Error in STDHEP file generation after pythia showering in mssm process

Hi Juhi,

> I am generating a mssm process p p > go go in MG5_aMC_v2_2_2 .I am generating 1million events by using multi_run process (each step being of 50k events).Initially I am generating the unweighted events .After the merged file is generated I am running pythia+Delphes on it. The problem is that the STDHEP tar file being generated at the end of this run is of the form : "tag_1_pythia_events.hep" instead of "tag_1_pythia_events.hep.gz" .The file is inextractable. I want to use the hep file for a standalone Delphes comparison hence I need to extract the .hep file . I tried giving a fresh run however the same error occurs again. I am confused as for shorter runs of 100 the proper “tag_1_pythia_events.hep.gz" file is being generated and is extractable but not so for the large number of events .

This is a know problem for file larger than 4 Gb. This problem will be fixed in 2.2.3 which should be release extremely soon.
The problem is just that the file name is wrong, if you rename your lhe file into lhe.gz everything should be fine (and then you can run Delphes on it).

> Another thing is that I am using a susyhit generated spectrum as the param card in mg5. I read up here in the answers section about a short python script to change the format of the susyhit output to make it compatible with mg5 param card.However after such a conversion , pythia is not reading any of the branching ratios og gluino.However directly using the susyhit output removes this error. I proceeded with this approach to generate the events followed by the pythia + delphes run.
> Am I correct in presuming that MG5_aMC_v2_2_2 automatically takes care of the compatibiltiy of the susy hit output to its param card format?Also is there an error in the pythia run due to the above case?

I’m not sure which thread you are referring to. If you speak about SLHA 1/2 format, we are suppose to handle both at the madgraph level but pythia6 only support SLHA1.
Now I do not know which kind of format SUSYHIT returns and how much this is compatible with our code. I know that a lot of SUSY spectrum generator forgets to specify the SM information
which typically makes our code to crash.

Cheers,

Olivier

On 07 Feb 2015, at 18:11, Juhi Dutta <email address hidden> wrote:

> New question #261817 on MadGraph5_aMC@NLO:
> https://answers.launchpad.net/mg5amcnlo/+question/261817
>
> Hi,
>
> I am generating a mssm process p p > go go in MG5_aMC_v2_2_2 .I am generating 1million events by using multi_run process (each step being of 50k events).Initially I am generating the unweighted events .After the merged file is generated I am running pythia+Delphes on it. The problem is that the STDHEP tar file being generated at the end of this run is of the form : "tag_1_pythia_events.hep" instead of "tag_1_pythia_events.hep.gz" .The file is inextractable. I want to use the hep file for a standalone Delphes comparison hence I need to extract the .hep file . I tried giving a fresh run however the same error occurs again. I am confused as for shorter runs of 100 the proper "tag_1_pythia_events.hep.gz" file is being generated and is extractable but not so for the large number of events .
>
> Another thing is that I am using a susyhit generated spectrum as the param card in mg5. I read up here in the answers section about a short python script to change the format of the susyhit output to make it compatible with mg5 param card.However after such a conversion , pythia is not reading any of the branching ratios og gluino.However directly using the susyhit output removes this error. I proceeded with this approach to generate the events followed by the pythia + delphes run.
> Am I correct in presuming that MG5_aMC_v2_2_2 automatically takes care of the compatibiltiy of the susy hit output to its param card format?Also is there an error in the pythia run due to the above case?
>
> Am I making a mistake somewhere? Kindly help me resolve this issue.Looking forward to your help.
>
> Thanks & Regards,
>
> Juhi
>
>
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

Hi,

Thanks for the reply! I tried renaming the hep tar file to hep.gz. This opens up correctly with archive manager.However the hep file inside shows a size of 3.1GB whereas the full tar file is 46.3GB. Also extraction hangs up midway. So then does this hep file has incomplete information of the process ? Subsequently the lhco file which was generated (as I had run Pythia+Delphes together) is not trustable or is it? How do I check whether it is correct ? Also can the delphes option be independently run on madgraph?I have so far found it to be clubbed with the pythia running option when I try to run via bin/madevent.

And for the susyhit case , I was referring to the following ticket id :226751.However that was in reference to pythia 8 showering in madgraph while I was using Pythia 6 all the while. Thanks for pointing out !

Regards,

Juhi

Hi Juhi,

> However the hep file inside shows a size of 3.1GB whereas the full tar file is 46.3GB. Also extraction hangs up midway.

Clearly your extraction fail. Probably a problem linked that you archive manager is compile in 32bit and therefore can not handle such large file. Actually one of the main motivation of the multi-run option was to not run pythia on the merge file to avoid to generate file larger than 4 Gb. So I would advise to run pythia on each of your multi-run sample independently.

> Also can the delphes option be independently run on madgraph?

Sure it can, exactly like pythia can be run from ./bin/madevent, you have a command delphes

Cheers,

Olivier

Thanks Olivier Mattelaer, that solved my question.

Hi,

Apologies for re-opening the question again. I have a doubt.

> I would advise to run pythia on each of your multi-run sample independently.

I did so .Each step of my multirun being 10k. The tag_1_pythia_events.hep (461.3mb size ) is being formed.I rename it and now the file is extractable. The "pythia.hep" file in it is of size 1.1GB. I am unsure whether this is the correct file.Is there a way to be sure about whether it is the correct file ? 10k is a short run and the same error showing up is what concerns me here.Please help.

Regards,

Juhi

Thanks Olivier Mattelaer, that solved my question.