Originally Posted by soulsaver_8229
that would be very interesting if they did this and it worked, however how large the proc would be, unless they can get lower than 65nm and get the gpu smaller than that, man thatd be amazing
Well, from what I've heard, integrating the GPU right on the processor die would cut the transistor "budget" by a third for the GPU part. The lowered latencies would also allow for more performance for less transistors.
I get the impression this is not going to be high-end GPUs though, some of them have higher transistor counts than dual-core CPUs. If anything, it'll probably start off with the low end and may move into the mid-range. It certainly will be interesting to watch.