Crash in PHOTONS

[ponyisi@gmail.com]

Sherpa 2 crashes when generating specific t tbar gamma* (Run.dat attached). The error message is as follows:

Exception_Handler::GenerateStackTrace(..): Generating stack trace {

0x2ba699616ed0 in 'ATOOLS::Terminate()' (Exception_Handler.C:135) 0x39134bd0fd in 'operator new(unsigned long)' 0x2ba6a3b86770 in 'PHOTONS::Dress_Blob_Base::GenerateNumberAndEnergies()' (new_allocator.h:89) 0x2ba6a3b862db in 'PHOTONS::Dress_Blob_Base::GeneratePhotons(double const&, double const&)' (Dress_Blob_Base.C:248) 0x2ba6a3b836ce in 'PHOTONS::Dipole_FI::AddRadiation()' (Dipole_FI.C:93) 0x2ba6a396e631 in 'PHOTONS::Define_Dipole::AddRadiation()' (Dress_Blob_Base.H:83) 0x2ba6a396d24a in 'PHOTONS::Photons::AddRadiation(ATOOLS::Blob*)' (Photons.C:111) 0x2ba69b39f32b in 'SHERPA::Soft_Photon_Handler::AddRadiation(ATOOLS::Blob*)' (Blob.H:159) 0x2ba698e824d0 in 'SHERPA::Signal_Process_FS_QED_Correction::Treat(ATOOLS::Blob_List*, double&)' (Signal_Process_FS_QED_Correction.C:210) 0x2ba698e665a8 in 'SHERPA::Event_Handler::IterateEventPhases(SHERPA::eventtype::code&, double&)' (Event_Handler.C:176) 0x2ba698e67838 in 'SHERPA::Event_Handler::GenerateStandardPerturbativeEvent(SHERPA::eventtype::code&)' (Event_Handler.C:245) 0x2ba698e651ff in 'SHERPA::Event_Handler::GenerateEvent(SHERPA::eventtype::code)' (Event_Handler.C:123) 0x2ba698c480c3 in 'SHERPA::Sherpa::GenerateOneEvent(bool)' (Sherpa.C:197) 0x404512 in 'main' (Main.C:37)

} Exception_Handler::Terminate(): Pre-crash status saved to '/work/02130/ponyisi/sherpawork/ttll-lowm/sherpa2/sample/StatusWed_Jan1_12-31-00_2014'. Event_Handler::MPISync() {

Memory usage increased by 4098 MB, now 4142 MB. This might indicate a memory leak! Please monitor this process closely.

} In Event_Handler::Finish : Summarizing the run may take some time. ERROR in Event_Handler::Finish:

After event : 15 / 3 particles / blobs undeleted ! Continue and hope for the best.

I'd guess it's a memory leak in the QED radiation generation?

[ponyisi]

Run.dat that causes crash

[marek]

Hi,

thanks for reporting this bug. Yes, this is a memory leak, the "new" in Photons merely is the one to run over the limit. We will investigate the source of this leak.

Cheers, Marek