Form factors with transposed spinors bug

Bug #1468429 reported by Khaled Teilab
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
MadGraph5_aMC@NLO
Fix Released
Undecided
Unassigned

Bug Description

Hi,
I try to include a form factor in a model having two scalar particles (sig, phi) and two fermions (f1 and f2).
The lagrangian used in feynrules looks like:
g_1 Inner[Dot,Inner[Dot,{f1bar,f2bar },sig IdentityMatrix[2]],{f1,f2 }] + I g_2 (phi(CC[f1bar].Ga[5].f1) + anti[phi](f1bar.Ga[5].CC[f1]))

In the attachment you find four models:
1) Model1: Includes ALL vertices coming from the above Lagrangian WITHOUT form factor => WORKS
2) Model4: Includes ONLY vertices involving the sig particle WITH form factor => WORKS
3) Model2: Includes ALL vertices coming from the above Lagrangian WITH form factor on ALL vertices => FAILS to produce output folder
4) Model3: Includes ALL vertices coming from the above Lagrangian WITH form factor on ONLY vertices involving the sig particle => FAILS to poruce output folder

The output looks like:

INFO: initialize a new directory: model3test
INFO: remove old information in model3test
INFO: Organizing processes into subprocess groups
INFO: Generating Helas calls for process: f1+ f1+ > f1+ f1+ WEIGHTED=2
INFO: Processing color information for process: f1+ f1+ > f1+ f1+
INFO: Creating files in directory P0_f1pf1p_f1pf1p
INFO: Generating Feynman diagrams for Process: f1+ f1+ > f1+ f1+ WEIGHTED=2
INFO: Finding symmetric diagrams for subprocess group f1pf1p_f1pf1p
History written to /home/teilab/mg/model3test/Cards/proc_card_mg5.dat
Generated helas calls for 1 subprocesses (3 diagrams) in 0.011 s
Wrote files for 12 helas calls in 0.154 s
Export UFO model to MG4 format
CRITICAL: aS not define as external parameter adding it!
WARNING: aEWM1 and Gf not define in MODEL. AQED will not be written correcty in LHE FILE
ALOHA: aloha creates FFS2 routines
Command "output model3test" interrupted with error:
NameError : name 'exp' is not defined
Please report this bug on https://bugs.launchpad.net/madgraph5
More information is found in 'MG5_debug'.
Please attach this file to your report.

Could you please have a look at the attached models and tell me what the problem is? The debug file is also attached

Thanks,
Khaled

Revision history for this message
Khaled Teilab (teilab) wrote :
Revision history for this message
Olivier Mattelaer (olivier-mattelaer) wrote : Re: [Bug 1468429] [NEW] Form factors with transposed spinors bug
Download full text (5.6 KiB)

Dear Khaled,

The problem is that first you are suppose to use
cmath.exp and not exp
cmath.sqrt and not sqrt
in most of the case this should go trough since those will be handle by fortran anyway but not in this context.

However, you have a four fermion interaction with a fermion flow violation (with identical particle) which is one case that we can not handle.
In order to generate events for your theory you need to introduce fake scalar to split the 4 point interaction in a succession of three point interaction.

Cheers,

Olivier

On 24 Jun 2015, at 17:11, Khaled Teilab <email address hidden> wrote:

