My preferred setup would be i9, 32GB ramp, Vega 20, like you, but bump to 2TB drive for the sake of future-proofing & long term life...I mean I'm on a 2011 still. Anyhow, I'm concerned about the heat that the i9 & Vega 20 would put off in the keyboard, palm rest, & the bottom, so I may go with the i7 Vega 20. I'll be waiting to see how your's turns out. I need to dig up the software I like to test with & send it your ways.
https://benchmark.unigine.com/heaven
Predictable does not give numeric info to actually compare. That's why I asked for the tests. The combo of both will show the worst case scenario of stress loading the MB. It also makes comparing against my current hardware because I can replicate the same conditions...100% CPU with high GPU load. Games go all over the place depending on the scene & what level of CPU is being used.
I still have the 2016 MBP with the 2.9GHz quad core – that one can maintain its base frequency and also maintain clocks on the 460 gpu, which uses 35 watts. Temperatures are in the low 70s range during heavy gaming sessions. Isnt the Vega 20 supposed to come with a 35 watts package as well?
Yes, it is. But the CPU easily exceeds the 45W, thanks to less strict Power Limit States.The question is, what would be considered uncomfortable temperatures
You can test it out: set a custom lower TDP power limit and run Cinebench etc. until your clocks settle at 2.6Ghz. Since i9 is higher binned SKUs, there is a high probability that its more power-efficient than a i7 running at the same clock. But again, this will depend on two individual CPUs. Put differently: if you test 1000 i7 vs 1000 i9, I am confident that i9 models will be more power efficient on average. But some of the i7 will outperform some of the i9s in this regard.
