Thanks for all the info.
I'm running OS 10.10.5, so it doesn't look like this is an OS problem.
I'm having this problem on the latest 2015 15" MBP with 16GB of ram and 1TB of flash storage. There shouldn't be ANY problem running this concert. But, the same situation applies to my 2011 15" MBP. The concert runs fine with 3.1.1 and starts popping and stuttering after 25 minutes. The behavior is exactly the same on both machines, so the extra horsepower doesn't help at all, which again points to something strange going on in software.
You and I both are using a lot of NI stuff. This could be the problem. But, I did make an NI free concert for testing and still ran into problems. I should go back and check this again.
In your video, do I understand that the CPU use goes up as you use more and more patches? This seems to be the case over here. It's as if everything I use gets added to the CPU use even though that isn't what seems to be happening when I look at the expanded CPU history meter. In other words, the expanded meter shows plug ins dropping out of use when I change patches and those plug ins aren't part of the new patch. At least, most of the time that's what happens. Sometimes certain plug ins seem to get stuck. But, I haven't found anything consistent about this and I've had dropouts when the CPU use isn't particularly high, so I don't know what's gong on. What's weird is, once the dropouts and stuttering start, they won't stop until I quit and restart MS. I see an unstable CPU meter - there are spikes and choppiness - but the CPU is well below 100% even during the spikes. It makes no sense that this sort of CPU use would make the audio so degraded.
I never used MS 2 since I only started using MS seriously around the middle of last year. The example you show of the CPU going to 100% when you make a patch change has been something that happens pretty consistently for me. But, it doesn't seem to cause any trouble - at least with 3.1.1. I make the patch change and start to play and I see a brief jump to 100%, but there aren't any dropouts and the CPU immediately drops back down to where it belongs a moment later. So, I'm not completely surprised to see your CPU staying at 100% when you race through the patches. What happens when you select the patch and start playing? Does everything work normally then?
What audio interface are you using? The problems occur for me even when I'm just using the internal audio, but I'm normally using a MOTU interface live. I'm not sure whether an audio driver can cause interference if the interface isn't hooked up (I doubt it), but I thought it was worth mentioning in case you're using one too.