[Halld-offline] [Halld-physics] pilot run for pwa: event generation
Matthew Shepherd
mashephe at indiana.edu
Wed Mar 17 18:00:31 EDT 2010
Hi Craig,
We have our own fitter and "amplitude manipulation package" that is independent of the ruby fitter. If I (not necessarily you) were going to do simulated GlueX PWA in the next six months I'd be using my own fitter and the corresponding Monte Carlo generation tools. It exists, and is currently being used by our group at Indiana to analyze CLEO and BESIII data. It is very capable of being used to analyze GlueX data (we just need to write the amplitudes, which is small task in comparison to writing the fitter, user interface, etc.). However our package is not polished and documented at a level where you can just download and do configure ; make ; make install and be off generating and fitting data.
One of the motivations for using a different fitter is performance. This may or may not be a limitation for the ruby fitter using simple isobar model amplitudes. I don't see any reason necessarily to not use the ruby fitter; however, we at Indiana don't have the experience to support that effort since we are not familiar with the code.
What will we use in the "long term"? It is hard to tell -- it wouldn't hurt to have a couple of fitters. This is how bugs are found and cross checks made. In addition certain fitting algorithms may be better optimized for certain types of fits.
-Matt
On Mar 17, 2010, at 5:39 PM, Craig Bookwalter wrote:
> I think what I am after here is this: if I'm going to do a PWA on simulated GlueX data sometime during the next six months, what tools will I be using? If the long-term plan is to use something other than the ruby fitter, and that something doesn't exist yet, then there's not much value in doing this. For my part, the ruby fitter exists and is functional, so I don't see why we shouldn't use it.
>
> --cb
>
> Matthew Shepherd wrote:
>> Hi Craig,
>>
>> I don't know -- I've never used the ruby fitter. Back in the old days there was a program called something like "gamps" that took in some data files and spit out a complex number for every amplitude evaluated for each event. These then got fed into a separate fitter. I believe the Ruby fitter was written in this style, with amplitude calculation independent of fitting. What I need is that recipe for turning the data into these complex numbers. In general this is a process that is completely independent of fitting, but this is where the "the physics" is. Once we have that model coded up, the everything else is pretty straightforward. I literally just need a C function that takes in an array of floats (4-vectors) and returns a single complex number.
>>
>> -Matt
>>
>>
>> On Mar 17, 2010, at 4:58 PM, Craig Bookwalter wrote:
>>
>>
>>> Please bear with me--is this not something that already exists in the Ruby fitter? Or is the Ruby fitter not going be used in GlueX?
>>>
>>>
>>> Matthew Shepherd wrote:
>>>
>>>> Hi Craig,
>>>>
>>>> I basically need a function that I can plug some four vectors into and it returns a complex number that equal to the amplitude for the particular reaction. The function likely has parameters that depend on spin and polarization of various particles in the decay and consists of the correct linear combination of d-functions evaluated with angles in the appropriate frames. It essentially form "the basis" for the PWA. GlueX-doc 29 (written by Curtis) is a start -- we should verify, understand, and code this up. Also, it might be easier to start with something like eta pi in the final state. In this case, how does the photon polarization manifest itself in the final state angular distributions?
>>>>
>>>> -Matt
>>>>
>>>>
>>>> On Mar 17, 2010, at 4:31 PM, Craig Bookwalter wrote:
>>>>
>>>>
>>>>> Hi Matt,
>>>>> Thanks for the quick reply--I'm confused as to what you mean by an `expression' for the amplitudes...can you elaborate?
>>>>>
>>>>> Thanks again.
>>>>>
>>>>> --cb
>>>>>
>>>>>
>>>>> Matthew Shepherd wrote:
>>>>>
>>>>>> Hi Craig,
>>>>>>
>>>>>> We just about finished with version 1.0 (or 0.1?) of our AmpTools package that will allow both MC generation and fitting. (In fact, I'm sitting in the BESIII control room writing code right now.) It will take some time to get together a documented package, but I'm happy to start to work with anyone who wants to use it. I think what we need most next is the expression for the amplitudes. I have code to generate all the kinematics. I've also done some generation with E852 amplitudes (spin 0 beam instead of spin 1).
