1 <chapter name="Advanced Usage">
3 <h2>Advanced Usage</h2>
5 On this page we collect information on a number of classes that
6 the normal user would not encounter. There are cases where the
7 information is essential, however, for instance to
8 <aloc href="ImplementNewShowers">implement your own showers</aloc>.
10 <h3>The subsystems</h3>
12 One aspect that complicates administration is that an event
13 can contain several subsystems, each consisting of one MPI and its
14 associated ISR and FSR. To first approximation these systems are
15 assumed to evolve independently, but to second they are connected by
16 the interleaved evolution, and potentially by rescattering effects.
17 The partons of a given subsystem therefore do not have to be stored
21 The <code>PartonSystems</code> class is primarily used to keep track
22 of the current positions of all partons belonging to each system,
23 represented by the index <code>iPos</code> for a parton stored in the
24 event-record slot <code>event[iPos]</code>. With "all" we mean the
25 currently defined two incoming partons, or none for a resonance decay,
26 and the current set of outgoing partons, but with all ISR and FSR
27 intermediate-state partons omitted. That is, it stores all partons
28 that could be subject to some action in the next step of the
29 combined MPI/ISR/FSR/BR description. As a special case, an outgoing
30 parton is stored even if it undergoes a rescattering, and thus no
31 longer belongs to the final state proper.
34 The <code>partonSystems</code> instance of <code>PartonSystems</code>
35 class is a public member of the <code>Pythia</code> top-level class,
36 but is also available as a pointer <code>partonSystemsPtr</code> in
37 various <code>PartonLevel</code> classes, e.g. inside the current
38 instances of the <code>TimeShower</code> and <code>SpaceShower</code>
42 A number of <code>PartonSystems</code> methods can be used to set or
43 get information on the subsystems:
45 <li><code>clear()</code> resets all the contents in preparation for the
47 <li><code>addSys()</code> add a new (initially empty) subsystem to the
48 current list and return its index <code>iSys</code> in the list,
49 where index 0 is the hardest subcollision and so on.
50 <li><code>sizeSys()</code> the number of separate subsystems.</li>
51 <li><code>setInA(iSys, iPos), setInB(iSys, iPos)</code> store position
52 <code>iPos</code> of the incoming parton from beam A or beam B to the
53 <code>iSys</code>'th subcollision. These values are 0 initially, and
54 should so remain if there are no beams, such as in resonance decays.
56 <li><code>addOut(iSys, iPos)</code> store position <code>iPos</code>
57 of a new outgoing parton in the <code>iSys</code>'th subcollision,
58 by appending it at the end of the current vector, with beginning in
60 <li><code>setOut(iSys, iMem, iPos)</code> store position <code>iPos</code>
61 in the <code>iMem</code>'th slot in the vector of outgoing partons in the
62 <code>iSys</code>'th subcollision. Here <code>iMem</code> must be in
63 the range already constructed by <code>addOut</code> calls. </li>
64 <li><code>replace(iSys, iPosOld, iPosNew)</code> replace the existing
65 incoming or outgoing parton position <code>iPosOld</code> by
66 <code>iPosNew</code> in the <code>iSys</code>'th subcollision.</li>
67 <li><code>setSHat(iSys, sHat)</code> set the invariant squared mass
68 <code>sHat</code> of the <code>iSys</code>'th subcollision.</li>
69 <li><code>hasInAB(iSys)</code> true if an incoming parton has been set
70 for beam A or beam B (and hence should have been set for both) in the
71 <code>iSys</code>'th subcollision, else false.</li>
72 <li><code>getInA(iSys), getInB(iSys)</code> the position <code>iPos</code>
73 of the incoming parton from beam A or beam B to the <code>iSys</code>'th
75 <li><code>sizeOut(iSys)</code> the number of outgoing partons
76 in the <code>iSys</code>'th subcollision.</li>
77 <li><code>getOut(iSys, iMem)</code> the position <code>iPos</code>
78 of an outgoing parton in the <code>iSys</code>'th subcollision,
79 with the <code>iMem</code> range limited by <code>sizeOut(iSys)</code>.
80 These partons are not guaranteed to appear in any particular order. </li>
81 <li><code>sizeAll(iSys)</code> the total number of incoming and outgoing
82 partons in the <code>iSys</code>'th subcollision.</li>
83 <li><code>getAll(iSys, iMem)</code> the position <code>iPos</code>
84 of an incoming or outgoing parton in the <code>iSys</code>'th subcollision.
85 In case there are beams it gives same as <code>getInA(iSys) </code> and
86 <code> getInB(iSys)</code> for indices 0 and 1, and thereafter agrees with
87 <code>getOut(iSys, iMem)</code> offset two positions. If there are no
88 beams it is identical with <code>getOut(iSys, iMem)</code>.</li>
89 <li><code>getSHat(iSys)</code> the invariant squared mass
90 <code>sHat</code> of the <code>iSys</code>'th subcollision.</li>
91 <li><code>list()</code> print a listing of all the system information,
92 except for the <code>sHat</code> values.</li>
96 New systems are created from the hard process and by the MPI, not from
97 any of the other components. Both FSR and ISR modify the position
98 of partons, however. Since an FSR or ISR branching typically implies
99 a new state with one more parton than before, an outgoing parton must
100 be added to the system. Furthermore, in a branching, several existing
101 partons may also be moved to new slots, including the incoming beam ones.
