[smpanaro]

For single batch inference of anything remotely LLM you'll hit the memory bound way before FLOPs, so I haven't actually looked at FLOPs much. For raw performance GPU is certainly better. ANE is more energy efficient, but you need larger batches to really benefit.

Maybe cache is the wrong word. This is a limit to how much can be mmap'd for the ANE at once. It's not too hard to hit on M1 if your model is in the GB range. Chunking the model into smaller pieces makes it more likely to "fit", but if it doesn't fit you have to unmap/remap in each forward pass which will be noticeable.

Awesome to hear about ModernBERT! Big fan of your work as well :)

[anemll]

Right.I was thinking about it, you still need batch refill, however, Apple Core ML tools were failing for attention activations quantization. Long context, pre-fill is still compute bound.