You have most of what you need to optimize for listed in the second paragraph:
- Adobe CC - lags with larger documents
- Adobe CC and Visual Studio - together with iOS/Android simulators
For these items, you'll want to fire up Activity Monitor and or the terminal to figure out what is the bottleneck. Is it RAM / virtual memory swapping to storage or is it purely CPU and thread execution waiting or is the GPU incapable of helping with some of the workload.
- Two terminal windows with
iostat running - one over a 60 second cycle and the other over a 1 second measurement window will let you know how bursty your IO is and whether you need IOPS or need bandwidth out of your storage. (My guess is you need iops - unless you are streaming / editing 4k video). Also, a third window with
top -u -s 60 can correlate the load averages you see on
iostat 60 with the CPU usage and what apps are really hitting the CPU.
Next, figure out how much RAM you really need - when you have all the apps open and the simulators - is your memory pressure in yellow or still in green. For most people in your case - money spent on more than 8 GB of RAM is usually better spent elsewhere on a faster CPU or more storage or new app upgrades or AppleCare. Even our Mac Pro / iMac 27 inch developers that run development and heavy Adobe apps are finding it hard to justify more than 16 GB of RAM when we measure actual workloads on machines that run 8 hours a shift - three shifts a day.
Once you know where your current workflow is hanging up, you can make a better judgement if you will be delighted with NVMe+PCI storage over the old and slower SATA/AHCI storage. My workload is similar to you and I have a 6 core MacPro (current model) with tons of RAM, a maxed out RAM/CPU 15 inch MacBook Pro with GPU and thunderbolt 2 ports (the last rev before the TB3 models) and a 2015 MacBook (NVMe storage and i3 / 1.1 GHz CPU and one USB C port) and I choose to do the vast majority of my work on the MacBook.
- for example, compiling mongodb from source (repeat
time brew reinstall mongodb --build-from-source two or three times) my MacPro gets the job done in 9m26s using 12 threads, the MacBook Pro takes 11m7s using 8 threads, the MacBook takes 41m25s to compile. The MacBook will run into thermal stress and slow down, but it's rare I have such a large full compile to run and incremental builds are the normal for me.
Yes the others can compile large software faster in benchmarks and Adobe runs a little faster on the large files, but I'm almost never waiting for the machines and I need to take my work with me and the hassle of lugging the 15 inch due to it's physical size and added weight (power brick, adapters, and the CPU/Display) are far more a downside than the few times I can do something a little faster. I'm totally going to upgrade to the 13 inch TouchBar model when the 2017 capital budget hits and decide if I even need to keep the MacBook.
I wouldn't look at any benchmark and instead just measure your actual workload and optimize for that. The truth is - any new MacBook with a touch bar is going to be better for you than anything Apple Sells unless you want to get a 27 inch iMac for max CPU/RAM. In practice, the NVMe storage makes your real life performance great and covers up for any measurable benchmark differences. I'm also not a fan of GPU in the MacBook line - the software glitches you suffer occasionally and the extra drain on the battery never seem to make me happy as a trade off for some limited faster filters in Photoshop or a 3D render that takes 5 minutes instead of 7 to 10.