Www.QinetiQ.com © Copyright QinetiQ limited 2007 1 DPMJET-2.5 & DPMJET-3 Treats hadron-nuclear and nuclear-nuclear interactions >5 GeV/nuc, with the upper.

www.QinetiQ.com

1© Copyright QinetiQ limited 2007

DPMJET-2.5 & DPMJET-3

• Treats hadron-nuclear and nuclear-nuclear interactions >5 GeV/nuc, with the upper limited reported to be of order 1000TeV

• Two versions currently available, both of which treat nuclear-nuclear interactions:

− DPMJET-2.5 (Johannes Ranft) - source code publicly released

− DPMJET-3 (Stefan Roesler) - access to source controlled by Roesler

• Both versions of DPMJET available in FLUKA-2006, but distributed as compiled libraries

• A version of DPMJET-II.5 has been implemented as the model G4DPMJET2_5Model and is being tested. There have been challenges:

− Very limited documentation on the code, none covering explanation of design / organisation

− How to handling Glauber data generated as a part of the DPMJET-II.5 initialisation process

www.QinetiQ.com


What do (stored) Glauber data refer to?

Target

Projectile

b

• Cumulative probability distribution function

• Used to sample impact parameter, once it is confirmed an impact has occurred

• Unique to each projectile (AZNP) and

target (AZNT) and must be generated as

a function of momentum

1.E-04

1.E-03

1.E-02

1.E-01

1.E+00

1 10 100 1000

Impact parameter bin #

Cu

mu

lati

ve

pro

ba

bil

ity

www.QinetiQ.com


Need to Generation Glauber Profile Data Has Added Significant Complexity

• DPMJET-II.5 expects the data to be a function of:

− A & Z of projectile

− A & Z of target

− Projectile momentum/nucleon

• If we created data for 2546 projectiles (~301 stable nuclides + ~2245 well-quantified radioactive nuclides) on 301 stable targets = 676,046 combinations ~100 GBytes is full-size ASCII files

• This is the reason this wasn’t pre-computed.

www.QinetiQ.com


DPMJET-II.5 interface (Old Version)theExcitationHandler->SetEvaporation(theEvaporation);

theExcitationHandler->SetFermiModel(theFermiBreakUp);

theExcitationHandler->SetMaxAandZForFermiBreakUp(12, 6);

G4DPMJET2_5Model *theDPM = new G4DPMJET2_5Model(theExcitationHandler);

theDPM->SetMinEnergy(9.0*GeV);

theDPM->SetVerboseLevel(2);

• Get or define Glauber Data for projectiles and target

G4GlauberDataSetHandler *theGDSHandler = theDPM->GetGlauberDataSetHandler();

theGDSHandler->AddGlauberDataDir(“/home/geant4/g4/geant4/data/GlauberDat”);

theGDSHandler->AddGlauberDataDir(“.”);

theGDSHandler->LoadGlauberData(12,6,false);

theGDSHandler->LoadGlauberData(56,26,27,13,false);

theGDSHandler->CreateGlauberData(24,12, true);

Carbon on all materials in geometry

Iron on aluminium

magnesium on all materials in geometry

www.QinetiQ.com


DPMJET-II.5 interface (Old Version)G4ParticleDefinition *anIon = G4ParticleTable::GetParticleTable()->GetIonTable()->

GetIon(8,16)

theGDSHandler->CreateGlauberData(anIon);

theGDSHandler->SetSaveGlauberDataType(1);

theGDSHandler->SetSaveGlauberDataDir(“/home/geant4/g4/geant4/data/GlauberDat”);

theGDSHandler->SaveGlauberData(anIon);

• Or for saving lots of data:

theGDSHandler->SaveAllGlauberData();

• Set limits for the amount of Glauber data which can be generated or loaded.

theGDSHandler->SetMaxGlauberDataSets (0,75);

theGDSHandler->SetMaxGlauberDataSets (-1,1000);

oxygen on all materials in geometry

oxygen in any material

Save as parameterised (not full) data

www.QinetiQ.com


DPMJET-II.5 Subroutine SHMAKI

Can new GDS be created (for 1 momentum)

generate new data for one momentum only, transfer to DPMJET

N N

Instantiate new GDS objects generate new data (all momenta)

Y

N

Transfer projectile to product vector

DPMJET-II.5 Subroutine DPMEVT

Simulate interaction Transfer information about secondary products to product vector, de-excite nuclear fragments if required and store nuclear de-excitation products

