Overview
The RMark package can be installed and run under recent distrubutions of the GNU/Linux operating system (hereafter, Linux), or recent flavors of OS/X.
The following assumes you have a basic familiarity with Linux and/or OS/X, and have at least sudo access to specific directories (if you don't know what that is, then you should probably stop now...). It also assumes that the machine you're planning on using has (i) the latest verion of R installed (at least 4.1.x) and running, and (ii) at least gcc 8 (or newer) installed (including all the libs).
There are a few steps to install MARK and RMark under Linux or OS/X, described below.
Step (1) - installing MARK executable
- download the g'zip'd tarball containing builds of the MARK
numerical routines executable - for Linux, grab the file archived here.
The archive contains both a 64-bit a statically linked compilation
(also 64-bit) -- this should work 'as is' on most Linux distributions.
For OS/X, grab the files archived here. The archive contains a 64-bit only version of the MARK executable.
- for Linux users, extract the file (mark64.dynamic, or mark54.static) from mark.64.zip (and rename as desired). For Mac users, extract the file mark.64.osx from mark.osx.zip
- rename the file 'mark' (i.e, drop the .64 part, or 64.osx parts), copy it to a directory in the path (typically, /usr/local/bin for Linux users, or wherever you prefer for OS/X types), and make it executable
- make sure it runs by typing mark from the command line (or ./mark if trying to run it from the directory in which you placed the executable)
- If you get an error (something like "dyld: lazy symbol binding failed..."), you'll most likely need to install gfortran from http://gcc.gnu.org/wiki/GFortranBinaries#MacOS. This will generally involve installing Xcode and the associated command line tools, and then gfortran.
Step (2) - installing RMark under Linux or OS/X...
Here are the basic steps...
- Install the RMark package from CRAN.
- For the Linux version of RMark (and probably for OS/X as well), it expects the
executable for mark to
be named 'mark' and to be found in the search path. The variable MarkPath is
ignored.
If you store it some place not in your path you can create a symbolic link. For example, you
can install the
executable for MARK where
you want (e.g.
~/Programs/MARK/ ) and
in /usr/local/bin you'd type
ln -s ~/Programs/MARK/mark mark
- One other setting is the name of the editor/file viewer. The default is pico. If you want to use something other than pico or your machine does not have pico. Simply type MarkViewer="myviewer" in your R session before using RMark (for OS/X, MarkViewer="open -a myviewer"), where myviewer is the name of an editor or file viewer that is in your path. You can put the library call and the MarkViewer assignment into a function .First() and they will be run each time you start the R session in that directory.
Misc notes...
The Linux-compiled version of MARK were built using gfortran 10.x. (Most modern distros will have gfortran >=9 installed -- for other less bleeding-edge (and thus way more stable) distros like CentOS or Redhat -- you might need to install gfortran 10.xx alongside the older version that comes bundled. This isn't hard. Search for 'devtoolset' and 'gfortran' and you should find the basic information you need). There is some chance that the libraries on your distribution may be in different locations than the MARK executable is expecting. You can LDD the MARK executable to find out what it 'requires', then tweak the LD_LIBRARY_PATH as needed (if the preceding sentence doesn't make sense to you, you should probably quit now, and ask a local 'guru' for help... ;-) '
On the Mac side of things, there are (unfortunately), many more steps that you'll need to follow (despite what Mac users might think, OS/X isn't as simple as a pretty GUI on a Linux kernel). A very nice + thorough collation can be found here and here. Note that the sequence of things you need to do to get things to work under OS/X might (and probably will) vary as OS/X evolves.