Yeah, I’ll wait and see what 15% means, but on the face of it, that’s a little disappointing. 15% on geekbench of M2 takes it to around 2900 which is hmmmm. Are we sure these aren’t A16?So 15% faster P-cores and 30% faster E-cores. Interesting. Was hoping for closer to 30% faster P-cores, but 15% isn't bad by any stretch.
It was scary fastWell... that was short.
I doubt it, but they didn't say the core names - but that's just like the A17. It'd be weird to have the same A17 GPU cores and not CPU cores on the same process as A17.Yeah, I’ll wait and see what 15% means, but on the face of it, that’s a little disappointing. 15% on geekbench of M2 takes it to around 2900 which is hmmmm. Are we sure these aren’t A16?
Odd, M3 Pro chip is not a Max chip cut-off this time
I’m sure you’re correct, but then I’m struggling to think why single core performance might only be 15%. Time will tell I suppose. Their measurements don’t always align with the kind of benchmarks we discuss here.I doubt it, but they didn't say the core names - but that's just like the A17. It'd be weird to have the same A17 GPU cores and not CPU cores on the same process as A17.
It never was...?
No, it really was in previous generations. You look at the die shot of M1 Pro or M2 Pro, up next to the corresponding Max die, it looks exactly like they just cut off half the GPU core set.
Not literally cut down, but designed so with minimal changes you could create a SoC similar to the Max but with less GPU cores, sharing most of the floor planning.I really do not think that is correct, the M1/M2 Pro were always their own die, not cut-down M1/M2 Max dies; and that would be a horrendous waste of wafer space...?
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.