My first computer i personally owned was a base model 13 inch 2015 macbook pro. It was a 5th gen 2.7ghz dual core config with 8gb of ram and a 128gb ssd. I simply couldnt afford any upcharges: it was a stretch even purchasing the computer, at $1000. What I found from that experience is that a dedicated GPU, a large internal hard drive, and at least 16gb of ram are necessary baselines for any computer. I know that, for me, the highest spec GPU option will be always be necessary for my main computing device. And with Apple, the highest end gpu option usually isnt even that impressive. The radeon 580 is a fine card, but in the gaming pc market, the 580 is widely considered a fairly low-midrange budget card. It will play games at 1080p pretty well, but it’s by no means a powerhouse. The 580 was a no-brainer for me, because I’ve already found its limitations. With the iMac, RAM is no problem. Get the 8gb model and add in a couple 16gb dimms—don’t get 8gb sticks, because if you ever want to go up to 64gb of ram, you will have to swap out every dimm in the computer. CPU intensive workloads don’t seem to ever become noticeably faster with higher end CPUs, as weird as that sounds. 8-cores is a pretty big deal, but when manipulating things in real-time I think there are simply limitations to UI and code bases that make higher end CPUs usually not any noticeably faster than their lower end counterparts. Usually a lower end CPU will just slow you down, not completely impede certain workflows, although this depends of course on how low end you go and what kind of software you use. I think configuring a 512gb internal ssd sku with vega 48 and the i5 6-core with 8gb of ram (through the education store) would be the decision I would probably make these days. However, we dont really know how much faster the vega is quite yet, so maybe I would wait until someone like Max Yuryev does a comparative benchmark
