Toolkit for partonic process data
storage and manipulation
Balatenichev S., Cherstnev A., Ilyin V.
(SINP MSU)
Reasons and ideas of the toolkit
New standard of partonic events storage
Toolkit utilities
Current status of the toolkit development
24-28 june 2002
ACAT'02
1
Why do we need this toolkit?
I. Modern complexity of MC generation has resulted
in separation of 2 stages of the generation: ME
(e.g. CompHEP) and SH (e.g. PYTHIA)
One needs interface between these stages
II. Many collaborations have and develop databases
of Monte-Carlo events at hard partonic level.
Since MC generation in many cases is heavy task,
changes of the data is faster to do by direct
operations with the events without re-generation.
Toolkit of utilities for these manipulation
24-28 june 2002
ACAT'02
2
Structure of the toolkit
Visualization (Root)
ME
(CompHEP)
Partonic
events
database
SH
(PYTHIA)
Data manipulation
(mixing, updating, cuts and etc.)
24-28 june 2002
ACAT'02
3
New data storage standard
Text file format is optimal for partonic events
storage, in our opinion. Thus, a some
standardization is needed for these files.
The standard we proposed gives flexibility for
data description and manipulations.
Specific points for this standard are:
• A physical information from ME to SH is highly
heterogeneous. Our standard allows to transfer different
types of the information within uniform syntax.
• We assume that SH generator or a manipulation utility
will take that information transferred which can
interpret.
24-28 june 2002
ACAT'02
•
4
New data storage standard
Text file of partonic events
Header of the file
Description
of physical information:
beams, processes, etc.
Partonic
event records
Event 1
Event 2
…
24-28 june 2002
ACAT'02
5
Header Syntax
Basic unit of information in the header – command
Syntax: command name = command value
Command value can be of numerical type or string
(in that case it encloses in quotes).
Some commands can be grouped under tags.
Tag syntax: ##Tag name: command 1,…, command N,;
1.Double # – the beginning of the tag
2. semicolon – the end of the tag
The header information is stored as a set of tags.
The tags are separated from event records by extra
semicolon.
24-28 june 2002
ACAT'02
6
Example of the tags set
##generator: name=‘CompHEP’, version=‘4.2.00’,;
##beam: energy=1.000E+03, KF=2212, name=‘proton', mass=9.380E-01,;
##strfun: name='CTEQ', version='5m1', PDFid=833, PDFgr=11,
PDFfile=‘$PDF/cteq5m1.pdf',;
##process: ID=1, name='u D -> E1 n1 b B ', CrosSec=3.69595E-02,
CrosSecErr=4.61224E-03, Nparton=6, master=3,;
##n_event: IDprocess=1, N=1000, multN=1, maxW=2.96610E+00,;
##parton: IDprocess=1, in=1, KF=2, name='u', mass=0.00000E+00,;
##parton: IDprocess=1, out=2, KF=11, name=‘n1', mass=0.00000E+00,;
##total: Nprocess=1, CrosSec=3.69595E-02, CrosSecErr=4.61224E-03,
Nevent=1000,;
##QCDprocess: IDprocess=1, NL=2, Qnorm=9.11870E+01,
QCDalpha=1.18108E-01, file_name='$GEN/c_source/num/alpha_s.c',
func_name='alpha_c(int q)',;;
24-28 june 2002
ACAT'02
7
Event record
Event record occupies one line in the data file.
1. For the interpretation of event records we introduce special
tag:
##format: IDprocess=5, ProcNumber='i', p1.3='13.6E', p2.3='13.6E',
p3.1='13.6E', p3.2='13.6E', p3.3='13.6E', p4.1='13.6E', p4.2='13.6E',
p4.3=‘13.6E', Q='10.3E', color_chain=‘string’,;
2. event record consists of statements divided with colons.
Event record example:
1: 2.825664E+02:-1.593711E+02: 8.204448E+01:
-1.468924E+00: 6.821391E+01:-1.024018E+01: 3.039867E+01:
2.797844E+01:9.119E+01:(1 2)(6 5)
24-28 june 2002
ACAT'02
8
Advantages of the standard
1. Our text file standard corresponds to Les
Houches generic format for events storage in
FORTRAN common blocks (hep/ph-0109068).
2. Immediate advantages are:
 Independence of programming languages for
ME and SH generators.
 readability
 Simple realization.
24-28 june 2002
ACAT'02
9
Connection to Les Houches format
We re-write FORTRAN Les Houches common
blocks to C structures. Example is:
#define maxproc 100
typedef struct procinfo
{ int id_beam[2];
/* KF of beam particle */
double ebmup[2];
/* Beam energy */
int pdf_group[2], pdf_set[2];
/* PDF information */
int master, n_proc, proc_list;
/* Master switch, number of subp
double cs[maxproc], cs_err[maxproc], cs_max[maxproc ] /* Cross sec
int list_prc[maxproc], max_weight[maxproc];
/* Extra info
} procinfo;
Toolkit writer(reader) writes(reads) data to(from) datafile
from(to) these structures or FORTRAN common blocks.
24-28 june 2002
ACAT'02
10
Toolkit utilities
Storage of partonic events data in text files of this
standard allows to manipulate easily with the data.
We suppose that the toolkit will include utilities:
• Mixing and de-mixing
in practice, we generate
several files with partonic events and would like to
mix them to one event flow and visa versa.
• Apply kinematical cuts
phenomenological
tasks require to apply kinematical cuts on events.
Storage of data in our format allows to develop a
utility for cut applying.
24-28 june 2002
ACAT'02
11
Toolkit utilities
• Event visualization
we intend to make some
utilities for remote visualization of process
information and histogamming of events and etc.
• Updating
this utility will allow to re-weigh
events by MC algorithm. If distribution of new
events weights is sufficiently “smooth” function,
not very large part of the events statistic will be
lost. For instance, this utility is convenient for
updating of structure functions.
• Rewriter of old events files
this utility rewrites
events generated by CompHEP of old versions
into new standard.
24-28 june 2002
ACAT'02
12
Current status of toolkit
At present the toolkit is in test stage.
We have developed:
• Reader and writer to Les Houches common
blocks (C structures) on C and FORTRAN
• Mixing utility
• CompHEP 4.2 outputs data in the standard
My supposition: the toolkit will be in site
http://theory.sinp.msu.ru/…
In next some months.
24-28 june 2002
ACAT'02
13
Descargar

Toolkit for parton process data manipulation