Hey Dave,
I'm not sure that you would get a HUGE performance gain, but you would realize SOME improvement if you try reducing the number of plugin projects you have. For example, you may find that 5 individual plugin projects (let's say portlets) can be collapsed into a single plugin. That alone would reduce the number of deployments from 20 - 15 which would not only speed up the deployment time, but also allow you to go longer between restarts. Personally, for my projects, I always start with just ONE plugin project per type and only break out the plugins into additional projects as necessary. For me this has been an effective way to ensure quick startup times (I have projects with 20, 30 portlet definitions but only one plugin project).
With that said, I am not using grails so I am not sure what you can do to help that. Depending on your version of Liferay, perhaps using Gradle as a build tool which intelligently only builds project items that have actually changed. And along that train of thought... 20 grails portlets build grails 20 times over, reduced to one or two grails portlet plugin projects would also mean X fewer grails builds
If you follow a common pattern (and naming convention) for your portlet development, it shouldn't take long to collapse the projects into a handful to try it out.
I'm not sure that you would get a HUGE performance gain, but you would realize SOME improvement if you try reducing the number of plugin projects you have. For example, you may find that 5 individual plugin projects (let's say portlets) can be collapsed into a single plugin. That alone would reduce the number of deployments from 20 - 15 which would not only speed up the deployment time, but also allow you to go longer between restarts. Personally, for my projects, I always start with just ONE plugin project per type and only break out the plugins into additional projects as necessary. For me this has been an effective way to ensure quick startup times (I have projects with 20, 30 portlet definitions but only one plugin project).
With that said, I am not using grails so I am not sure what you can do to help that. Depending on your version of Liferay, perhaps using Gradle as a build tool which intelligently only builds project items that have actually changed. And along that train of thought... 20 grails portlets build grails 20 times over, reduced to one or two grails portlet plugin projects would also mean X fewer grails builds
If you follow a common pattern (and naming convention) for your portlet development, it shouldn't take long to collapse the projects into a handful to try it out.