source: HiSusy/trunk/Pythia8/pythia8170/htmldoc/Tunes.html @ 1

Last change on this file since 1 was 1, checked in by zerwas, 11 years ago

first import of structure, PYTHIA8 and DELPHES
File size: 15.4 KB
Line 
1<html>
2<head>
3<title>Tunes</title>
4<link rel="stylesheet" type="text/css" href="pythia.css"/>
5<link rel="shortcut icon" href="pythia32.gif"/>
6</head>
7<body>
8
9<h2>Tunes</h2>
10
11Since some physics aspects cannot be derived from first principles,
12this program contains many parameters that represent a true
13uncertainty in our understanding of nature. Particularly afflicted
14are the areas of hadronization and multiparton interactions, which both
15involve nonperturbative QCD physics.
16
17<p/>
18Technically, PYTHIA  parameters can be varied independently of each
19other, but the physical requirement of a sensible description of a set
20of data leads to correlations and anticorrelations between the
21parameters. Hence the need to produce tunes, not of one parameter at 
22a time, but simultaneously for a group of them. A well-known (separate)
23such example is parton densities, where combined tunes to a wide range
24of data have been produced, that can then be obtained prepackaged. 
25
26<p/>
27Given the many PYTHIA parameters to be tuned, it is convenient to
28divide the task into subtasks. Firstly, if we assume jet universality,
29hadronization and final-state parton showers should be tuned to
30<i>e^+e^-</i> annihilation data, notably from LEP1, since this
31offers the cleanest environment. Secondly, with such parameters fixed,
32hadron collider data should be studied to pin down multiparton interactions
33and other further aspects, such as initial-state radiation. Ideally this
34would be done separately for diffractive and non-diffractive events,
35although it is not possible to have a clean separation. (Thirdly
36would come anything else, such as physics with photon beams, which
37involve further parameters, but that is beyond the current scope.)
38
39<p/>
40The first step in this program has now been taken, with a tune to LEP1
41data by Hendrik Hoeth, using the Rivet + Professor framework. Starting
42with version 8.125 it defines the default values for hadronization
43parameters and timelike showers.
44
45<p/>
46The situation is more complicated for hadronic interactions in general
47and multiparton interactions in particular, where PYTHIA 8 is more
48different from PYTHIA 6, and therefore more work is needed. Specifically,
49it is <i>not</i> possible to "port" a PYTHIA 6 tune to PYTHIA 8.
50
51<p/>
52A first simple tune, appropriately called "Tune 1", became default
53starting with version 8.127. It was noted, in particular by Hendrik
54Hoeth, that this tune had a tension between parameters needed to
55describe minimum-bias and underlying-event activity. Therefore some
56further physics features were introduced in the code itself
57[<a href="Bibliography.html" target="page">Cor10a</a>], which were made default as of 8.140. This version
58also included two new tunes, 2C and 2M, based on the CTEQ 6L1 and the
59MRST LO** PDF sets, respectively. These have been made by hand, as a
60prequel to complete Professor-style tunings.
61
62<p/>
63The very first data to come out of the LHC showed a higher rapidity
64plateau than predicted for current PYTHIA 6 tunes, also for the lower
65energies. This may suggest some tension in the data. Two alternatives,
663C and 3M, were produced by a few brute-force changes of 2C and 2M.
67These were introduced in 8.140, but discontinued in 8.145 in favour of
68the new 4C tune, that is based on a more serious study of some early
69LHC data, see [<a href="Bibliography.html" target="page">Cor10a</a>]. Following the comparative studies in
70[<a href="Bibliography.html" target="page">Buc11</a>], which independently confirmed a reasonable agreement
71with LHC data, tune 4C was made the default as of 8.150. A variant is
72tune 4Cx, where the Gaussian matter profile has an <i>x</i>-dependent
73width [<a href="Bibliography.html" target="page">Cor11</a>].
74
75<p/>
76Several ATLAS tunes have now been included, obtained with different PDFs
77and with different emphasis on minimum-bias and underlying-event data
78[<a href="Bibliography.html" target="page">ATL12</a>]. These typically require LHAPDF to be linked, but this
79can be avoided in cases where the same PDF set is implemented internally.
