Interested to hear your guys thoughts (based on the extremely limited information) that Apple shared today ?
If benchmarks back up (which I have no doubt they will) Apples numbers (which in recent years have tended to sandbag just how good improvements have been), then this is a nice step up.
I’m particularly interested to learn how they got a 2X ray tracing performance improvement - new GPU architecture? They claim a bandwidth improvment of 20% but that wouldn’t alone account for it.
The other area that I’m interested to learn more about is the direct programmable ML accelerators…
The next M4 Pro/Max/Bodacious chips could be shaping up to be beastly.
If benchmarks back up (which I have no doubt they will) Apples numbers (which in recent years have tended to sandbag just how good improvements have been), then this is a nice step up.
I’m particularly interested to learn how they got a 2X ray tracing performance improvement - new GPU architecture? They claim a bandwidth improvment of 20% but that wouldn’t alone account for it.
The other area that I’m interested to learn more about is the direct programmable ML accelerators…
The next M4 Pro/Max/Bodacious chips could be shaping up to be beastly.
Attachments
Last edited: