Build a test suite of large models to maintain/improve speed

[dirkschumacher]

All unit tests cover relatively small models. It would be great to have a set of large models that test model building speed and memory consumption (especially of API functions). Those tests should not be part of the regular tests, but could be a simple rmarkdown document somewhere that can be executed whenever something is submitted to CRAN.

[rickyars]

Do you have any idea of what is the main bottleneck? I'm building relatively small models, just a couple hundred decision variables, and sometimes the build (MIPModel) can take 30 - 60 seconds to build.

Edit: Looks like the slowdown happens when I start adding constraints.

[dirkschumacher]

See #170