Instantiate new GDS objects & load with data

Instantiate G4DPMJET2_5Model Initialise C++ variables and DPMJET-II.5 common block data Instantiation G4DPMJET2_5Model

Instantiate Glauber DSH object Initialise Glauber DSH variables & clear map of ptr to GDS objects

Define Glauber Data Sets Retrieve ptr to GDSH

Load existing GDS based on projectile – target

Create any new GDS based on projectile – target

Set limits on max # GDS data to be loaded/generated

Instantiate GDS objects & generate new data

Save GDS data to file if required

Unload GDS based on projectile – target

Delete GDS objects

ApplyYourself

DPMJET-II.5 Common blocks accessed through C Struct declarations

Determine projectile and target

Initialise DPMJET-II.5 common block data for interaction

DPMJET-II.5 Common blocks

Get GDS and transfer Glauber data to DPMJET-II.5

GDS already loaded for interaction?


Does filed GDS exist and can it be loaded?

Can new GDS be created (for all momenta)?


Y Y Y

N

LOGICAL FLOW DIAGRAM – Old Version (part 1 of 2)

www.QinetiQ.com



Can new GDS be created (for 1 momentum)

generate new data for one momentum only, transfer to DPMJET

N N

Instantiate new GDS objects generate new data (all momenta)

Y

N

Transfer projectile to product vector

DPMJET-II.5 Subroutine DPMEVT

Simulate interaction Transfer information about secondary products to product vector, de-excite nuclear fragments if required and store nuclear de-excitation products


Instantiate G4DPMJET2_5Model Initialise C++ variables and DPMJET-II.5 common block data Instantiation G4DPMJET2_5Model

Instantiate Glauber DSH object Initialise Glauber DSH variables & clear map of ptr to GDS objects

Define Glauber Data Sets Retrieve ptr to GDSH

Load existing GDS based on projectile – target

Create any new GDS based on projectile – target

Set limits on max # GDS data to be loaded/generated

Instantiate GDS objects & generate new data

Save GDS data to file if required

Unload GDS based on projectile – target

Delete GDS objects

ApplyYourself

DPMJET-II.5 Common blocks accessed through C Struct declarations

Determine projectile and target

Initialise DPMJET-II.5 common block data for interaction

DPMJET-II.5 Common blocks

Get GDS and transfer Glauber data to DPMJET-II.5

GDS already loaded for interaction?


Does filed GDS exist and can it be loaded?

Can new GDS be created (for all momenta)?


Y Y Y

N

LOGICAL FLOW DIAGRAM – Old Version (part 2 of 2)

www.QinetiQ.com


Review of Existing Sources of Error and Assessment of Approximations to Reduce Demand on No. Data Sets

• There are three sources of error already within the Glauber profile data as used/generated by DPMJET-II.5

− Stochastic errors: generated by MC calculation (up to ~3%)

− MODB interpolation errors i.e. interpolation between different probabilities->impact parameters shown in page 2 (80% of data within error of ~10% for light proj/targets, rising to 90% for others)

− Errors in interpolation between different projectile momenta (90% with 3%)

• There is a better way:

− Only generate and store Glauber profile data as a function of Aprojectile and Atarget (typically <3% error)

− Instead of MODB interpolation, fit impact parameter as a function of integral probability: 200 datapoints -> 10 parameters

− ~300 Mbytes ASCII (uncompressed)

9999.095.0)1ln(

95.003.0

1

03.0

)(

3

0

2

1

2

2

1

PPc

P

P

Pb

PPb

PX

i

ii

M

M

M

www.QinetiQ.com


DPMJET-II.5 interface (New Version)

theExcitationHandler->SetEvaporation(theEvaporation);

theExcitationHandler->SetFermiModel(theFermiBreakUp);

theExcitationHandler->SetMaxAandZForFermiBreakUp(12, 6);

G4DPMJET2_5Model *theDPM = new G4DPMJET2_5Model(theExcitationHandler);

theDPM->SetMinEnergy(9.0*GeV);

theDPM->SetVerboseLevel(2);

G4GlaubAADataSetHandler *theGDSHandler =theDPM->GetGlauberDataSetHandler();

theGDSHandler->SetMaxGlauberDataSets (0,75);

www.QinetiQ.com


LOGICAL FLOW DIAGRAM – New Version (part 1 of 2)

www.QinetiQ.com


LOGICAL FLOW DIAGRAM – New Version (part 2 of 2)

www.QinetiQ.com


Data Generation

• Just completed re-writing of code; need to test/debug

• Currently generating Glauber profiles for 2AP 58 and 2 AT 58

• Also, as part of this process, generating cross-sections for all inelastic interactions and intend to produce updated cross-section class for E>2GeV/nuc

