Frequently Asked Questions¶
Q: My cells are not initialized at the right position¶
This is possible due to the lattice spacing usually (unless defined otherwise)
being 0.5 µm while the positions in the
.pos file are in µm. This means
that a cell center of
10 12 24 in a
.pos file will have a center of
20 24 48 in the simulation.
Q: Can the cells lie out of the domain?¶
Yes, cells are bound inside the domain by their centerpoints, so parts of the cell might reach over the domain. These cells are deleted during initialisation, unless the system is run on a single core.
Q: What is the exact centerpoint of the domain, how big is my domain?¶
When you specify a domain of for example 10 cells, the exact middle will be BETWEEN cell 5 and 6. imagine that dx is 0.5µm then the domain is 9.5µm long, and the middle is 4.75µm. periodic boundaries of course add 1 dx back to the length.
Q: Compiling with the singularity images gives an error within Cmake¶
We have had reports that the clang (or llvm) compiler can cause problems with compiling the singularity image. Disabling it (removing it from the environment) or using another system might work. We are looking into this issue.
Q: Compiling with singularity gives errors¶
Maybe you have an old
hemocell/ folder in the same directory and are trying to compile with a newer version. If this is the case you can rename
hemocell_old and the newer singularity image should create a new folder
Q: HemoCell seems to be leaking memory over time¶
We have experienced similar issues when compiling and running with OpenMPI version 2.0.x, older or newer versions generally seem to be ok. We have tried to track down the issue with heaptrack and only found memory leaks within the MCA part of OpenMPI.
Do you have a problem that is not listed here? please mail: email@example.com or create an issue at the github repository, you might help someone else as well!