102 In a FSR <ei>1 -> 2</ei> branching it is irrelevant which parton position
103 you let overwrite the existing slot and which is added to the end of
107 The system information must be kept up-to-date. Both the MPI, ISR, FSR and
108 BR descriptions make extensive use of the existing information. As an
109 example, the introduction of primordial <ei>kT</ei> in the beam remnants
110 will fail if the information on which final-state partons belong to which
111 system is out-of-date. The introduction of rescattering as part of the
112 MPI framework adds further complications, where an outgoing parton of one
113 subsystem may be the incoming one of another system. This part of the code
114 is still under development.
117 Currently the system information is kept throughout the continued
118 history of the event. Specifically, resonance decays create new systems,
119 appended to the existing ones. This could be useful during the
120 hadronization stage, to collect the partons that belong to a resonace
121 with preserved mass when a small string collapses to one particle,
122 but is not yet used for that.
126 The different subsystems are tied together by them sharing the same
127 initial beam particles, and thereby being restricted by energy-momentum
128 and flavour conservation issues. The information stored in the two
129 beam particles, here called <code>beamA</code> and <code>beamB</code>,
130 is therefore as crucial to keep correct as the above subsystem list.
133 Both beam objects are of the <code>BeamParticle</code> class.
134 Each such object contains a vector with the partons extracted from it.
135 The number of such partons, <code>beamX.size()</code> (X = A or B),
136 of course is the same as the above number of subsystems in the event
137 record. (The two diverge at the BR step, where further beam remnants
138 are added to the beams without corresponding to new subsystems.)
139 The individual partons are accessed by an overloaded indexing
140 operator to a vector of <code>ResolvedParton</code> objects. The
141 <code>iPos()</code> property corresponds to the <code>iPos</code>
142 one above, i.e. providing the position in the main event record of
143 a parton. In particular,
144 <code>beamA[iSys].iPos() = partonSystemsPtr->getInA(iSys)</code> and
145 <code>beamB[iSys].iPos() = partonSystemsPtr->getInB(iSys)</code>.
146 Whereas thus the indices of the two incoming partons to a subsystem
147 are stored in two places, the ones of the outgoing partons only
148 appear in the system part of the <code>PartonSystems</code> class.
151 Just as the subsystems in <code>PartonSystems</code> must be updated,
152 so must the information in the two <code>BeamParticle</code>'s, e.g.
153 with methods<code>beamX[iSys].iPos( iPosIn)</code> when an incoming
154 parton is replaced by a new one in line <code>iPosIn</code>. Furthermore
155 the new parton identity should be set by <code>beamX[iSys].id( idIn)</code>
156 and the new <ei>x</ei> energy-momentum fraction by
157 <code>beamX[iSys].x( xIn)</code>. The three can be combined in one go
158 by <code>beamX[iSys].update( iPosIn, idIn, xIn)</code>.
161 To be specific, it is assumed that, at each step, the two incoming
162 partons are moving along the <ei>+-z</ei> axis and are massless.
163 Since the event is constructed in the c.m. frame of the incoming
164 beams this implies that <ei>x = 2 E / E_cm</ei>.
165 If the <ei>x</ei> values are not defined accordingly or not kept
166 up-to-date the BR treatment will not conserve energy-momentum.
169 In return, the <code>BeamParticle</code> objects give access to some
170 useful methods. The <code>beamX.xf( id, x, Q2)</code> returns the
171 standard PDF weight <ei>x f_id(x, Q^2)</ei>. More intererstingly,
172 <code>beamX.xfISR( iSys, id, x, Q2)</code> returns the modified weight
173 required when several subsystems have to share the energy and flavours.
174 Thus <code>iSys</code> is added as an extra argument, and the momentum
175 already assigned to the other subsystems is not available for evolution,
176 i.e. the maximal <ei>x</ei> is correspondingly smaller than unity.
177 Also flavour issues are handled in a similar spirit.
180 An additional complication is that a parton can be either valence or
181 sea, and in the latter case the BR treatment also distinguishes
182 companion quarks, i.e. quark-antiquark pairs that are assumed to
183 come from the same original <ei>g -> q qbar</ei> branching, whether
184 perturbative or not. This can be queried either with the
185 <code>beamX[iSys].companion()</code> method, for detailed information,
186 or with the <code>beamX[iSys].isValence()</code>,
187 <code>beamX[iSys].isUnmatched()</code> and
188 <code>beamX[iSys].isCompanion()</code> metods for yes/no answers
189 whether a parton is valence, unmatched sea or matched sea.
190 This choice should affect the ISR evolution; e.g., a valence quark
191 cannot be constructed back to come from a gluon.
194 To keep this info up-to-date, the <code>beamX.pickValSeaComp()</code>
195 method should be called whenever a parton of a new flavour has been
196 picked in the ISR backwards evolution, but not if the flavour has not
197 been changed, since then one should not be allowed to switch back and
198 forth between the same quark being considered as valence or as sea.
199 Since the <code>pickValSeaComp()</code> method makes use of the
200 current parton-density values, it should be preceded by a call
201 to <code>beamX.xfISR( iSys, id, x, Q2)</code>, where the values in
202 the call are the now finally accepted ones for the newly-found mother.
203 (Such a call is likely to have been made before, during the evolution,
204 but is not likely to be the most recent one, i.e. still in memory, and
205 therefore had better be redone.)
209 <!-- Copyright (C) 2012 Torbjorn Sjostrand -->