GSI Forum - RDF feed
https://forum.gsi.de/index.php
Minimizer Interface
https://forum.gsi.de/index.phpindex.php?t=rview&goto=12720&th=3377#msg_12720
in this post i'd like to discuss ideas for and requirements of an interface to minimizers like Geneva and Minuit2.
As a starter i will show a small first test i did for a toy-application:
Attached you can find a pdf with a visualization of the basic class structure and an example. As i do not have very much experience with general analysis packages, this interface i show here is meant as a playground.
For this first interface-draft i mainly tried to separate the interaction with the minimizers from the user-level. Therefore i have a virtual base-class for Minimizers called MIBase, which takes care that every implementation of it (with whatever minimizer) is able to deal with the same "data". In particular this means that the function that has to be minimized is provided by the user by implementing a class of type MIData. There he has to provide a function which then can be minimized. The minimizer-implementations (like MIMinuit in the example) need a MIData class to work with.
On the attached slide you have an small example on the left side where Geneva and/or Minuit2 are called to perform a polynomial fit to some smeared toy-data using the same parameters. Next to the example you find the classes used together with their base classes, with blue classes being part of the interface and just the orange one on the bottom is part of the user side. You can find this example also in the git-repository of the panda-pwa-tool in: Examples/Tutorial/FitIF/
As you can see in the example the parameters are just a list of doubles at the moment, i am lacking of a more flexible way to deal with the parameters of the minimization up to now. Also i didn't implement functionality to control the minimizers, which is technically not a problem but one has to check what options should be general for all minimizers and which ones are specific.
If you have questions, suggestions or some requirements (to make it suitable for the BES-software a.e.), feel free to discuss it here, i'd be glad about some feedback and ideas.
Cheers
Mathias ]]>Mathias Michel2011-10-14T10:36:11-00:00PANDA PWA EVO Meeting: Tuesday (11.10.2011) at 10h30
https://forum.gsi.de/index.phpindex.php?t=rview&goto=12701&th=3370#msg_12701
on tuesday next week (11.10.2011), there will be an EVO meeting at 10h30 (Central European Summer Time). This meeting is focussed on the joint PWA activities between PANDA and BES. The major topic is therefore a discussion on how to make use of GENEVA on GPU based PWA applications. If you are insterested in this discussion you are very welcome to attend this meeting. Please make sure that EVO is working properly on your machine and test it beforehand. The relevant booking information can be found below.
With kind regards,
Bertram.
p.s. EVO booking info
Title: PANDA BES PWA
Description: PANDA BES PWA
Community: Panda
Password: pawian
- Skype (tm) (World-wide)
evo.phone
See: http://evo.caltech.edu/evoGate/Documentation/extclient/skype/skype.html]]>Bertram Kopf2011-10-06T09:59:45-00:00(PANDA) PWA Lecture on Thursday (16.12.2010) at GSI
https://forum.gsi.de/index.phpindex.php?t=rview&goto=11305&th=2998#msg_11305
as decided few weeks ago, the second part of the (PANDA) PWA Lecture, given by Klaus Peters, will take place on next thursday at GSI in room KP1/3.
The major topics are the different PWA formalisms. The lecture is divided in two sessions, a morning session (10h30-12h30) and an afternoon session from 14h00 till 16h00.
Afterwards an one hour PWA meeting will follow, where status reports of the recent PWA activities will be presented. The meeting will therefore be closed at 17h00.
Best regards,
Bertram. ]]>Bertram Kopf2010-12-13T13:14:08-00:00PANDA PWA EVO Meeting: Tomorrow (20.07.2010) at 14h00
https://forum.gsi.de/index.phpindex.php?t=rview&goto=10876&th=2890#msg_10876
the next PANDA PWA EVO Meeting will take place tomorrow (20.07.2010) at 14h00. Since we have to close the meeting not later than 15h00, I would like to ask everyone to login on time.
Tentative agenda:
o recent software development activities
o upcoming tasks
o first PWA results
o upcoming PWA Workshop at Mainz
Best regards,
Bertram.
p.s.: EVO booking info:
Title: PWA
Description: PANDA PWA EVO Meeting
Community: Panda
]]>Bertram Kopf2010-07-19T09:09:38-00:00Next PANDA PWA EVO Meeting on Tuesday, 6.07.2010, at 14h00
https://forum.gsi.de/index.phpindex.php?t=rview&goto=10848&th=2879#msg_10848
as already announced one week ago, the next PWA EVO Meeting will take place on next Tuesday at 14h00.