80
81<p/>
82Central diffraction is a recent addition to the "soft QCD" process palette,
83and is thus not yet included in tunes; indeed its cross section is actively
84zeroed. You can switch it back on <i>after</i> you have selected your tune,
85with <code>SigmaTotal:zeroAXB = off</code>. But note that, since the
86total cross section is assumed unchanged, the minbias cross section
87is reduced and thus also the MPI machinery affected, even if effects
88should not be big (for a small central diffractive cross section).
89
90<p/>
91Note that comparisons with data also require that other aspects agree,
92such as that decay chains are stopped at an agreed-on level. For instance,
93in the ATLAS tunes all particles with a lifetime above 10 mm
94are considered stable, <code>ParticleDecays:limitTau0 = on</code>,
95<code>ParticleDecays:tau0Max = 10</code>. We have chosen not to
96include this as part of the tune settings itself, since the tune as
97such could still be used with any other choice of stable and
98unstable particles.
99
100<p/>
101Further comparisons have been posted on the
102<a href="http://mcplots.cern.ch/">MCPLOTS</a> pages.
103They have been produced with help of the
104<a href="http://projects.hepforge.org/rivet/">Rivet</a> package
105[<a href="Bibliography.html" target="page">Buc10</a>].
106
107<p/>
108In the future we hope to see further PYTHIA 8 tunes appear. Like with
109parton distributions, there is likely to be several tunes, because
110different sets of data will pull in different directions, by imperfections   
111in the model or in the data, and by differences in the chosen
112tuning strategies. We therefore propose to collect some of these tunes
113here, in a prepackaged form. Of course, in all cases it is a matter
114of setting values for parameters already defined elsewhere, so the
115tunes offer no new functionality, only a more convenient setup.
116
117<p/>
118You should be aware that the evolution of the program will not guarantee
119complete backwards compatibility between versions. Most obviously this
120concerns bug fixes. But also for some other major changes, like the
121introduction of the new diffractive machinery, the default behaviour
122of old tunes has been changed retroactively. (Which should be fine for
123diffraction, since previous tunes were not based on data strongly
124influenced by diffraction.) 
125
126<p/>
127The setup of the tunes is special, in that the choice of a tune forces
128the change of several different flags, modes and parameters. Furthermore
129a design principle has been that it should be possible to start out
130from a tune and then change a few of its settings. This gives power
131and flexibility at the expense of requiring a more careful ordering
132of commands. We therefore here sketch the order in which operations
133are carried out.
134<ol>
135<li>The constructor of a <code>Pythia</code> instance will read in
136all settings, and initialize them with their default values.
137</li>
138<li>At the end of this operation, the <code>Tune:ee</code> and
139<code>Tune:pp</code> modes (see further below) are checked. If either
140of them are positive the methods <code>Settings::initTuneEE(...)</code>
141and <code>Settings::initTunePP(...)</code>, respectively, are called
142to overwrite the whole collection of settings in the relevant tune.
143Zero (or negative) means that nothing will be done.
144</li> 
145<li>After the <code>Pythia</code> constructor all the relevant values
146for the default tune(s) have thus been set up.
147</li>
148<li>You as a user can now start to overwrite the values at will,
149using <code>Pythia::readFile(...)</code> to read a configuration file,
150or a list of <code>Pythia::readString(...)</code> commands,
151or the lower-level <code>Settings</code> methods. All changes
152are made in the order in which the commands are encountered during
153the execution. A given variable can be changed multiparton times,
154but it is the latest change that sets the current value.
155</li>
156<li>The two <code>Tune:ee</code> and <code>Tune:pp</code> modes can also
157be changed in exactly the same way as described for all other settings
158above. Unique for them, however, is that when one of them is encountered
159it also initiates a call to the <code>initTuneEE(...)</code> or
160<code>initTunePP(...)</code> method, respectively. In such cases all
161settings affected by the <i>e^+e^-</i> or <i>pp/ppbar</i> tune
162are first reset to the default values (the <code>-1</code> options)
163and thereafter the relevant tune is set up.
