With Apple now moving to ARM, I would hope that there is some behind the scenes discussion about what should be done. There seems to still be questions about whether Apple will continue to add AMD (presumably) GPUs in their more powerful units, or scale up the one on the SOC. LLR being made outside of PG also makes it a little more complicated.
The other fun comes in wondering how much performance there really is available when the high-bit vector instructions that really speed up the math aren't present. Of course, the lack of these also keeps power usage down, which is part of their advantage.
Good luck, admin team!
Eating more cheese on Thursdays.