In my experience and opinion, software models of hardware of compressors all sound about the same. The differences are subtle at best and the color or character they supposedly impart on the signal are also subtle at best unless and until you get into extreme settings (i.e. the 1176 all button mode etc).
 
Here's a thought experiment to show what I mean. Take an individual source track and clone it several times. Then use an 1176 on one track an LA2A on the next and Sonitus on the next and a Focusrite on the next etc. Next set all tracks to roughly the same (not extreme) amount of compression and same output level. If you did all that, I would bet that 99.999999% of the DAW users out there could not do any better than random chance at guessing which was which in a blind listening test.
 
I've said this before. The reason I choose one compressor plugin over another has nothing to do with sound. It has to do with the controls on that particular model and my understanding of how to use them. I gravitate toward the 1176 because I know how to get what I want out of it. I could get the same results with lots of other compressors, but it would take me longer because I don't know their controls as well. Also to be fair, I gravitate toward both the 1176 and the LA2A because they're in the prochannel which makes them quicker and easier to use and adjust.

So the bottom line is use which ever compressor you want because at the end of the day, compression is compression and it all sounds more or less the same regardless of which model you use. 

That being said, when I'm doing guitar usually go with the stomp box models rather than inserting a 'high end' model before the amp. I have a couple of reasons for that. First off, when I'm using compression on a guitar signal chain going into an amp, I'm usually trying to different things with the compressor than I when I'm using compression in a mix down setting.
 
Mixdown compression is often about taming dynamic range (keep the peaks under control) and sometimes also a bit of fattening. With guitar, its usually more about fattening and sustain than about keeping peaks under control. And like I said above, for me choosing which to use is about knowing how to get what I'm looking for from the control on that particular unit. When it comes to guitar compression, its faster and easier for me to use the stomp box models because I know how to set them. I could use an 1176, but chances are no one will hear the difference when its all said and done so why bother?

The modeled stomp boxes are based on compressors that were specifically designed to work with the characteristics of an electric guitar whereas the other compressors are general purpose studio compressors.  There are decades worth of countless recordings with those classic sounds baked in that people are used to hearing.  You'd drop a stomp box compressor into the chain in TH3 rather than use a separate LA-2A or similar if you wanted to get as close to that classic sound as possible without much fuss.  
 
An emulation of a classic studio signal chain might actually be a compressor stomp box upfront followed by the amp/cab/mic models in TH3 followed by a separate LA-2A or similar.  You can think of the stomp box compressor as a guitar effect and the LA-2A as the glue that pulls things together.  
 
That said, there are no rules.  If you put an LA-2A up front an like it, go for it.  There are stomp boxes out there that are actually based on LA-2A.  
 
Realize that compressors, particularly emulated compressors like the TH3 stomp boxes, CA-2A, etc., impart characteristics on the sound beyond just the theoretical compression.  You'd pick one or the other for a particular spot in the chain because you like those extra characteristics in that context.  
 

Kamikaze
I'm pretty new to the world of Guitars and Amp, I've never played through a real amp, so just have TH2 (3 sort of) and MarkBass. These come with stomps of compressors. What would be the benefit of using these before and amp over those in the DAW; CA2A, PC76 and Boz+10 for example. I'm guessing having a modelled high end compressor over a modeled stomp box.
 
I've also been investigating pick ups and note comments like 'more attack, more sustain, deeper sound, more even dynamics, more presence etc. These all seem the subtle qualities you can achieve processing the sound pre Amp Sim, to mimic a better sounding an instrument. Of course a better sounding instrument would be better, but I've wondered about using transient designers, Channel eq's etc before the amp sim to achieve these qualities

There are no strict rules in this area. You may find that your workflow helps you to achieve a very unique sound/tone.
Some suggestions to consider:
1) Try Tube Saturation PC module before amp-sim.
2) Use studio quality compressors, like the ones you mention after the amp sim, or between "amp" and "speaker" (this may or may not require two instances of TH2, depending on your choice of compressor).
3) Try EQ anywhere in the chain. Though I'd avoid using EQ before amp-sims, yet there are no strict rules.
 

