There are some technical things I feel are somewhat misrepresented:
1. Why don't AMD/Intel use a unified GPU/CPU memory setup like Apple? Because the big discrete gpus would be too big and hot to be part of the same chip as a cpu, and need faster memory than DDR4.
2. Zen3 is clocked at 5ghz: the best of them can turbo to 4.9 for a little bit, base clocks are ~3.4 to 3.8 depending on core count / model
There has been talk of using eGPUs with the M1 chips. Obviously there would be downside to having to move stuff over a slow Thunderbolt 3 connection but given that the GPU can be much bigger, how do you think that would play out?
Would that offer higher performance? If so on what kind of workloads?
1. Why don't AMD/Intel use a unified GPU/CPU memory setup like Apple? Because the big discrete gpus would be too big and hot to be part of the same chip as a cpu, and need faster memory than DDR4.
2. Zen3 is clocked at 5ghz: the best of them can turbo to 4.9 for a little bit, base clocks are ~3.4 to 3.8 depending on core count / model