ttbb 5FS simulation: m_2^2 is negative in t-channel branching

[T. Verlage]

I tried to simulate ttbb using the 5fs in MG5_aMC@NLO (version 2.3.2.2) several times. Unfortunately, it always crashes during the event simulation in MadSpin with the error: "ValueError : could not convert string to float: Warning: m_2^2 is negative in t-channel branching"
Please find attached the log file: pilotrun_tag_1_debug.log
Is there anything I can do to avoid this error?

[Pierre Artoisenet (partois)]

Hi ,

Could you send me a sample of the event file (+ the banner ) ?
(100 events would be ok)

Pierre

[T. Verlage (tverlage)]

Hi Pierre,

Please find attached the event file and the corresponding log file. In addition I also uploaded the whole tar file. You can find it here: https://gigamove.rz.rwth-aachen.de/d/id/vK2BWKA9WiKEiQ I hope this helps.

Best regards
Tobias

[T. Verlage (tverlage)]

and here the corresponding logifle

[Olivier Mattelaer (olivier-mattelaer)]

Hi Pierre,

I have taken a look myself.
What do you think about the following fix?
The code goes trough after that but I'm not sure if the fix is correct or not.

it crash for a value of m2 squared is quite small -2.6432186714373529E-003 so it might be a numerical accuracy problem

Cheers,

Olivier

=== modified file 'MadSpin/src/driver.f'
--- MadSpin/src/driver.f 2015-01-29 14:35:06 +0000
+++ MadSpin/src/driver.f 2016-04-04 18:57:15 +0000
@@ -907,7 +907,8 @@
                 if (index_p2.gt.0) then
                    m2_tchan(i)=m(index_p2)
                 else
- write(*,*) 'Warning: m_2^2 is negative in t-channel branching '
+ m2_tchan(i)=0
+c write(*,*) 'Warning: m_2^2 is negative in t-channel branching '
                 endif
             endif
          ! extract phi

[Olivier Mattelaer (olivier-mattelaer)]

Pierre? Any update on this?

Olivier

[Olivier Mattelaer (olivier-mattelaer)]

Ok I have implemented the above patch in 2.5.2.

Cheers,

Olivier