164</li>
165<li>It is possible to mix commands of type 4 and 5 in any order; it
166is always the last change that counts. That is, any changes you have
167made to variables of a tune <i>before</i> a <code>Tune:ee</code> or
168<code>Tune:pp</code> command are overwritten by it, while variables
169you set <i>after</i> will overwrite the tune values. As a rule,
170therefore, you want to begin with the tune choice, and thereafter
171modify only a small part of its settings.
172</li>
173<li>Needless to say, the flexibility can lead to unwanted setups if
174you do not exercise some discipline. It is therefore recommended that
175you always check the listing obtained with
176<code>Pythia::settings.listChanged()</code> to confirm that the
177final set of changes is the intended one.
178</li>
179</ol>
180
181<p/><code>mode&nbsp; </code><strong> Tune:ee &nbsp;</strong> 
182 (<code>default = <strong>0</strong></code>; <code>minimum = -1</code>; <code>maximum = 3</code>)<br/>
183Choice of tune to <i>e^+e^-</i> data, mainly for the hadronization
184and timelike-showering aspects of PYTHIA. You should study the
185<code>Settings::initTuneEE(...)</code> method to find exactly which
186are the settings for the respective tune.
187<br/><code>option </code><strong> -1</strong> : reset all values that are affected by any of the
188<i>e^+e^-</i> tunes to the default values. This option can be used
189on its own, but is also automatically used as a first step for either
190of the positive tune values below, to undo the effect of previous tune
191settings.
192 
193<br/><code>option </code><strong> 0</strong> : no values are overwritten during the initial setup,
194step 2 above. Note that changing to <code>0</code> in the user code
195has no effect; if you want to restore the individual settings you
196should instead use <code>-1</code>.
197 
198<br/><code>option </code><strong> 1</strong> : the original PYTHIA 8 parameter set, based on some
199very old flavour studies (with JETSET around 1990) and a simple tune
200<i>of alpha_strong</i> to three-jet shapes to the new
201<i>pT</i>-ordered shower. These were the default values before
202version 8.125.
203 
204<br/><code>option </code><strong> 2</strong> : a tune by Marc Montull to the LEP 1 particle
205composition, as published in the RPP (August 2007). No related (re)tune
206to event shapes has been performed, however. 
207 
208<br/><code>option </code><strong> 3</strong> : a tune to a wide selection of LEP1 data by Hendrik
209Hoeth within the Rivet + Professor framework, both to hadronization and
210timelike-shower parameters (June 2009). These are the default values
211starting from version 8.125, so currently there is no need for this
212option.
213 
214 
215
216<p/><code>flag&nbsp; </code><strong> Tune:preferLHAPDF &nbsp;</strong> 
217 (<code>default = <strong>on</strong></code>)<br/>
218Tunes made by experimental collaborations typically use the LHAPDF
219package to obtain their PDF values, and so PYTHIA has to be linked
220accordingly. For PDFs implemanted natively in PYTHIA it is possible
221to use the respective tunes, without having to link to LHAPDF, if you
222set <code>Tune:preferLHAPDF = off</code> <i>before</i> the
223<code>Tune:pp</code> choice.
224 
225
226<p/><code>mode&nbsp; </code><strong> Tune:pp &nbsp;</strong> 
227 (<code>default = <strong>5</strong></code>; <code>minimum = -1</code>; <code>maximum = 11</code>)<br/>
228Choice of tune to <i>pp/ppbar</i> data, mainly for the
229initial-state-radiation, multiparton-interactions and  beam-remnants
230aspects of PYTHIA. Note that the previous crude (non-)tunes
2313C and 3M are removed as of 8.145, superseded by the 4C tune.
232You should study the <code>Settings::initTunePP(...)</code> method
233to find exactly which are the settings for the respective tune.
234<br/><code>option </code><strong> -1</strong> : reset all values that are affected by any of the
235<i>pp/ppbar</i> tunes to the default values. This option can be used
236on its own, but is also automatically used as a first step for either
237of the positive tune values below, to undo the effect of previous tune
238settings.