>>>>>>
>>>>>> Once we get the expressions for the amplitudes, things should go pretty quickly. I was generating and fitting (with E852 amplitudes) just last night. Using the NVIDIA card on my laptop I was easily able to generate a couple hundred toy data sets with moderate statistics and fit them to examine pull distributions. With graphics acceleration, this pretty much becomes an interactive task, like fitting a histogram.
>>>>>>
>>>>>> If there is a small group of interested people, perhaps we could arrange a PWA powwow around the May mtg. I would be happy to organize a tutorial on the software. The more that people use it and complain, the faster it will get debugged!
>>>>>>
>>>>>> -Matt
>>>>>>
>>>>>>
>>>>>> On Mar 17, 2010, at 4:02 PM, Craig Bookwalter wrote:
>>>>>>
>>>>>>
>>>>>>> Hi folks,
>>>>>>> Last week at the offline software meeting there was some discussion about attempting a partial-wave analysis with currently existing software and simulated data, both to build some expertise locally at JLab as well as to flesh out where effort needs to be focused in the PWA development. Assuming that the folks at Indiana doing the bulk of this work are in favor of this idea, it seems the first order of business would be to generate some monte carlo with a set of resonances decaying with all the appropriate angular information. I seem to remember Adam Szczepaniak doing something like this for me a few years ago when I first started playing around with the Ruby fitter--does that code still exist? Can we bring it into the GlueX repository?
>>>>>>>
>>>>>>> Thanks in advance.
>>>>>>>
>>>>>>> --cb
>>>>>>>
>>>>>>> --
>>>>>>> -------------------------------------------------------------------------
>>>>>>> Craig Bookwalter FSU Office: (850) 644 3808
>>>>>>> Department of Physics JLab Office: (757) 269 5465
>>>>>>> Florida State University craigb at hadron.physics.fsu.edu
>>>>>>> Tallahasse, FL 32306 craigb at jlab.org
>>>>>>>
>>>>>>>
>>>>>>> "One toke? You poor fool. Just wait till you see those (expletive) bats."
>>>>>>> -------------------------------------------------------------------------
>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> Halld-physics mailing list
>>>>>>> Halld-physics at jlab.org
>>>>>>> https://mailman.jlab.org/mailman/listinfo/halld-physics
>>>>>>>
>>>>>>
>>>>> --
>>>>> -------------------------------------------------------------------------
>>>>> Craig Bookwalter FSU Office: (850) 644 3808
>>>>> Department of Physics JLab Office: (757) 269 5465
>>>>> Florida State University craigb at hadron.physics.fsu.edu
>>>>> Tallahasse, FL 32306 craigb at jlab.org
>>>>>
>>>>>
>>>>> "One toke? You poor fool. Just wait till you see those (expletive) bats."
>>>>> -------------------------------------------------------------------------
>>>>>
>>>>>
>>>>
>>> --
>>> -------------------------------------------------------------------------
>>> Craig Bookwalter FSU Office: (850) 644 3808
>>> Department of Physics JLab Office: (757) 269 5465
>>> Florida State University craigb at hadron.physics.fsu.edu
>>> Tallahasse, FL 32306 craigb at jlab.org
>>>
>>>
>>> "One toke? You poor fool. Just wait till you see those (expletive) bats."
>>> -------------------------------------------------------------------------
>>>
>>>
>>
>>
>>
>
>
> --
> -------------------------------------------------------------------------
> Craig Bookwalter FSU Office: (850) 644 3808
> Department of Physics JLab Office: (757) 269 5465
> Florida State University craigb at hadron.physics.fsu.edu
> Tallahasse, FL 32306 craigb at jlab.org
>
>
> "One toke? You poor fool. Just wait till you see those (expletive) bats."
> -------------------------------------------------------------------------
>
More information about the Halld-offline
mailing list