I believe every modeler has some specific restrictions.
As far as I remember the FM3 had some limit on the amp blocks you can use in a preset.
However, @HonestOpinion made a very good statement.
I also believe in 202x they should all be able to distribute the CPU load on the cores without bothering or restricting the user.
You should be able to put blocks and connections wherever you want until you’re at 90+ CPU load overall and the unit should distribute it cleverly internally.
The idea is that you get more processing power for each signal flow - so if you are using Scenes in your presets then you could have distortion with rhythm and lead as scenes on 1 and 3 and have clean amps and other things on 2 and 4. Personally I use the volume utility to turn off for the unused path in a scene and on for the one I am using. This way there is no latency in channel switching - other scenes have different settings (obviously) for different effects
Revisiting this as I’ve been playing around with a Headrush Core recently. It has some nice features, one is how flexible the signal path options are if you are going to have only a couple of devices/models on one particular line.
Bear in mind that the Core is designed very much for singer/guitar, there are some path options that have only 3 devices for the mic input, which frees up more for the guitar input.
It would be nice if this could be done on the QC; as things stand, IIRC, if you are using all 4 inputs on the QC, you are limited to only 8 blocks per line, EVEN if one of the lines only has 3 blocks on it (say for example the line for a mic input has a compressor, a reverb and an EQ) the other 5 spaces on that row aren’t available to any of the other inputs. A great way to make the QC even more flexible would be to be able to add an input or a split halfway along a row if the input that is using that row doesn’t need all the block slots on that row…
Not sure if that makes sense??