#64 Segmentation Fault (core dumped) | PhysiCell 1.13.1 & 1.7.1 | Virtual Box running Ubuntu 22.04.3 LTS
Hi PhysiCell Team,
I am extremely new with PhysiCell, and aim to use and build simulations surrounding metabolic growth, phenotype tracking, and resource allocation of cells within a spheroid environment. There is a paper by Pierre Jacquet et al. that I am using as a reference; however, while I am able to compile and build the simulation, I seem to hit a "Segmentation fault (core dumped)" error after the first time step is printed in my terminal (text attached). It appears the version of PhysiCell used in their repository (https://gricad-gitlab.univ-grenoble-alpes.fr/timc-bcm/angelique/cancersimulationphysicell) is 1.7.1. The error occurred when running the simulation with their PhysiCell environment as well as when I migrated the scripts, custom modules, and settings into a new project directory in my PhysiCell 1.13.1 environment. I have also noticed the same error occurring when running an advanced sample project (PhysiMeSS).
From this, I can only gather that it might be a hardware issue on my end (50GB base memory, 6 processors), or simple user error from my lack of experience. If so, is there any way to remedy this issue in order to properly run and modify runs for this project? Attempts to reduce the number of threads and domain size yield the same error. I am hoping to get some guidance on adjustments/strategies I can make/use to get these simulations running. Thanks in advance!
Related
This is quite specialized. Do you have a model that doesn't require Boost?
In my efforts to get the model from the literature working, I installed any packages that patched errors returned when attempting to compile/build/run the simulation. I am trying to slowly climb the learning curve, so I'm not sure what packages or model parameters directly affect the performance of my machine. At this point, it seems like I may need to build the model from scratch (something I hoped to do after more experience) to better capture the behavior of our cells anyway. What puzzled me was the inability for the simulations to actually run on my machine, given I tried running the sample projects as well.
Let's start with something simpler. Can you provide your terminal output when compiling and trying to run the "template" project, i.e., these commands:
make reset
make template
make
project
You might also provide "g++ --version"
g++ (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0
I ran the heterogeneity-sample project as well and it ran smoothly. I realize the simulation from the literature had several custom modules and adjusted settings for their model; however, when running that simulation, I noticed the output file sizes were considerably large for the time-steps it was able to run (>2.0GB for a "cells.mat" file). The entire output folder for the heterogeneity-sample simulation was right at 2.0GB. In my case, does the segmentation fault (core dumped) error simply trigger due to the excessive size of the output files? I can re-install my VM with a larger disk-space, but even that has its limits. I was curious to see if there is a way to either compress or reduce the output file size returned during a simulation of growing complexity?
For starters try running that model with the full output flag to false.
Just save the SVG.
On Wed, Jan 10, 2024 at 9:06 AM Steven S ssummey@users.sourceforge.net
wrote:
Related
Tickets: #64
Unless there are multiple locations in that repository that reference output flags, I wasn't able to get the output to actually change when running the simulation. Before the first iteration (in my 1.13.1 folder) I am now seeing a new error:
ERROR : Unknown parameter oncoprotein_min ! Quitting.Looking through the custom modules, it looks like it is called in the heterogeneity_coloring_function in heterogeneity.cpp, specifically:
` static int oncoprotein_i = pCell->custom_data.find_variable_index("oncoprotein");
static double p_min = parameters.doubles("oncoprotein_min");
static double p_max = parameters.doubles("oncoprotein_max");`
but I cannot find an instance or assignment of those values anywhere else in the files (including PhysiCell and bioFVM). It looks like pieces left over from the heterogeneity sample project?
As for the segmentation fault (in my 1.7.1 folder), would you recommend going through the files that contain "#include <signal.h>" and commenting it out? </signal.h>