3 ms·
The M7 Max and M7 Ultra will likely prefill-bottlenecked at 100GB+ scale inference. Layered 6090s would not be.
by bigyabai 4mo ago
The M7 Max and M7 Ultra will likely prefill-bottlenecked at 100GB+ scale inference. Layered 6090s would not be.
- aurareturn 4mo agoNeural Accelerators in M5 are already 4x faster than M4 at prefill. With M7, especially if they focus on AI like this article claims, it likely will have excellent prefill compute.