Kamikaze
By breaking up, I mean distortion coming in.

Be careful with gain staging. For clean signals you definitely do not want to run hot anywhere, especially at the interface or the front end of TH2.

FWIW, do not rely on meters... I cannot get that "auto volume" feature in TH2 above mid yellow without clipping my interface. Be mindful of those specific stages, and compensate in the rest of the chain.

As mentioned above, FX are stupid and blindly process what they get fed. Be sure you consciously think about what is going into the next stage (can use multiple FX with SPAN inserted between each to get a feel for the chains you create).

"No set rules" is sage advice... experiment, learn, have fun.

realize that the better sims, use the natural signal from a guitar's pickups, as the 'starting point' to define the dynamics and touch sensitivity of what happens once it gets into the sim.
 
if you mess with the signal coming in, the sim will probably not react the way it was intended.
 
look into the REAMP system.
 

batsbrew
realize that the better sims, use the natural signal from a guitar's pickups, as the 'starting point' to define the dynamics and touch sensitivity of what happens once it gets into the sim.

AFAIC, only ReValver 4 has a dedicated control to set your pickup output, besides ACT profiles for different instruments and/or pickup configurations.

yep.
those machines want to control the input,
not have it driven thru some other option before hitting their controls.
 

Dave76
The modeled stomp boxes are based on compressors that were specifically designed to work with the characteristics of an electric guitar whereas the other compressors are general purpose studio compressors.  

The most common compressor pedals (MXR Dynacomp, Keeley and so on) are based on a Ross design. It's actually a high-ratio limiter with fast attack and the two knobs on the pedal set the threshold and makeup gain. I'm not sure to what extent Ross had the electric guitar in mind versus developing a circuit that would fit in a pedal and run on a 9V battery.

Unfortunately the basic Ross-derived 'two knob' designs seem to be the ones included in most amp sim packages. There are far better pedal designs out there. In my opinion anyway.

Thanks for the input guys.
 
Regards the gain structure, I'm still learning this, being I like clean tones, TH2 meter stays in the yellow, and I just make sure I don't clip at the input. Coming from physical synthesizers and drum machines, I'm inclined to get the best signal out of my guitar for noise floor reasons, and deal with gain inside sonar, than
 
I investigated PUP heights too, and adjusted the neck height. top and bottom strings were the same hieight and I'm liking the bottom strings a little further away
 
I inclined to agree with TLW, I don't see what would be so special about stomp box compressor apart from it's portability and ease of use on a stage. Which isn't the scenario here. If I want a fast compressor before a sim, it seems I have some better tools in sonar for the job.
 
Being players often do put stomp boxes before an amp, I don't see why it would care if it's from an external VST, or a purpose built stomp that comes with the modeled amp.
 
The conditioning I was thinking about would be subtle changes, in the same way different PUPs create a dsubtle difference. I'm not trying to model a PUP. But if an UltraJazz PUP gives my bass say deeper bass, and stronger attack, I can bring these qualities with a transient designer and EQ. It's like the common question "EQ before or after the compressor". Seems the same scenario here to me. If there is a resonance going on with my Guitar, why not deal with this first?  I can see how a compressor before is going to affect how the amp responds, but evening out the a touch rather than squashing it was what I had in mind. So stong nots are more consistant in how the amp is affecting them, and softer notes are bought up.
 
I guess it's just not commonly done. I thought I would get some suggestions and handy tricks. Seeing at least Soundwise is experimenting with it, it doesn't feel so out there.  I'll play with it some more I guess.
 
I noted there were no comments about bass, thought there would be more players of both

as far as compressors go,
i used to own an original grey box Ross,
and a script dynacomp,
and now i use a Barber Tone Press,
and for a good reason.