12  Questions and Answers

I’m using this chapter as a place to provide miscellaneous information that might not have an obvious place to live. I’d call this a FAQ, but not all of the questions are asked frequently or at all.

12.1 Can I interrupt a simulation?

Starting with mrgsolve version 0.11.1, you can interrupt a long simulation by pressing Control-C or Esc, the standard way to pass an interrupt signal through R. mrgsolve will stop every so often to look for the interrupt signal.

You can control the frequency with which mrgsolve looks for the interrupt signal through an argument to mrgsim (default: 256 simulation records). Increase to check less frequently, increase to check more frequently (this might be needed for a model where a large amount of work is required to advance one step) or set to negative number to never check.

12.2 Can I pass compiler flags to my model?

Compiler flags can be passed to your model by setting PKG_CXXFLAGS in $ENV. For example

PKG_CXXFLAGS = "-std=c++11"

will compile your model according to C++11 standard (but note that there is a special plugin that will do this automatically for you; see Section 9.4.

12.3 Can I compile my model with C++11?

Yes, you can do this by invoking the CXX11 plugin (Section 9.4).

12.4 How can I calculate time after dose?

There are three approaches

tad argument to mrgsim()

To get time after dose into your output you can call

mrgsim(mod, tad = TRUE)

and the output will have a tad column. Note this does not let you interact with the tad value inside your model.

Simple calculation in the model

Most applications will call self.tad() (Section 2.3.21). For example

[ main ] 
double tad = self.tad();

More complicated calculation in the model

You can get more control and track tad in a specific compartment by using the tad plugin. See Section 9.3 for details.

12.5 My model failed to compile; what now?

The model can fail to compile for a variety of reasons, including an error in the C++ code or inability of R to find the compiler and other pieces of the tool chain.

If your model is not compiling, try the recover argument to mread() (or mcode())

mod <- mread(..., recover = TRUE)

You will see a warning on the console and mread() will return a list of build information. You can look into that information or share it in the mrgsolve issue tracker on GitHub.

If your model has C++ syntax problems, the errors should be printed on the console. If you possibly have problems with the compiler or the rest of the toolchain, take a look at the pkgbuild package, which provides some helpful tools, especially if you are working on a Windows platform


12.6 Can I run mrgsolve on a network drive?

No; do not run mrgsolve on a network drive. Your R installation, mrgsolve installation, and R working directory should be on a local hard disk.

12.7 Can I run mrgsolve on a cloud-synced folder?

No; do not run mrgsolve in a synced folder for cloud services like OneDrive, GoogleDrive, DropBox etc. Your R installation, mrgsolve installation, and R working directory should be on a local hard disk.

12.8 Can I run mrgsolve in a path that includes spaces?

No; do not run mrgsolve in a path that includes spaces. Your R installation, mrgsolve installation, and R working directory should all be in locations whose paths do not include spaces.