When to be rendering clips?

I learned the hard way to render after AudioSnap (uses a lot of mem.) but do we need to render after Automation too for the same reason?

I'm with Karl.

There is no benefit whatsoever in rendering your automation.

You can put Sonar in Offset mode to do that, if I recall correctly.

Ok thanks, I'll search for that for volume only. "O" seems to affect Panning at the same time.
I was typing this while you were answering Thanks for elaborating
.
John, I get the idea this does not come up for you often?
Maybe my process is out of order?

I'm getting my (3) stereo acoustic guitar tracks tweaked before recording the Bass, Vox or any other padding. 
That meant (After some Pro Ch presets) maxing and balancing the 6 levels and then reducing the stray peaks (set "show peak marker") with Process/Apply/Gain or an automated envelope.

I can see the need to adjust faders after new tracks of course.
Do folks get all the tracks in before peak tweaking?
Should I be using a limiter instead for this? (I just got the concrete limiter a few weeks ago and not sure if this is a good use for it?). 

Maybe not as conventional - but I sometimes use the gain/output of one of my fx (usually compressor) to bump down volumes in some instances.

I don't render it but as I work setting up the nodes and editing I do hit SAVE every few minutes.

I have a very stable system but still.....why take chances.

I have once or twice in the past bounced the track to another track with automation enabled and ended up with just the audible parts of the envelopes in the new track. 

I have not used it that way in quite a long time now. You never know when you might want to start that guitar part a few beats sooner.