Tentative topics:
o recent software development activities
o upcoming PWA Workshop at Mainz
o miscellaneous
With best regards,
Bertram.
p.s.: EVO booking info:
Title: PWA
Description: PANDA PWA EVO Meeting
Community: Panda
- France (RAP, Paris)
+33 144 27 81 50
]]>Bertram Kopf2010-07-04T10:48:58-00:00Next PANDA PWA EVO Meeting postponed by one week
https://forum.gsi.de/index.phpindex.php?t=rview&goto=10835&th=2874#msg_10835
the next PANDA PWA Meeting was planned for tomorrow (29.06.2010) at 14h00. Due to some obligations we have to postpone it by one week.
The meeting is now scheduled for tuesday, 06.07.20010, at 14h00.
If you are not able to take part in this meeting, please send me a short email.
Best regards,
Bertram.]]>Bertram Kopf2010-06-28T08:57:31-00:00Geneva available in PANDA PWA software
https://forum.gsi.de/index.phpindex.php?t=rview&goto=10757&th=2852#msg_10757
the latest HEAD version contains "Geneva" as an external package. I.e. everything related to Geneva should be accessible inside our code. Up to now you can start one Geneva test application by typing: ./GMultiPopulation -c $CMAKE_SOURCE_DIR/Examples/geneva/GMultiPopulation/GMultiPopulation.cfg
Note: It is now mandatory to install Geneva and to set the environment variable "GENEVA_ROOT" before starting the cmake procedure.
Best regards,
Bertram.]]>Bertram Kopf2010-06-02T09:48:30-00:00PANDA PWA EVO Meeting: Tomorrow (01.06.10) at 14h00
https://forum.gsi.de/index.phpindex.php?t=rview&goto=10754&th=2850#msg_10754
the next PANDA PWA EVO Meeting will take place tomorrow, June 1st, at 14h00. The booking information can be found below.
Tentative agenda:
o discussion on the name of the software
o status reports of the recent activities
o discussion on the event class concept
o AOB
Any other topics to discuss?
Best regards,
Bertram.
p.s. EVO booking info
Title: PWA
Description: PANDA PWA EVO Meeting
Community: Panda
---------------
If you would like to add a telephone bridge at your institute, please contact EVO at evocontact@vrvs.org]]>Bertram Kopf2010-05-31T09:04:55-00:00Event Class Brainstorming
https://forum.gsi.de/index.phpindex.php?t=rview&goto=10717&th=2840#msg_10717
... and sorry for starting this topic so late...
Here are some of the ideas and concepts that we at COMPASS came up with for the design of an event class for PWA:
* Prime responsibility is kinematics
* event-data SEPARATED from amplitude calculator
* Event weight
* Meta event information is useful
In more detail:
Prime responsibility of the event class is to contain/provide all the kinematical information that is necessary to calculate amplitudes (in a given formalism). This obviously includes the 4-vectors of all involved particles and some meta-information which further defines the particles (species, quantum numbers, pid-info etc).
It also makes sense to put kinematical manipulations of subsystems of the particles here. Including the necessary lorentz-transforms (possibly with a caching mechanism).
At compass so far we use simple ascii files to store the info. This is read into a custom class. We will switch (rootpwa) to root-persistence in the near future.
What should be SEPARATED from the event is the actual computation of an amplitude, because this might depend on the formalism.
One issue that complicates the story a bit is the fact that for different formalisms but especially for different production mechanisms one needs different information (especially on the initial state.) This means that the amplitude calculator in principle decides which information it needs from the event. So the event class (if you do not want to write a special one for each case) needs some flexibility. There are obviously different technical solutions to this.
The event class must be able to carry a weight needed for several statistical tricks...
The event may contain meta-information such as trigger-flags, event selection cut settings etc which might be useful to still have available on the level of PWA.
So far so good...
Happy discussion!
Cheers! Sebastian.
]]>Sebastian Neubert2010-05-20T16:11:59-00:00Voting: Name for the PANDA PWA software
https://forum.gsi.de/index.phpindex.php?t=rview&goto=10716&th=2839#msg_10716
due to the fact that no further ideas have been raised, I would suggest to start here the poll on the name of the PWA software.
At the last EVO meeting we decided to have two stages of voting.
In the first stage - it will start right now until next Thursday - we will vote for three out of six candidates. Then in the second stage, we will finally vote for the best name of the remaining three candidates.