Its difficult to make midi drums sound realistic,so having the different sounds on separate tracks lets you go back at mixdown and say"I want a bit more level on just the kick" or whatever.
Effects are what eats up the processor,because all the data resides on the hard drive and not in RAM like other programs.If your system starts to bog down because of the # of tracks,that is when a sub-mix comes in handy.
As an example say you have three-part harmony vocals we want to reduce to a submix.Mute everything else and adjust your harmony voices to get a good blend.Mixdown those blended three and call the wave "backup vocals".Reimport that backup wave to the program and mute the three original tracks it came from.Now you can adjust levels on the backups relative to the main vocal but all the harmony voices will maintain their relative positions as you do so.
This same idea works on drums.Do the individual tracks,make up a punchy submix and mixdown,reimport the drum wave and now you have the ability to do levels on the entire kit with one fader.If you later change your mind and want a reverse gate on the snare,say;dump the submix and unmute the original tracks for a fresh submix.All the muted tracks won't affect the processor at all (except a small bit of horsepower required to draw the graphics).
This post has gotten long so I'll sign off.Good luck on your midi adventures!
Tom