MadGraph5_aMC@NLO

about iseed on run_card.dat

Asked by ozgun

Dear Madgraph Team,

I created a process with 600k events using Madgraph. I generate 1200 different events, 500 for each event. For example, I run a script from run_t8_001_e1 to run_t8_001_e60, which generates 3000 events with 60 different root files, and then use the hadd command to combine them. Then, I run another script where I define different directories, which generates events from run_t8_001_e61 to run_t8_001_e120. By continuing in this way, I created 600k events. Although the run names and directories were different in each script, I noticed an anomaly when checking the number of events and cross-section values ​​for each execution.

For example, for the first 500 event generation process (run_t8_001_e1), the cross-section value is 7.71 and the number of events is 350. When I run the second script, and for the first 500 event generation process of that script (run_t8_001_e61), the cross-section value is 7.71 and the number of events is 350. Similarly, when I check run_t8_001_e900 (the first 500 run in the script), its cross-section value is 7.71 and the number of events is 350.

In all the scripts I use in this way, the cross-section value and the number of events for the first 500 events are the same, for the 30th are the same, and for the 45th are the same. For example, in every event production from 1 to 60, the cross-section value of the 9th production (run_t8_001_e9, run_t8_001_e18, run_t8_001_e27, ...) is 7.753 ± 0.12, and the number of event value is 472. The banner.txt file is created for every 500 production processes. When I check, I see different values ​​such as 21 = iseed, 367 = iseed.

Is this normal? If it is abnormal, how can I solve it?

Question information

Language:
English Edit question
Status:
Answered
For:
MadGraph5_aMC@NLO Edit question
Assignee:
No assignee Edit question
Last query:
Last reply:
Revision history for this message
Olivier Mattelaer (olivier-mattelaer) said : 		#1

Not sure to fully understand what you are doing here.

If you generate different directory you can not use iseed=0 in the run_card and be sure that the value of iseed does not overlap between the directories/run.
If you use multiple times the launch command (or the multi_run command) then the issed value is automatically update between run but the sequence will be the same for two different directories.

Does it answer your question?

Cheers,

Olivier

Revision history for this message
ozgun (ozgunkdz) said : 		#2

Here is a sample of my generation script

output /home/desktop/data/ft9/t9_02
launch -n run_t9_0.2_e1
shower=Pythia8
detector=Delphes
analysis=OFF
set nevents 500
set lpp1 1
set lpp2 1
set ebeam1 50000
set ebeam2 50000
set FT8 0.0
set FT9 0.2e-13
set pta 25
set ptj 0
set ptj1min 30
set ptj2min 30
set mmjj 800
set deltaeta 2.5
set maxjetflavor 4
set etaamin -2.5
set etajmin -4.7
set etaj 4.7
#set xetamin 4.7
#set ptb 25
#set zerowidth_tchannel True
set sde_strategy 2
/home/programFiles/MG5_aMC_v3_4_2/Delphes/cards/FCC/FCChh_aqgc.tcl
done
launch -n run_t9_0.2_e2
shower=Pythia8
detector=Delphes
analysis=OFF
set nevents 500
set lpp1 1
set lpp2 1
set ebeam1 50000
set ebeam2 50000
set FT8 0.0
set FT9 0.2e-13
set pta 25
set ptj 0
set ptj1min 30
set ptj2min 30
set mmjj 800
set deltaeta 2.5
set maxjetflavor 4
set etaamin -2.5
set etajmin -4.7
set etaj 4.7
#set xetamin 4.7
#set ptb 25
#set zerowidth_tchannel True
set sde_strategy 2
/home/programFiles/MG5_aMC_v3_4_2/Delphes/cards/FCC/FCChh_aqgc.tcl
.
.
.
.
launch -n run_t9_0.2_e60
shower=Pythia8
detector=Delphes
analysis=OFF
set nevents 500
set lpp1 1
set lpp2 1
set ebeam1 50000
set ebeam2 50000
set FT8 0.0
set FT9 0.2e-13
set pta 25
set ptj 0
set ptj1min 30
set ptj2min 30
set mmjj 800
set deltaeta 2.5
set maxjetflavor 4
set etaamin -2.5
set etajmin -4.7
set etaj 4.7
#set xetamin 4.7
#set ptb 25
#set zerowidth_tchannel True
set sde_strategy 2
/home//programFiles/MG5_aMC_v3_4_2/Delphes/cards/FCC/FCChh_aqgc.tcl
done
launch /home/desktop/data/ft9/t9_02 -i
print_results --path=/home/desktop/data/ft9/t9_02/t9_0.2_e1_e60.txt --format=short

When it finishes I run another script.

output /home/desktop/data/ft9/t9_02
launch -n run_t9_0.2_e61
shower=Pythia8
detector=Delphes
analysis=OFF
set nevents 500
set lpp1 1
set lpp2 1
set ebeam1 50000
set ebeam2 50000
set FT8 0.0
set FT9 0.2e-13
set pta 25
set ptj 0
set ptj1min 30
set ptj2min 30
set mmjj 800
set deltaeta 2.5
set maxjetflavor 4
set etaamin -2.5
set etajmin -4.7
set etaj 4.7
#set xetamin 4.7
#set ptb 25
#set zerowidth_tchannel True
set sde_strategy 2
/home/programFiles/MG5_aMC_v3_4_2/Delphes/cards/FCC/FCChh_aqgc.tcl
done
launch -n run_t9_0.2_e62
shower=Pythia8
detector=Delphes
analysis=OFF
set nevents 500
set lpp1 1
set lpp2 1
set ebeam1 50000
set ebeam2 50000
set FT8 0.0
set FT9 0.2e-13
set pta 25
set ptj 0
set ptj1min 30
set ptj2min 30
set mmjj 800
set deltaeta 2.5
set maxjetflavor 4
set etaamin -2.5
set etajmin -4.7
set etaj 4.7
#set xetamin 4.7
#set ptb 25
#set zerowidth_tchannel True
set sde_strategy 2
/home/programFiles/MG5_aMC_v3_4_2/Delphes/cards/FCC/FCChh_aqgc.tcl
.
.
.
.
launch -n run_t9_0.2_e120
shower=Pythia8
detector=Delphes
analysis=OFF
set nevents 500
set lpp1 1
set lpp2 1
set ebeam1 50000
set ebeam2 50000
set FT8 0.0
set FT9 0.2e-13
set pta 25
set ptj 0
set ptj1min 30
set ptj2min 30
set mmjj 800
set deltaeta 2.5
set maxjetflavor 4
set etaamin -2.5
set etajmin -4.7
set etaj 4.7
#set xetamin 4.7
#set ptb 25
#set zerowidth_tchannel True
set sde_strategy 2
/home//programFiles/MG5_aMC_v3_4_2/Delphes/cards/FCC/FCChh_aqgc.tcl
done
launch /home/desktop/data/ft9/t9_02 -i
print_results --path=/home/desktop/data/ft9/t9_02/t9_0.2_e61_e120.txt --format=short

When I check the cross-section values of run_t9_0.2_e1 and run_t9_0.2_e61 they same.

Revision history for this message
Olivier Mattelaer (olivier-mattelaer) said : 		#3

So yes then this is what i said above,
The iseed value need to be controlled by hand otherwise the same value of iseed will be taken for each of the
directory meaning you are generating the exact same random sequence in all cases.

Cheers,

Olivier

Can you help with this problem?

Provide an answer of your own, or ask ozgun for more information if necessary.

To post a message you must log in.