I have put three new GSIM files in the simulation repository:
1) 2.4gev_p_tgt_e_phi=0b.evt 16.2 Mbytes
2) 2.4gev_p_tgt_1.evt 47.6 Mbytes
3) electron_p_theta.evt 37.6 Mbytes
I also renamed 2.4gev_p_tgt_e_phi=0.evt to 2.4gev_p_tgt_e_phi=0a.evt.
Number 1 above is a rerun of the electron phi = 0 run with the
new SC geometry and digitization. It has 5000 events. Experience has
shown that a crash is much less probable with the electron phi = 0
than with it unconstrained; this run did not crash. The event
generator was CELEG, with all resonances and elastic scattering
turned on.
Number 2 is a run on the SUN computers (!!). The event generator
was CELEG, with no constraints on electron phi angle. An earlier
HP-UX run crashed on event 3000. On the SUN IT DID NOT CRASH!!
The run terminated normally at event 9914 when the internal time
limit expired. (8.7 seconds/event for 9914 events is 24 cpu hours).
I have made several runs on the Suns which ended in various ways,
but none have been because the program crashed on its own.
The executible for this run is several weeks old and so does not
contain the new SC geometry.
Number 3 is a diagnostic run with electrons which is described
in its documentation file. It has the new SC geometry.
I heard from Alex that he has some clues about what causes the
crashes of GSIM on the HP's. He has been hotly pursuing this.
The ntuple versions of the above files will be available as
soon as TY creates them. This requires somehow keeping track
of whether the SC geometry and digitization is a new or old
version.
I heard from John Ficenec that he has put in both the lead and
the stainless steel into the SC simulation, and is testing this.
Progress progress progress! I assume he'll announce this when
it's ready for public consumption.
- Will