Restarting a filament simulation from a previous run results in a similar but not identical rerun.

[bascee]

As the random number generator does show continuity there is most likely a difference in a filial link being loaded from xml or it being continued from the previous model state, investigate..

[bascee]

The issue also occurs for other protocol files. The issue seems to revolve around a substrate small substrate concentration jump and for both the transient and the steady state solver. The issue does not occur if the initial state is written to XML and then that file is used to run the sim. There must be something that is determined in the previous time step and then used in the following without it being saved to the XML file.