Hi,
I checked in new codes into DGen in the last few days. I was a little
reluctant to do so since I was still checking the codes. But I was asked
about it several times lately, I decided to commit it.
DGen is fully revised in hope for easier extensibility as there are still
major reaction channels yet implemented and it seems likely more than
myself will be involve in further developement of the code.
New data and formalism for electro-disintegration channel are incoporated.
Please see: Arenhovel, PRC 55, 2214, 1997 and Arenhovel, nucl-th/9905029
for theoretical detail. I just want to point out that our set up will
measure unpolarized final states, which is a type-A spin component defined
by Arenhovel. One needs this fact to explicitly compute the cross section.
Also, the response functions in the new formalism are identical to or are
linear combinations of the old ones. A few examples are: fL(old) = fL(00),
fTT11-(old) = (fTT11--fTT11+)/2, fTT11+(old) = (fTT11- + fTT11+)/2. (Sign
may be incorrect here).
I decide the 10 KINE parameters are far from enough yet a lot of variables
need to be adjustable without recompilation. Therefore a file kine.init
was added. It is heavily commented, hope by reading it one can get a
feeling on how to run the new version of dgen before full documentation
comes out.
I will be available for questions. Also, this release is rather hasty, I
am sure there are a lot of bugs. I will be collecting bug reports and
correct the codes continuously. The codes are still under construction. If
you spot bugs that are easy to correct, please go ahead and check in the
correction. But please do be kine to send me (or blast_anaware) a message.
Since it is very different from the previous release, I tag the earlier
one as v_1. One must specify -r v_1 to check out the previous release.
highlight of new features:
*new(correct) electro-disintegration cross section.
*flexibility: more adjustable parameters without recompilation
*extensibility: adding new channels should be earier. only a Deuteron
class needs to be defined with almost no modification to existing codes.
(with the only exception of DGen.cc)
*event generation with mixed channels, alternating target and beam
polarization. Total number of events for a certain channel,
certain spin orientation is weighted by the totat cross section.
It is now possible to generate Asymmetry data with one dgen or
blast session.
*more decoupled to ROOT. The codes can be compiled with or without support
for ROOT interpreter session. The binary generated without
ROOTCINT are much smaller.
*efficiency: a hashing machinism is introduced to extract response
functions for EEP cross sections. The result is that
initialization of EEP channel now takes only 20-30 seconds with a
rather dense grid in phase space. The run time of old version
would be untolerable with a grid of the same size. Some efforts
are made to optimize the code for run time and memory usage. In
EEP channel, it normaly guarantees 500events/sec. Negligible compare
to the run time of blastmc.
Chi
This archive was generated by hypermail 2.1.2 : Mon Feb 24 2014 - 14:07:29 EST