Can someone with extensive knowledge is GPU explain what would keep apple to put the 765m in the rMBP ? Also I'm scared of apple not even bothering to use a dedicated GPU because of iris/GT3e wich would make the laptop pretty useless to me any chance of that happening in the 15 inch model ?
How would iris pro make the laptop useless? It is not that far away from a current 650M. Is that one also almost useless in your opinion?
Also Intel's GPU beats Kepler in GPGPU stuff and OpenCL. Because Kepler unlike Fermi (or AMD's GCN) was severely handicapped in that department.
Also one wouldn't have to deal with Apple's GPU switching which switches an unnecessary GPU on way to often and keeps it on even if the App that might use it is currently only in the background doing nothing. The real life battery improvement in OSX compared to the current 15" would be big.
For all OSX needs the Iris Pro is probably a better GPU than the 650M because with the right driver it is more geared to those uses rather than DirectX gaming as the Nvidia Kepler is all focused on. Kepler is great for gaming but in order to make a very efficient GPU Nvidia cut away almost all the stuff that wasn't needed for gaming.
In any case I doubt it is going to happen. OEM's say a non Iris Pro CPU + 750M is cheaper and well it is faster too. Iris Pro while surprisingly close in games and all practical metrics is still slower than a 650M @ 900Mhz and as this thread shows Intel's reputation needs some more time. Apple might not do it just because of the bad press they would get.
In truth though all that a swtich would really hurt is those people that want to game in bootcamp. It is more a marketing thing then an Apple thing. If Apple does as it used to under Jobs and care about OSX first and objective value, they will probably do as Anand expects and deprecate the dedicated GPU.
The reason why not 765M is simply. Because Apple doesn't go for maximum performance. They never did. They don't want to redesign anything about the rMBP just switch the logic board. A 765M can work in small design because the 700M turbo 2.0 starts out at a rather low clock speed and does what the thermals allow. So you can more easily fit them into smaller designs.
If anything they would more likely use a 760M. I wouldn't rule that out but who knows.
750M + HD 4600 Quad is the most likely.
760M + Quad is less likely.
an HD 5200 Quad would make a lot of sense as an option but from what I read the price might be a problem for that. I should check what a 4702HQ or the lower end HQ models cost. A 4950HQ is definitely to expensive but that is how it always is the XM Quads were for idiots with too much money.
The thing Nvidia has going foritself is that those GPUs don't cost much and most designs are geared toward it anyway. I assume most OEMs didn't really expect Intel to do as good as it did. I suspect that HD 5200 only notebooks are more likely to turn up 6 months from now. It is also a brand thing. Except for some Anandtech readers most people don't see Intel GPUs as favourable and on the package one can claim equal battery life with a nvidia GPU as that one will be off due to Optimus for all the battery life tests.
A Razer 14" does not get 6h with the Nvidia GPU active.