Why is it that when I place audio effects on recorded audio tracks (real world) versus digital to audio tracks, why does it use so much cpu?  Is there a way to remediate after the fact and/or mitigate during recording?  I''d say my cpu usage is cut in half when I take the audio effects off of those tracks.(??)

I don't understand what you mean with "digital to audio tracks" but any effect you place in an FX bin, ProChannel FX chain or on a clip has to be calculated in real time by your CPU. By definition that will take up CPU cycles. Various things influence how much;
1. How CPU intensive the plugin is (e.g. generally reverbs use more CPU than EQ and more recent plugins tend to be heavier than older ones). You could try substituting lighter plugs. 
2. How fast/recent your CPU is, not much you can do about this without getting a new computer.
3. The samplerate you're working at (higher samplerate=>more CPU), you could go down to 44.1KHz if you're at something higher. 
4. The buffer size/latency settings for your soundcard. With a good soundcard with proper ASIO drivers you can get buffer sizes as small as 64 samples or even 32 samples. This allows for incredibly low latency which is nice if you want to play aoftware synths or monitor through your DAW. But it also puts a lot more strain on your CPU than a higher buffer size. If you're in the mixing stage, you could try setting a buffer size of 512 or even 1024 samples. This means the computer will be slower to respond to real time input but during the mixing phase this is less of an issue and it will allow your computer to "look ahead" a bit more and spread out the needed CPU processing, leading to lower strain.

eph221
Why is it that when I place audio effects on recorded audio tracks (real world) versus digital to audio tracks, why does it use so much cpu?  Is there a way to remediate after the fact and/or mitigate during recording?  I''d say my cpu usage is cut in half when I take the audio effects off of those tracks.(??)

I have no idea what you are saying. What is real world? 

One way to remediate after the fact is to freeze those tracks as soon as you're done tweaking the recording and the effect settings.

Some idea of the OP's system would be a good idea, 'cos we don't have a clue what he's running nor what on!

Haha nm.

Thanks bvideo, that's what I was looking 4.  But does freezing an audio track (one that you've recorded from a real instrument) really free up resources?  what does feezing actually do?!  I mean I kind of understand that the computer would then be working with an audio wave file instead of a midi file(?) but isn't that what the sequencer is using when you record a real instrument.  I think I may have some misconceptions that need clearing up.
 

If I put the exact same effect on the fx bin of an audio track and a midi instrument track...the audio track tacks up vastly more cpu cycles.  That's what puzzles me.  (This is how my original question should have been framed, I apologize)
 
 

The FX plugin itself should use the same amount of CPU in either case, but the virtual instrument also uses CPU - probably more than the FX - so you would need to get a baseline with the instrument by itself before adding the FX. what instruments and FX are you referring to specifically, and does it matter whether the track is mono or stereo?
 
 

I realize it's counterintuitive, that's why I'm asking.   The audio track is usually mono, and the midi instrument tracks are usually stereo.  The midi instrument tracks use vastly less cpu cycles.  Same effect in both fx bins.  Maybe it has something to do with how they're routed through the master bus.