> Public bug reported:
>
> Hi,
> I try to include a form factor in a model having two scalar particles (sig, phi) and two fermions (f1 and f2).
> The lagrangian used in feynrules looks like:
> g_1 Inner[Dot,Inner[Dot,{f1bar,f2bar },sig IdentityMatrix[2]],{f1,f2 }] + I g_2 (phi(CC[f1bar].Ga[5].f1) + anti[phi](f1bar.Ga[5].CC[f1]))
>
> In the attachment you find four models:
> 1) Model1: Includes ALL vertices coming from the above Lagrangian WITHOUT form factor => WORKS
> 2) Model4: Includes ONLY vertices involving the sig particle WITH form factor => WORKS
> 3) Model2: Includes ALL vertices coming from the above Lagrangian WITH form factor on ALL vertices => FAILS to produce output folder
> 4) Model3: Includes ALL vertices coming from the above Lagrangian WITH form factor on ONLY vertices involving the sig particle => FAILS to poruce output folder
>
> The output looks like:
>
> INFO: initialize a new directory: model3test
> INFO: remove old information in model3test
> INFO: Organizing processes into subprocess groups
> INFO: Generating Helas calls for process: f1+ f1+ > f1+ f1+ WEIGHTED=2
> INFO: Processing color information for process: f1+ f1+ > f1+ f1+
> INFO: Creating files in directory P0_f1pf1p_f1pf1p
> INFO: Generating Feynman diagrams for Process: f1+ f1+ > f1+ f1+ WEIGHTED=2
> INFO: Finding symmetric diagrams for subprocess group f1pf1p_f1pf1p
> History written to /home/teilab/mg/model3test/Cards/proc_card_mg5.dat
> Generated helas calls for 1 subprocesses (3 diagrams) in 0.011 s
> Wrote files for 12 helas calls in 0.154 s
> Export UFO model to MG4 format
> CRITICAL: aS not define as external parameter adding it!
> WARNING: aEWM1 and Gf not define in MODEL. AQED will not be written correcty in LHE FILE
> ALOHA: aloha creates FFS2 routines
> Command "output model3test" interrupted with error:
> NameError : name 'exp' is not defined
> Please report this bug on https://bugs.launchpad.net/madgraph5
> More information is found in 'MG5_debug'.
> Please attach this file to your report.
>
> Could you please have a look at the attached models and tell me what the
> problem is? The debug file is also attached
>
> Thanks,
> Khaled
>
> ** Affects: mg5amcnlo
> Importance: Undecided
> Status: New
>
> ** Attachment added: "models and debug file"
> https://bugs.launchpad.net/bugs/1468429/+attachment/4420064/+files/spiegeltest.tar.gz
>
> --
> You received this bug notification because you are subscribed to
> MadGraph5_aMC@NLO.
> https://bugs.launchpad.net/b...

Read more...

Changed in mg5amcnlo:
status: New → Won't Fix
Revision history for this message
Khaled Teilab (teilab) wrote : Re: [Bug 1468429] Re: Form factors with transposed spinors bug
Download full text (6.6 KiB)

Dear Olivier,
thanks for your reply. I understand that your last sentence is a comment
on my model (not an instruction to overcome the problem). If so, I have
two questions:
1) Why model1 (the one with no form factors) is working and producing
results which are in agreement with independent calculations?
2) Since model1 is working, I expect that model3 should also be working.
Here, the problematic vertex with two identical fermions is treated the
same way as in model1 (no form factor)? The form factor is only used for
the normal t-channel vertices.
Cheers,
Khaled

