madgraph fails to reach target number of event (outputs less events than asked)

[Sungwoo Hong]

Hi MadGraph developers,

I have been trying to generate events for the following process:

set group_subprocesses Auto
set ignore_six_quark_processes False
set loop_color_flows False
set gauge unitary
set complex_mass_scheme False
set max_npoint_for_channel 0
import model sm
define p = g u c d s u~ c~ d~ s~
define j = g u c d s u~ c~ d~ s~
define l+ = e+ mu+
define l- = e- mu-
define vl = ve vm vt
define vl~ = ve~ vm~ vt~
generate p p > j j Z W+, (Z > l- l+), (W+ > l+ vl) @1
add process p p > j j Z W-, (W- > l- vl~), (Z > l- l+) @2
output

This process has many diagrams (14240 diagrams (2048 independent)) to evaluate.
At the beginning, I hope to generate events for

generate p p > j j l- l+ W+, W+ > l+ vl @1
add process p p > j j l- l+ W-, W- > l- vl~ @2

but this process has even more diagrams (31040 diagrams (4464 independent)) and it gives me similar but worse error that I am going to describe now.

The error is simply that I am getting much less events than I asked. For example, if I asked 10k, sometimes it gives ~O(3000) and sometimes ~O(4000) something like that. The log says that

"INFO: fail to reach target 10000 "

Another, perhaps related error is that when I run this in the cluster in "multi-run" mode,

(1) it generates unweighted_event.lhe file but NO event.lhe (weighted) one.
(2) if I run, say, 10 multi_run mode, in some run it generates all files (unweighted_event, lhco for delphes, hep.gz, but still no weighted event (parton) file) but with much less events than asked, however in some run, it generates only pythia.log (and this log file often times huge ~ few TB), banner.txt, unweighted_event.lhe.

I thought at the beginning it is error that occurs during hadronization (since I got full unweighted_event, but less events in lhco file) but now I think (after some experiments running only in parton level) it is actually errors occurring before (or perhaps both) that.

Any help will be greatly appreciated !
Thank you.
Sungwoo Hong

[Sungwoo Hong (shong710)]

Hi Olivier,

It seems that you replied to my question, but I do not understand any of them.
Am I supposed to download and do something in order to be able to see your comments ?

Many thanks !
Sungwoo

[Olivier Mattelaer (olivier-mattelaer)]

launchpad was suppose to uncrypt the message automatically. But looks like it failed to do so this time.

My message was:

Hi,

Did you use 2.5.0.beta?

Cheers,

Olivier

(or more generally which version are you using?)

[Sungwoo Hong (shong710)]

Hi,

Thank you for getting back to me soon.

I am using MG5_aMC_v2_4_3 (in cluster).
I wonder if this error occurs when the number of diagram is very large.
Right now, I am running another process which also has big number of diagrams ~O(50k) and I think I am getting the same type of issues.

Sungwoo

[Olivier Mattelaer (olivier-mattelaer)]

Hi,

Can you try with the beta version?
https://launchpad.net/mg5amcnlo/2.0/2.4.x/+download/MG5_aMC_v2_5_0_beta2.tar.gz

I guess that the problem is already solved in that version.

Cheers,

Olivier

> On Oct 11, 2016, at 19:53, Sungwoo Hong <email address hidden> wrote:
>
> Hi,
>
> Thank you for getting back to me soon.
>
> I am using MG5_aMC_v2_4_3 (in cluster).
> I wonder if this error occurs when the number of diagram is very large.
> Right now, I am running another process which also has big number of diagrams ~O(50k) and I think I am getting the same type of issues.
>
> Sungwoo
>
> --
> You received this bug notification because you are subscribed to
> MadGraph5_aMC@NLO.
> https://bugs.launchpad.net/bugs/1631465
>
> Title:
> madgraph fails to reach target number of event (outputs less events
> than asked)
>
> Status in MadGraph5_aMC@NLO:
> New
>
> Bug description:
> Hi MadGraph developers,
>
> I have been trying to generate events for the following process:
>
> set group_subprocesses Auto
> set ignore_six_quark_processes False
> set loop_color_flows False
> set gauge unitary
> set complex_mass_scheme False
> set max_npoint_for_channel 0
> import model sm
> define p = g u c d s u~ c~ d~ s~
> define j = g u c d s u~ c~ d~ s~
> define l+ = e+ mu+
> define l- = e- mu-
> define vl = ve vm vt
> define vl~ = ve~ vm~ vt~
> generate p p > j j Z W+, (Z > l- l+), (W+ > l+ vl) @1
> add process p p > j j Z W-, (W- > l- vl~), (Z > l- l+) @2
> output
>
> This process has many diagrams (14240 diagrams (2048 independent)) to evaluate.
> At the beginning, I hope to generate events for
>
> generate p p > j j l- l+ W+, W+ > l+ vl @1
> add process p p > j j l- l+ W-, W- > l- vl~ @2
>
> but this process has even more diagrams (31040 diagrams (4464
> independent)) and it gives me similar but worse error that I am going
> to describe now.
>
> The error is simply that I am getting much less events than I asked.
> For example, if I asked 10k, sometimes it gives ~O(3000) and sometimes
> ~O(4000) something like that. The log says that
>
> "INFO: fail to reach target 10000 "
>
> Another, perhaps related error is that when I run this in the cluster
> in "multi-run" mode,
>
> (1) it generates unweighted_event.lhe file but NO event.lhe (weighted) one.
> (2) if I run, say, 10 multi_run mode, in some run it generates all files (unweighted_event, lhco for delphes, hep.gz, but still no weighted event (parton) file) but with much less events than asked, however in some run, it generates only pythia.log (and this log file often times huge ~ few TB), banner.txt, unweighted_event.lhe.
>
> I thought at the beginning it is error that occurs during
> hadronization (since I got full unweighted_event, but less events in
> lhco file) but now I think (after some experiments running only in
> parton level) it is actually errors occurring before (or perhaps both)
> that.
>
> Any help will be greatly appreciated !
> Thank you.
> Sungwoo Hong
>
> To manage notifications about this bug go to:
> https://bugs.launchpad.net/mg5amcnlo/+bug/1631465/+subscriptions

