[yunnpp]

UMAs aren't made for speed, but for power savings. You are ignoring the fact that a discrete GPU accesses VRAM and caches at much higher bandwidths (and power) than an iGPU does RAM. Shared mem also comes at the cost of keeping it coherent between CPU/GPU. So you can't just look at one part of the system and then claim that UMAs must be faster because there are no data transfers.

And by the way, even on UMAs, the iGPU can still have a dedicated segment of memory not readable by the CPU. Therefore UMA does not imply there won't be data transfers.

[alwillis]

In the case of the A18 Pro in the Neo, the memory is integrated directly into the package which is shared between the CPU, GPU and Neural Engine.

Naturally it's faster to have all of this in the same package, with memory bandwidth up to 400 GB/s.

Intel and AMD are heading in the same direction.

[yunnpp]

I had not picked up on these guys doing 400 GB/s; that's nearing discrete levels of bandwidth.