− Necessary since currently using Tripathi, Shen or Kox approximations, and Tripathi only seems to be tested to ~2GeV/nuc

• Taking ~4 CPU-months

• Need to extent to 2 AP 240 and 2 AT 240

• Need for proton-nuclear interactions acknowledged

www.QinetiQ.com


WP1 DPMJET-II.5 interface design constraints

• Will NOT be able to instantiate G4DPMJET2_5Model more than once, due to reliance on FORTRAN subroutines

• In order to run G4DPMJET2_5Model, you require cernlib, kernlib, mathlib,

− Currently uses g77 + g2c

− Looking into gfortran

www.QinetiQ.com© Copyright QinetiQ limited 2007

QinetiQ Proprietary - MarsREM Project-in-Confidence

Back-up slides

www.QinetiQ.com


Error in Glauber Profiles from Monte Carlo statistics (1 of 2)

0%

20%

40%

60%

80%

100%

0.0% 1.0% 2.0% 3.0%

Error, E [%]

% d

ata

wit

h e

rro

r <

= E

[%

]

3H+3H3He+3H3He+3He6He+6He6Li+6Li6Be+6Be

www.QinetiQ.com


Error in Glauber Profiles from Monte Carlo statistics (2 of 2)

0%

20%

40%

60%

80%

100%

0.0% 0.5% 1.0% 1.5%

Error, E [%]

% d

ata

wit

h e

rro

r <

= E

[%

]

12C+12C12C+24Mg12C+56Fe24Mg+24Mg24Mg+56Fe56Fe+56Fe

www.QinetiQ.com


Error in determination of impact parameter due to existing MODB interpolation

40%

60%

80%

100%

0.001% 0.010% 0.100% 1.000% 10.000% 100.000%

Error, E [%]

% D

ata

wit

h e

rro

r >

= E

[%

]

MODB 3H+3H

MODB 6Li+6Li

MODB 12C+12C

MODB 24Mg+24Mg

MODB 56Fe+56Fe

www.QinetiQ.com


Errors from interpolating in projectile momentum

0%

20%

40%

60%

80%

100%

0.0% 1.0% 2.0% 3.0%

Error, E [%]

% d

ata

wit

h e

rro

r <

=E

[%

]

3H+3H

6He+6He

12B+12Be

24Na+24Si

www.QinetiQ.com


0%

20%

40%

60%

80%

100%

0.0% 1.0% 2.0% 3.0%

Error, E [%]

% d

ata

wit

h e

rro

r <

= E

[%

]

A=3 on A=3

A=3 on A=6

A=3 on A=12

A=3 on A=24

A=3 on A=56

Each Curve Represents the Variations in Glauber profiles for fixed Aprojectile and Atarget but range of Zprojectile and Ztarget

www.QinetiQ.com


0%

20%

40%

60%

80%

100%

0.0% 1.0% 2.0% 3.0%

Error, E [%]

% d

ata

wit

h e

rro

r <

=E

[%

]

A=6 on A=6

A=6 on A=12

A=6 on A=24

A=6 on A=56


www.QinetiQ.com


0%

20%

40%

60%

80%

100%

0.0% 1.0% 2.0% 3.0%

Error, E [%]

% d

ata

wit

h e

rro

r <

=E

[%

]

A=12 on A=12

A=12 on A=24

A=12 on A=56


www.QinetiQ.com


0%

20%

40%

60%

80%

100%

0.0% 1.0% 2.0% 3.0%

Error, E [%]

% d

ata

wit

h e

rro

r <

=E

[%

]

A=24 on A=24

A=24 on A=56

A=56 on A=56


www.QinetiQ.com


WP1 Parameterisation of Glauber data - inverse function

74.071119

1.901494

El

x Dl n n

11.514841 10

3 Dl n

1 103

0.01 0.1 1

1

10

100

Cumulative probability, f

i, im

pact

par

amet

er

74.071119

60

El

x Dl n n

10.99 Dl n

0.99 0.992 0.994 0.996 0.998 160

62

64

66

68

70

72

74

76

Cumulative probability , f

i, im

pact

par

amet

er

www.QinetiQ.com


Errors due to use of parametric fit of impact parameter to integral probability

40%

60%

80%

100%

0.000% 2.000% 4.000% 6.000% 8.000% 10.000%

Error, E [%]

% D

ata

wit

h e

rro

r >

= E

[%

]

Interp F 3H+3H

Interp F 6Li+6Li

Interp F 12C+12C

Interp F 24Mg+24Mg

Interp F 56Fe+56Fe

www.QinetiQ.com


1.E-04

1.E-03

1.E-02

1.E-01

0 2 4 6 8 10 12 14

Energy [GeV]

Cro

ss

-se

cti

on

[m

b]

Original DPMJET2.5

Geant4 DPMJET without de-ex.

