Dosxyz error : Segmentation fault - invalid memory reference.

[ChillError404]

Hi everyone,

I’m trying to score surface dose in DOSXYZnrc using an extremely thin z voxel (~0.002 mm).

The simulation runs fine for smaller jobs (e.g. ~3 million histories).
However, when I increase the number of histories to about 7.84 × 10¹⁰, the run crashes with a segmentation fault at the end of the job.

Error message:

Program received signal SIGSEGV: Segmentation fault - invalid memory reference.

Backtrace for this error:
#0 0x7163dce23e59 in ???
#1 0x7163dce22e75 in ???
#2 0x7163dca4532f in ???
#3 0x5d1554e9c8cb in write_dose_
#4 0x5d1554f10447 in MAIN__
#5 0x5d1554e736fe in main
Segmentation fault (core dumped)

any possible way to fix this, thank you in advance

Dosxyz egsinp file
TB10_Surface #!GUI1.0
1
H2O700ICRU
0.521, 0.01, 0, 0
-1, -1, -3, 1
-7.25
0.5, 29
-7.25
0.5, 29
0
0.002, 5
0.09, 1
0.1, 19
0, 0, 0, 0, 0, 0, 0, 0
0, 0, 0, 0, 0, 0, 0, 0
0, 0, 0, 0, 0, 0, 0, 1
2, 2, 0, 0, 0, 180, 0, 0, 0, 0, 10, 100, 100, 50
2, 0, 1, 30, 1, 30, 38, 0.1
/home/admin/EGSnrc/egs_home/BEAM_TrueBeam/TB_FS10.egsphsp1
78400000000, 0, 99, 33, 97, 20, 0, 0, 1, 1, 1, 0, 0, 0, 50, 1, 0, 0
#########################
:Start MC Transport Parameter:

Global ECUT= 0.521
Global PCUT= 0.01
Global SMAX= 1e10
ESTEPE= 0.25
XIMAX= 0.5
Boundary crossing algorithm= EXACT
Skin depth for BCA= 3
Electron-step algorithm= PRESTA-II
Spin effects= On
Brems angular sampling= Simple
Brems cross sections= BH
Bound Compton scattering= Off
Compton cross sections= default
Pair angular sampling= Simple
Pair cross sections= BH
Photoelectron angular sampling= Off
Rayleigh scattering= Off
Atomic relaxations= Off
Electron impact ionization= Off
Photon cross sections= xcom
Photon cross-sections output= Off

:Stop MC Transport Parameter:
#########################

[rtownson]

If the simulation just crashes at the end it, after finishing the last batch, it might be worth trying to run again after changing the "run option" to "analyze previous", or "combine parallel run" just to see if it finishes properly when you try again. Are you running on a single core or in parallel?

A workaround would be to run a number of slightly shorter simulations and then combine the 3ddose files together (but you'd need to write a script to combine those 3ddose files, there's not one in egs).

[ChillError404]

Sorry for late reply, The problem occur with single core, I have tried with the parallel run but not finish yet.