On 24.06.2015 23:59, Olivier Mattelaer wrote:
> Dear Khaled,
>
> The problem is that first you are suppose to use
> cmath.exp and not exp
> cmath.sqrt and not sqrt
> in most of the case this should go trough since those will be handle by fortran anyway but not in this context.
>
> However, you have a four fermion interaction with a fermion flow violation (with identical particle) which is one case that we can not handle.
> In order to generate events for your theory you need to introduce fake scalar to split the 4 point interaction in a succession of three point interaction.
>
> Cheers,
>
> Olivier
>
>
> On 24 Jun 2015, at 17:11, Khaled Teilab <email address hidden> wrote:
>
>> Public bug reported:
>>
>> Hi,
>> I try to include a form factor in a model having two scalar particles (sig, phi) and two fermions (f1 and f2).
>> The lagrangian used in feynrules looks like:
>> g_1 Inner[Dot,Inner[Dot,{f1bar,f2bar },sig IdentityMatrix[2]],{f1,f2 }] + I g_2 (phi(CC[f1bar].Ga[5].f1) + anti[phi](f1bar.Ga[5].CC[f1]))
>>
>> In the attachment you find four models:
>> 1) Model1: Includes ALL vertices coming from the above Lagrangian WITHOUT form factor => WORKS
>> 2) Model4: Includes ONLY vertices involving the sig particle WITH form factor => WORKS
>> 3) Model2: Includes ALL vertices coming from the above Lagrangian WITH form factor on ALL vertices => FAILS to produce output folder
>> 4) Model3: Includes ALL vertices coming from the above Lagrangian WITH form factor on ONLY vertices involving the sig particle => FAILS to poruce output folder
>>
>> The output looks like:
>>
>> INFO: initialize a new directory: model3test
>> INFO: remove old information in model3test
>> INFO: Organizing processes into subprocess groups
>> INFO: Generating Helas calls for process: f1+ f1+ > f1+ f1+ WEIGHTED=2
>> INFO: Processing color information for process: f1+ f1+ > f1+ f1+
>> INFO: Creating files in directory P0_f1pf1p_f1pf1p
>> INFO: Generating Feynman diagrams for Process: f1+ f1+ > f1+ f1+ WEIGHTED=2
>> INFO: Finding symmetric diagrams for subprocess group f1pf1p_f1pf1p
>> History written to /home/teilab/mg/model3test/Cards/proc_card_mg5.dat
>> Generated helas calls for 1 subprocesses (3 diagrams) in 0.011 s
>> Wrote files for 12 helas calls in 0.154 s
>> Export UFO model to MG4 format
>> CRITICAL: aS not define as external parameter adding it!
>> WARNING: aEWM1 and Gf not define in MODEL. AQED will not be written correcty in LHE FILE
>> ALOHA: aloha creates FFS2 routines
>> Command "output model3test" interrupted with error:
>> NameEr...

Read more...

Revision history for this message
Olivier Mattelaer (olivier-mattelaer) wrote :
Download full text (9.8 KiB)

Dear Khaled,

I have not try model1 so i can not comment on it.
When I ran model3 after the modification that I mention before. MG refuses to make the computation for the process that you mention
because it can not handle it. This is process specific, so it is possible that you test for another process, I do not know but if they are indeed the same model then I would expect both to fail.

Cheers,

Olivier

On 27 Jun 2015, at 15:02, Khaled Teilab <email address hidden> wrote:

> Dear Olivier,
> thanks for your reply. I understand that your last sentence is a comment
> on my model (not an instruction to overcome the problem). If so, I have
> two questions:
> 1) Why model1 (the one with no form factors) is working and producing
> results which are in agreement with independent calculations?
> 2) Since model1 is working, I expect that model3 should also be working.
> Here, the problematic vertex with two identical fermions is treated the
> same way as in model1 (no form factor)? The form factor is only used for
> the normal t-channel vertices.
> Cheers,
> Khaled
>
>
> On 24.06.2015 23:59, Olivier Mattelaer wrote:
>> Dear Khaled,
>>
>> The problem is that first you are suppose to use
>> cmath.exp and not exp
>> cmath.sqrt and not sqrt
>> in most of the case this should go trough since those will be handle by fortran anyway but not in this context.
>>
>> However, you have a four fermion interaction with a fermion flow violation (with identical particle) which is one case that we can not handle.
>> In order to generate events for your theory you need to introduce fake scalar to split the 4 point interaction in a succession of three point interaction.
>>
>> Cheers,
>>
>> Olivier
>>
>>
>> On 24 Jun 2015, at 17:11, Khaled Teilab <email address hidden> wrote:
>>
>>> Public bug reported:
>>>
>>> Hi,
>>> I try to include a form factor in a model having two scalar particles (sig, phi) and two fermions (f1 and f2).
>>> The lagrangian used in feynrules looks like:
>>> g_1 Inner[Dot,Inner[Dot,{f1bar,f2bar },sig IdentityMatrix[2]],{f1,f2 }] + I g_2 (phi(CC[f1bar].Ga[5].f1) + anti[phi](f1bar.Ga[5].CC[f1]))
>>>
>>> In the attachment you find four models:
>>> 1) Model1: Includes ALL vertices coming from the above Lagrangian WITHOUT form factor => WORKS
>>> 2) Model4: Includes ONLY vertices involving the sig particle WITH form factor => WORKS
>>> 3) Model2: Includes ALL vertices coming from the above Lagrangian WITH form factor on ALL vertices => FAILS to produce output folder
>>> 4) Model3: Includes ALL vertices coming from the above Lagrangian WITH form factor on ONLY vertices involving the sig particle => FAILS to poruce output folder
>>>
>>> The output looks like:
>>>
>>> INFO: initialize a new directory: model3test
>>> INFO: remove old information in model3test
>>> INFO: Organizing processes into subprocess groups
>>> INFO: Generating Helas calls for process: f1+ f1+ > f1+ f1+ WEIGHTED=2
>>> INFO: Processing color information for process: f1+ f1+ > f1+ f1+
>>> INFO: Creating files in directory P0_f1pf1p_f1pf1p
>>> INFO: Generating Feynman diagrams for Process: ...

