...and I mean less CPU-intensive...
In gereral, if I have a situation where I could use "audio"-signals or "data", which one is faster, or better, or whatever?!?

Let's say I have to read/ write an entire array. The arraysize is smaller than 1 block-size. Now I could use [tabread]/[tabwrite] or [tabread(4)~]/ [tabwrite~].
Is it possible to determine whats better???
Btw. what is the main difference between "audio"- and "data"-signals in terms of being processed "internally"??
What's your oppinion?

Greets Flipp

i think the only difference is the rate at which the messages are processed.

at control rate (data) i think the refresh rate is something around 70hz, but at audio rate, it's whatever you set your samplerate too (so well into the tens of thousands). audio rate therefore will be more intensive, but also more accurate.

If you can't hear a difference then control rate is gonna save you those CPU cycles.

What exactly do you mean with "refresh rate" ?! As far as I know, "data"-data is sent out as soon as you press a button.. but it can only interact with audio-data at each block-start (is that what you mean?).
Just load this patch and press the "10000"-Msg. (it sends out left to right order). The time you see is what it takes to bang 10000-times (in ms!). Just connect a [bang] or anything else to the [until] to use it as a stopwatch for your "subpatch"...
Actually I'd like to have something to measure performance for "audio" too (on windows)..

#N canvas 293 291 450 300 10;
#X obj 169 70 until;
#X obj 179 194 b;
#X obj 224 194 b;
#X obj 179 217 realtime;
#X floatatom 179 243 10 0 0 0 - - -;
#X msg 167 41 10000;
#X connect 1 0 3 0;
#X connect 2 0 3 1;
#X connect 3 0 4 0;
#X connect 5 0 1 0;
#X connect 5 0 0 0;
#X connect 5 0 2 0;

Is there maybe a difference in terms of where the data is processed?? or how, or what?!

Lets take a [count~]. It can be switched off. When it is off, will it still send or rather process data, so is the pc actually computing it at each sample?? ...since this isn't done with data-objects, as if they were "invisible" when not banged..I guess.

From what I understand, messages are essentially queued up when you send them, and they are computed and applied to the next block. They don't get re-computed on every block, though. They're event based.

Audio objects send and calculate a vector of values at every block. If you want to turn audio processing off for one or more objects, put them in a subpatch with the [switch~] object. Send it a [1( to turn on processing, and [0( to turn it off. It will only apply to the subpatch itself and it's child patches.