last 2 weeks:

Analysis tasks

 * did estimated for total MC needed: agreeing with those presented by RTJ at the Collaboration Meeting.
 * redid the estimates for MC contamination: ~1/3 of events will be background! Improvement possible?
 * Will better tracking chi-square come to the rescue?
 * Tighter cut on b1 mass, even cutting into it? (Pythia has many omegas but b1 distribution is flat) What will be the the effect on the systematics?
 * Fits succeeding in reaching the number of generated points. Sharing between waves not fully converging yet.
 * Adapted true, flat, 5-body phase space generator I started with.
 * Issue: Matt's phase space generator decayed the X and the b1 via 2-body decay (with unstable child mass picked from a BW distribution), followed by a 3-body decay of the &omega;.
 * 1) Discovered the source of the "echos" or "overtones" seen in the old spectra: breakup momenta vanished pion-mass away from resonance mass.
 * 2) Redid the scaling of batch-size to yield to find the point safe from clipping. Yield lower but
 * 3) Actually faster by an order of magnitude despite lower yield
 * Found that much due to low (flat) MC acceptance, most of the time during fits was spent computing the final integral with optimized wave parameters on the huge generated sample. Number integrated capped at 10k for the moment in AmpTools
 * Looked into compiling the ThreePi examples on the GPU: crash due to "invalid device pointer" when freeing GPU memory, apparently after the first amplitude computation cycle. Jon looking into it.
 * Understood how "parametrized" amplitudes are treated and how to control fitting with these
 * Not convinced the control is robust: "fixed" parameters sometimes got unfixed? Setting not remembered from fit to fit? Have to check more carefully
 * Not clear whether a "registered" parameter whose "fixing" is not specified causes amplitude recalculations
 * Packaging to get ready for production on OSG, but:
 * sim-recon factories less stable than used to be: found some "trap" events, submitted to Dave Lawrence et al.
 * more mysterious: mcsmear (now evidently multi-threaded) hangs and tries to replace threads, bloating in memory size

Papers

 * waiting for rocking curve temperature maps for diamond paper
 * Some text is there, heavy revision/direction reassessment necessary

Graduation

 * 2nd prospectus draft returned to RTJ. Doing another round and submitting to other committee members?
 * Needed for upcoming (?) Doctoral Dissertation Fellowship.
 * papers vital?
 * some postings up, many deadlines in January, I hear
 * beginning to contact employers, asking our collaborators for letters.
 * trying to find private time for thesis writing. May need a longer retreat, like fiancee's place over Thanksgiving/last week of semester.
 * stuff on electronics going resonably
 * diamond stuff: hope to make inroads on the paper to help with that chapter