1.E-04

1.E-03

1.E-02

1.E-01

0 2 4 6 8 10 12 14

Energy [GeV]

Cro

ss

-se

cti

on

[m

b]

DPMJET2.5 FLUKA

Geant4 DPMJET with de-ex.

1.E-04

1.E-03

1.E-02

1.E-01

0 2 4 6 8 10 12 14

Energy [GeV]

Cro

ss-s

ecti

on

[m

b]

Original DPMJET2.5

Geant4 DPMJET2 without de-ex.

12C on C @10 GeV/n

1.E-04

1.E-03

1.E-02

1.E-01

0 2 4 6 8 10 12 14

Energy [GeV]

Cro

ss-s

ecti

on

[m

b]

DPMJET2.5 FLUKA


protons protons

neutrons neutrons

Without de-excitation With de-excitation

www.QinetiQ.com


1.E-06

1.E-05

1.E-04

1.E-03

0 50 100 150 200 250

Energy [GeV]

Cro

ss

-se

cti

on

[m

b]

Original DPMJET2.5


1.E-06

1.E-05

1.E-04

1.E-03

0 50 100 150 200 250

Energy [GeV]

Cro

ss

-se

cti

on

[m

b]

DPMJET2.5 FLUKA


1.E-06

1.E-05

1.E-04

1.E-03

0 50 100 150 200 250

Energy [GeV]

Cro

ss-s

ecti

on

[m

b]

Original DPMJET-2.5

Geant4 DPMJET wo de-ex.

1.E-06

1.E-05

1.E-04

1.E-03

0 50 100 150 200 250

Energy [GeV]

Cro

ss-s

ecti

on

[m

b]

DPMJET2.5 FLUKA


12C on C @200 GeV/n

protons protons

neutrons neutrons


www.QinetiQ.com


1.E-04

1.E-03

1.E-02

1.E-01

0 2 4 6 8 10 12 14

Energy [GeV]

Cro

ss-s

ecti

on

[m

b]

Original DPMJET2.5

Geant4 DPMJET without de-ex

1.E-04

1.E-03

1.E-02

1.E-01

0 2 4 6 8 10 12 14

Energy [GeV]

Cro

ss

-se

cti

on

[m

b]

DPMJET2.5 FLUKA


1.E-04

1.E-03

1.E-02

1.E-01

0 2 4 6 8 10 12 14

Energy [GeV]

Cro

ss-s

ecti

on

[m

b]

Original DPMJET2.5


1.E-04

1.E-03

1.E-02

1.E-01

0 2 4 6 8 10 12 14

Energy [GeV]

Cro

ss-s

ecti

on

[m

b]

DPMJET2.5 FLUKA


12C on Al @10 GeV/n

protons protons

neutrons neutrons


www.QinetiQ.com


1.E-06

1.E-05

1.E-04

1.E-03

0 50 100 150 200 250

Energy [GeV]

Cro

ss

-sec

tio

n [

mb

]

Original DPMJET2.5


1.E-06

1.E-05

1.E-04

1.E-03

0 50 100 150 200 250

Energy [GeV]

Cro

ss

-sec

tio

n [

mb

]

DPMJET2.5 FLUKA


1.E-06

1.E-05

1.E-04

1.E-03

0 50 100 150 200 250

Energy [GeV]

Cro

ss-s

ecti

on

[m

b]

Original DPMJET2.5


1.E-06

1.E-05

1.E-04

1.E-03

0 50 100 150 200 250

Energy [GeV]

Cro

ss-s

ecti

on

[m

b]

DPMJET2.5 FLUKA


12C on Al @200 GeV/n

protons protons

neutrons neutrons


Www.QinetiQ.com © Copyright QinetiQ limited 2007 1 DPMJET-2.5 & DPMJET-3 Treats hadron-nuclear and nuclear-nuclear interactions >5 GeV/nuc, with the upper.

Documents

data slide

copyright qinetiq

glauber data

version of dpmjet

data generation

geometry slide

data sets

limited documentation