[Sungwoo Hong (shong710)]

Thank you for your answer Olivier.

So, I tried v2_5_0_beta2 version you suggested and it seems to give me full parton events.
However, now I have a problem with pythia running. I noticed that in this new version, it is recommending to use pythia 8. Still I installed old pythia-pgs since I've been using it together with Delphes and when I run the MG together with pythia and Delphes , it does not run pythia successfully. It gives me the error message:

INFO: can not run systematics since can not link python to lhapdf
store_events
INFO: Storing parton level results
INFO: End Parton
reweight -from_cards
decay_events -from_cards
^[[1;31mpythia-pgs require event_norm to be on sum. Do not run pythia6^[[0m
INFO: storing files of previous run
INFO: Done
quit
INFO:

Do you have any idea about this ?
Also, do you think even order version (older than 2_4_3) will not have the error I had earlier (giving much less events than asked) ? If so, I may tried one.

Thank you !
Sungwoo

[Olivier Mattelaer (olivier-mattelaer)]

Hi,

you should edit the run_card in order to have the parameter
> event_norm = sum

in order to be able to run successfully pythia6

> Also, do you think even order version (older than 2_4_3) will not have the error I had earlier (giving much less events than asked) ? If so, I may tried one.

Sure 2.4.2 and earlier should not have that problem if I’m correct.

Cheers,

Olivier

> On Oct 13, 2016, at 23:55, Sungwoo Hong <email address hidden> wrote:
>
> Thank you for your answer Olivier.
>
> So, I tried v2_5_0_beta2 version you suggested and it seems to give me full parton events.
> However, now I have a problem with pythia running. I noticed that in this new version, it is recommending to use pythia 8. Still I installed old pythia-pgs since I've been using it together with Delphes and when I run the MG together with pythia and Delphes , it does not run pythia successfully. It gives me the error message:
>
> INFO: can not run systematics since can not link python to lhapdf
> store_events
> INFO: Storing parton level results
> INFO: End Parton
> reweight -from_cards
> decay_events -from_cards
> ^[[1;31mpythia-pgs require event_norm to be on sum. Do not run pythia6^[[0m
> INFO: storing files of previous run
> INFO: Done
> quit
> INFO:
>
> Do you have any idea about this ?
> Also, do you think even order version (older than 2_4_3) will not have the error I had earlier (giving much less events than asked) ? If so, I may tried one.
>
> Thank you !
> Sungwoo
>
> --
> You received this bug notification because you are subscribed to
> MadGraph5_aMC@NLO.
> https://bugs.launchpad.net/bugs/1631465
>
> Title:
> madgraph fails to reach target number of event (outputs less events
> than asked)
>
> Status in MadGraph5_aMC@NLO:
> New
>
> Bug description:
> Hi MadGraph developers,
>
> I have been trying to generate events for the following process:
>
> set group_subprocesses Auto
> set ignore_six_quark_processes False
> set loop_color_flows False
> set gauge unitary
> set complex_mass_scheme False
> set max_npoint_for_channel 0
> import model sm
> define p = g u c d s u~ c~ d~ s~
> define j = g u c d s u~ c~ d~ s~
> define l+ = e+ mu+
> define l- = e- mu-
> define vl = ve vm vt
> define vl~ = ve~ vm~ vt~
> generate p p > j j Z W+, (Z > l- l+), (W+ > l+ vl) @1
> add process p p > j j Z W-, (W- > l- vl~), (Z > l- l+) @2
> output
>
> This process has many diagrams (14240 diagrams (2048 independent)) to evaluate.
> At the beginning, I hope to generate events for
>
> generate p p > j j l- l+ W+, W+ > l+ vl @1
> add process p p > j j l- l+ W-, W- > l- vl~ @2
>
> but this process has even more diagrams (31040 diagrams (4464
> independent)) and it gives me similar but worse error that I am going
> to describe now.
>
> The error is simply that I am getting much less events than I asked.
> For example, if I asked 10k, sometimes it gives ~O(3000) and sometimes
> ~O(4000) something like that. The log says that
>
> "INFO: fail to reach target 10000 "
>
> Another, perhaps related error is that when I run this in the cluster...

[Sungwoo Hong (shong710)]

Thank you Olivier for quick answer !

I will try either options and will let you know how it goes !

Thank you again !
best,
Sungwoo

[Graham (grahamtriumf)]

Hi!

I have the same bug so perhaps its most useful to comment in the same place.

I'm trying to generate events for the Z prime model with Zp > tau+ tau- > neutrinos and electrons

I want 1000 events but the number of events stops at 33 and 8 so far in my two attempts to run.

I downloaded and am using the v3.0.1 beta version

best, Graham