Revision history for this message
Khaled Teilab (teilab) wrote :

Hello Olivier,
actually -apart from the form factor- both models are exactly the same
...and... for the same process model1 produces very reasonable results !!!
MG only fails to calculate the process when form factors are included.
Cheers,
Khaled

On 27.06.2015 16:59, Olivier Mattelaer wrote:
> Dear Khaled,
>
> I have not try model1 so i can not comment on it.
> When I ran model3 after the modification that I mention before. MG refuses to make the computation for the process that you mention
> because it can not handle it. This is process specific, so it is possible that you test for another process, I do not know but if they are indeed the same model then I would expect both to fail.
>
> Cheers,
>
> Olivier
>
>
>

Revision history for this message
Olivier Mattelaer (olivier-mattelaer) wrote :
Download full text (3.5 KiB)

Hi,

So in principle the fix that I mention should fix the problem.

Cheers,

Olivier

On 27 Jun 2015, at 17:00, Khaled Teilab <email address hidden> wrote:

> Hello Olivier,
> actually -apart from the form factor- both models are exactly the same
> ...and... for the same process model1 produces very reasonable results !!!
> MG only fails to calculate the process when form factors are included.
> Cheers,
> Khaled
>
> On 27.06.2015 16:59, Olivier Mattelaer wrote:
>> Dear Khaled,
>>
>> I have not try model1 so i can not comment on it.
>> When I ran model3 after the modification that I mention before. MG refuses to make the computation for the process that you mention
>> because it can not handle it. This is process specific, so it is possible that you test for another process, I do not know but if they are indeed the same model then I would expect both to fail.
>>
>> Cheers,
>>
>> Olivier
>>
>>
>>
>
> --
> You received this bug notification because you are subscribed to
> MadGraph5_aMC@NLO.
> https://bugs.launchpad.net/bugs/1468429
>
> Title:
> Form factors with transposed spinors bug
>
> Status in MadGraph5_aMC@NLO Generator:
> Won't Fix
>
> Bug description:
> Hi,
> I try to include a form factor in a model having two scalar particles (sig, phi) and two fermions (f1 and f2).
> The lagrangian used in feynrules looks like:
> g_1 Inner[Dot,Inner[Dot,{f1bar,f2bar },sig IdentityMatrix[2]],{f1,f2 }] + I g_2 (phi(CC[f1bar].Ga[5].f1) + anti[phi](f1bar.Ga[5].CC[f1]))
>
> In the attachment you find four models:
> 1) Model1: Includes ALL vertices coming from the above Lagrangian WITHOUT form factor => WORKS
> 2) Model4: Includes ONLY vertices involving the sig particle WITH form factor => WORKS
> 3) Model2: Includes ALL vertices coming from the above Lagrangian WITH form factor on ALL vertices => FAILS to produce output folder
> 4) Model3: Includes ALL vertices coming from the above Lagrangian WITH form factor on ONLY vertices involving the sig particle => FAILS to poruce output folder
>
> The output looks like:
>
> INFO: initialize a new directory: model3test
> INFO: remove old information in model3test
> INFO: Organizing processes into subprocess groups
> INFO: Generating Helas calls for process: f1+ f1+ > f1+ f1+ WEIGHTED=2
> INFO: Processing color information for process: f1+ f1+ > f1+ f1+
> INFO: Creating files in directory P0_f1pf1p_f1pf1p
> INFO: Generating Feynman diagrams for Process: f1+ f1+ > f1+ f1+ WEIGHTED=2
> INFO: Finding symmetric diagrams for subprocess group f1pf1p_f1pf1p
> History written to /home/teilab/mg/model3test/Cards/proc_card_mg5.dat
> Generated helas calls for 1 subprocesses (3 diagrams) in 0.011 s
> Wrote files for 12 helas calls in 0.154 s
> Export UFO model to MG4 format
> CRITICAL: aS not define as external parameter adding it!
> WARNING: aEWM1 and Gf not define in MODEL. AQED will not be written correcty in LHE FILE
> ALOHA: aloha creates FFS2 routines
> Command "output model3test" interrupted with error:
> NameError : name 'exp' is not defined
> Please report this bug on https://bugs.launchpad.net/madgr...

