Fluka runtime error: SYNSTP reason SRAPHO overflow

Dear colleagues,
I am using latest FLUKA version 4.4.1 with a custom executable. In my simulation, I get the following error:

 Abort called from SYNSTP reason SRAPHO overflow Run stopped!
 STOP SRAPHO overflow

In my simulation, I have a primary 100 GeV/c pi- propagating in a setup that includes a vacuum region with magnetic field, where synchrotron radiation emission is enabled.

From the name of the routine causing the abort, SYNSTP, it came to my mind another error that I observed in the past (Fluka fortran runtime error, error cascade/synstp.f). As I did at that time, I tried to disable synchrotron radiation, and the error is no longer present.

I attach below the input card (with synchrotron radiation enabled) and the necessary files to compile the executable. The card is called “na64_2023A_h.inp”

Thanks,
Bests,
Andrea

data.tar.tgz (55.3 KB)

Hello Andrea,

Apparently the header NA64StrawHit.hh is missing in the tarball. Likewise, the Makefile seems to invoke a rndcrd.o with a missing .f counterpart in the tarball.

We’re trying to proceed in the meantime with a trimmed down version hoping to reproduce the abort you witness, thus far running without crashing.

Feel free to share the missing headers/files meanwhile so as to be in identical (as much as possible) lab conditions.

Thanks in advance,

Cesc

Hi Cesc,
apologies for my mistake, please find attached the new data.tar.gz with all files.
Concerning rndcrd.o, this is an object file that was kindly provided by @ceruttif (see: Re-simulate a specific primary history - #7 by ceruttif) in order to get the value of WHAT(2) from RANDOMIZ.

Thanks,
Bests,

Andrea

data.tar.tgz (36.0 KB)

Dear Andrea,

While we have not been able to explicitly reproduce the crash you report with your full simulation input (now complete, thanks), we have a distinct suspicion.

You start with ~100 GeV \pi^- (with a small angular and momentum divergence) traversing mostly air and vacuum, until they eventually reach a dipolar magnetic field in your MAG0 and MAG1 regions.

You might produce electrons (be it through pion decay, subsequent muon decay, etc) with an energy low enough that they spiral in your 1.85 T field along -y (this is needless to say a physical effect).

The crux of your crash is likely a combination of this effect with synchrotron radiation (SR) emission: you might have e- with energies low enough energy that they spiral but still high enough that they emit copious SR photons (down to 1 eV as per your input request). This vicious physical circle is what floods the SR photon stack.

The cure (besides a great band) is twofold:

  1. Raise the SR photon cut from 1e-9 GeV to, say, 1e-7 GeV. This kills the flood, but still leaves you with the spiraling e-. Sufficiently high-energy e- have a bending radius which does not allow a full circle, whereupon e- exit the B-field region without problem. At sufficiently low energies, however, they will start to spiral - without emitting SR photons.
  2. You’ll then run into the issue of low-energy e- spiraling forevermore in an everlasting uninterrupted fateful way, thus stalling your simulation. The cure to that is a glorious TIME-CUT card with a reasonable time window (maybe some 100 ns or whatever you deem necessary in your geometry).

Let us know if this allows you to overcome the issue.

Cheers to a nice weekend,

Cesc

PS: kudos to @dcalzola and @smarin for assistance and support.

1 Like

Dear Cesc,
I thank you for your detailed answer. I’ll try to apply the suggestion you pointed me. In case I’ll still observe the problem, I’ll come back to the forum.

Thanks,
Bests,
Andrea

1 Like