...

public class MyMojo
    extends AbstractMojo
{

  private static final Object lock = new Object();

  public void execute()
  {
     synchronized( lock) 
     {
         // Main mojo code
     }
  }
}

How Execution is evaluated

Image Added

For simplicity; let's assume all modules have an equal running time.

This build should have level 0 running first, then a fanout of up to 5 parallel on level 1. On level 2 you'll be running 3 parallel modules, and 7 on 3, 5 on level 4.

This goes by declared dependencies in the pom, and there is no good log of how this graph is actually evaluated. (I was hoping to render the actual execution graph, but never got around to finding a cool tool/way to do it - plaintext ascii in the -X log would be one option).

Of course, in real life your modules do not take equal amounts of time. Significant gains are common when the project has one or more "api" modules and dependencies on the "api" modules (and just bring the "impl" version of the module into the actual assembly that will be started). This design normally means your big chunky modules depend on lightweight "api" modules that build quickly.