Read more...

Revision history for this message
Khaled Teilab (teilab) wrote :

On 27.06.2015 18:37, Olivier Mattelaer wrote:
> Hi,
>
> So in principle the fix that I mention should fix the problem.
>
> Cheers,
>
> Olivier
>
Hi Olivier,

you mean the cmath functions? Unfortunately it doesn't :(
However, now the error message has changed to:

ALOHA: aloha creates FFS2 routines
unsupported operand type(s) for ** or pow(): 'AddVariable' and 'int'
Command "output" interrupted with error:
ALOHAERROR : Unable to deal with 4(or more) point interactions
        in presence of majorana particle/flow violation
Please report this bug on https://bugs.launchpad.net/madgraph5

Greetings,
Khaled

Revision history for this message
Olivier Mattelaer (olivier-mattelaer) wrote :
Download full text (3.8 KiB)

Hi Khaled,

> unsupported operand type(s) for ** or pow(): ‘AddVariable' and 'int'

This is probably something that you can handle by writing explicitly the product.
The square is sometimes difficult for the code since it needs to know on which index to contract.

> ALOHAERROR : Unable to deal with 4(or more) point interactions
> in presence of majorana particle/flow violation

This is the error I mentioned before that MG can not handle 4 fermion interaction with fermion flow violation.

Cheers,

Olivier

On 27 Jun 2015, at 18:16, Khaled Teilab <email address hidden> wrote:

