The PMP side of LINKPlanner has evolved from our successful PTP planning tool and the performance will start to slow when you increase the number of network features. There are no in-built limits, but users may start to experience problems under the following conditions:
A large number (> 1000) of one type of feature (site, link, AP, SM), particularly when displaying the Google Map view or table views
Bulk editing large numbers of objects in a table view (editing 20 APs could cause recalculations in 1000s of SMs)
Saving/loading projects with a large number of items (slow save/load or possibly even out of memory errors). This is a particular problem if there are 1000s of long links.
Exporting a project to Google Earth with path profile previews can cause out of memory errors
Some users are able to work with larger projects and other users see errors with smaller networks (particularly with complex PTP link types or very long links). Some of this comes down to the speed of the computer and the patience of the user. Multiple CPU cores do not help and more available memory doesn't necessarily help since LINKPlanner is a 32 bit application. A faster CPU and a solid-state drive will help.
There are some things that you can do to work-around these issues:
Switch off automatic calculations. This means that SMs will only update when you have finished editing an AP rather than after each change.
Split the network into multiple project files with small clusters of network sites and hubs (this also allows multiple users to work on the network at the same time)
Hide the Google Map
Reduce the number of formatting rules
Disable profile previews when exporting to Google Earth
We are constantly making incremental improvements to the performance, unfortunately certain new features may then cause things to slow down again. We are working on a new version of LINKPlanner that will work with very large networks (amongst other things), but this has required a complete redesign of the underlying code and we don't have a release date planned yet. We'll continue to release versions based on the current code base until that is ready, so unfortunately small performance improvements are all that we can offer in the short-term.