239 
240<br/><code>option </code><strong> 0</strong> : no values are overwritten during the initial setup,
241step 2 above. Note that changing to <code>0</code> in the user code
242has no effect; if you want to restore the individual settings you
243should instead use <code>-1</code>.
244 
245<br/><code>option </code><strong> 1</strong> : default used up to version 8.126, based on
246some early and primitive comparisons with data.
247 
248<br/><code>option </code><strong> 2</strong> : "Tune 1", default in 8.127 - 8.139, based on some
249data comparisons by Peter Skands. Largely but not wholly overlaps
250with the default option 0.
251 
252<br/><code>option </code><strong> 3</strong> : "Tune 2C", introduced with 8.140 [<a href="Bibliography.html" target="page">Cor10a</a>].
253It uses the CTEQ 6L1 PDF, and is intended to give good agreement with
254much of the published CDF data.
255 
256<br/><code>option </code><strong> 4</strong> : "Tune 2M", introduced with 8.140 [<a href="Bibliography.html" target="page">Cor10a</a>].
257It is uses the MRST LO** PDF, which has a momentum sum somewhat above
258unity, which is compensated by a smaller <i>alpha_s</i> than in the
259previous tune. Again it is intended to give good agreement with much of
260the published CDF data.
261 
262<br/><code>option </code><strong> 5</strong> : "Tune 4C", new tune, introduced with 8.145
263[<a href="Bibliography.html" target="page">Cor10a</a>]. Starts out from tune 2C, but with a reduced cross
264section for diffraction, plus modified multiparton interactions parameters
265to give a higher and more rapidly increasing charged pseudorapidity
266plateau, for better agreement with some early key LHC numbers.
267See also the comparative study in [<a href="Bibliography.html" target="page">Buc11</a>].
268 
269<br/><code>option </code><strong> 6</strong> : "Tune 4Cx", based on tune 4C, but using the x-dependent
270matter profile, <code>MultipartonInteractions:bProfile = 4</code> and an
271increased <code>MultipartonInteractions:pT0Ref</code> [<a href="Bibliography.html" target="page">Cor11</a>].
272 
273<br/><code>option </code><strong> 7</strong> : "ATLAS MB Tune A2-CTEQ6L1", a minimum-bias tune based
274on tune 4Cx, but without rapidity-ordered spacelike emissions
275[<a href="Bibliography.html" target="page">ATL12</a>]. Uses CTEQ 6L1, by default  from LHAPDF.
276 
277<br/><code>option </code><strong> 8</strong> : "ATLAS MB Tune A2-MSTW2008LO", as above,
278but uses MSTW 2008 LO, by default from LHAPDF.
279 
280<br/><code>option </code><strong> 9</strong> : "ATLAS UE Tune AU2-CTEQ6L1", an underlying-event tune
281based on tune 4Cx, but without rapidity-ordered spacelike emissions
282[<a href="Bibliography.html" target="page">ATL12</a>]. Uses CTEQ 6L1, by default from LHAPDF.
283 
284<br/><code>option </code><strong> 10</strong> : "ATLAS UE Tune AU2-MSTW2008LO", as above,
285but uses MSTW 2008 LO, by default from LHAPDF.
286 
287<br/><code>option </code><strong> 11</strong> : "ATLAS UE Tune AU2-CT10", as above,
288but uses CT 10, which is not currently implemented in PYTHIA,
289so you must link LHAPDF.
290 
291<br/><code>option </code><strong> 12</strong> : "ATLAS UE Tune AU2-MRST2007LO*", as above,
292but uses MRST 2007 LO*, by default from LHAPDF.
293 
294<br/><code>option </code><strong> 13</strong> : "ATLAS UE Tune AU2-MRST2007LO**", as above,
295but uses MRST 2007 LO**, by default from LHAPDF.
296 
297 
298
299</body>
300</html>
301
302<!-- Copyright (C) 2012 Torbjorn Sjostrand -->
Note: See TracBrowser for help on using the repository browser.