> On 27.06.2015 18:37, Olivier Mattelaer wrote:
>> Hi,
>>
>> So in principle the fix that I mention should fix the problem.
>>
>> Cheers,
>>
>> Olivier
>>
> Hi Olivier,
>
> you mean the cmath functions? Unfortunately it doesn't :(
> However, now the error message has changed to:
>
> ALOHA: aloha creates FFS2 routines
> unsupported operand type(s) for ** or pow(): 'AddVariable' and 'int'
> Command "output" interrupted with error:
> ALOHAERROR : Unable to deal with 4(or more) point interactions
> in presence of majorana particle/flow violation
> Please report this bug on https://bugs.launchpad.net/madgraph5
>
> Greetings,
> Khaled
>
> --
> You received this bug notification because you are subscribed to
> MadGraph5_aMC@NLO.
> https://bugs.launchpad.net/bugs/1468429
>
> Title:
> Form factors with transposed spinors bug
>
> Status in MadGraph5_aMC@NLO Generator:
> Won't Fix
>
> Bug description:
> Hi,
> I try to include a form factor in a model having two scalar particles (sig, phi) and two fermions (f1 and f2).
> The lagrangian used in feynrules looks like:
> g_1 Inner[Dot,Inner[Dot,{f1bar,f2bar },sig IdentityMatrix[2]],{f1,f2 }] + I g_2 (phi(CC[f1bar].Ga[5].f1) + anti[phi](f1bar.Ga[5].CC[f1]))
>
> In the attachment you find four models:
> 1) Model1: Includes ALL vertices coming from the above Lagrangian WITHOUT form factor => WORKS
> 2) Model4: Includes ONLY vertices involving the sig particle WITH form factor => WORKS
> 3) Model2: Includes ALL vertices coming from the above Lagrangian WITH form factor on ALL vertices => FAILS to produce output folder
> 4) Model3: Includes ALL vertices coming from the above Lagrangian WITH form factor on ONLY vertices involving the sig particle => FAILS to poruce output folder
>
> The output looks like:
>
> INFO: initialize a new directory: model3test
> INFO: remove old information in model3test
> INFO: Organizing processes into subprocess groups
> INFO: Generating Helas calls for process: f1+ f1+ > f1+ f1+ WEIGHTED=2
> INFO: Processing color information for process: f1+ f1+ > f1+ f1+
> INFO: Creating files in directory P0_f1pf1p_f1pf1p
> INFO: Generating Feynman diagrams for Process: f1+ f1+ > f1+ f1+ WEIGHTED=2
> INFO: Finding symmetric diagrams for subprocess group f1pf1p_f1pf1p
> History written to /home/teilab/mg/model3test/Cards/proc_card_mg5.dat
> Generated helas calls for 1 subprocesses (3 diagrams) in 0.011 s
> Wrote files for 12 helas calls in 0.154 s
> Export UFO model to MG4 format
> CRITICAL: aS not define...

Read more...

Revision history for this message
Khaled Teilab (teilab) wrote :

Hi Olivier,

>> ALOHAERROR : Unable to deal with 4(or more) point interactions
>> in presence of majorana particle/flow violation
>
> This is the error I mentioned before that MG can not handle 4 fermion
> interaction with fermion flow violation.

The point is that "model1" which has exactly the same process is just
working nicely without even a warning. Maybe you could try it yourself.

Greetings,
Khaled

On Sat, June 27, 2015 10:04 pm, Olivier Mattelaer wrote:
> Hi Khaled,
>
>> unsupported operand type(s) for ** or pow(): ‘AddVariable' and 'int'
>
> This is probably something that you can handle by writing explicitly the
> product.
> The square is sometimes difficult for the code since it needs to know on
> which index to contract.
>
>> ALOHAERROR : Unable to deal with 4(or more) point interactions
>> in presence of majorana particle/flow violation
>
> This is the error I mentioned before that MG can not handle 4 fermion
> interaction with fermion flow violation.
>
> Cheers,
>
> Olivier
>

Revision history for this message
Olivier Mattelaer (olivier-mattelaer) wrote :

HI,

Looks like you do not have a 4 fermion vertex in your model.
the ALOHA package (arXiv:1108.2041) was missing the support form-factor in presence of fermion flow violation.

in attachment you will find a patch to that package which fix the problem.

Cheers,

Olivier

Changed in mg5amcnlo:
status: Won't Fix → Fix Committed
Revision history for this message
Khaled Teilab (teilab) wrote :

On 28.06.2015 16:48, Olivier Mattelaer wrote:
> HI,
>
> Looks like you do not have a 4 fermion vertex in your model.
> the ALOHA package (arXiv:1108.2041) was missing the support form-factor in presence of fermion flow violation.
>
> in attachment you will find a patch to that package which fix the
> problem.
>
> Cheers,
>
> Olivier
>
>
> ** Changed in: mg5amcnlo
> Status: Won't Fix => Fix Committed
>
> ** Attachment added: "aloha_patch.2.3.1"
> https://bugs.launchpad.net/mg5amcnlo/+bug/1468429/+attachment/4421539/+files/aloha_patch.2.3.1
>
Hi Olivier,
thank you. It works now.
Cheers,
Khaled

Changed in mg5amcnlo:
status: Fix Committed → Fix Released
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.