Leaked Presentation Reveals AMD's Fusion Strategy

[Borealis7]

i3s are the weak lower bins, i7s are overpriced. i'd say i5 is the ONLY option for the mainstream when using discrete GPU.

 

[podspi]

Yup, which matches up quite well with an earlier comparison of the nice looking cropped die shot that AMD released to a manipulation of a Llano wafer shot. Well, except for the fact that this is nowhere near so blurry =D

Either way, going with the 226mm^2 total die area, the total GPU logic ends up being something around 88mm^2. Which is important to keep in mind when comparing performance to Sandy Bridge at its ~41mm^2. (Looking at AMD/NVIDIA products, they typically get 1.65-1.8x the performance from doubling the number of execution resources, eg 5770 -> 5870.)

Is that for HD2000 and HD3000?

 

[AtenRa]

From a reliable source, NDA for AMD Fusio Llano end's 14th of June 2011.

 

[Khato]

Is that for HD2000 and HD3000?

The ~41mm^2 figure is for both. I'm pretty sure that there are only 2 sandy bridge dies at current, the 4 core + GT2 and 2 core + GT2, which would put HD2000 as a partially disabled version of HD3000... Probably just there to offer some amount of die recovery.

But if you think about it, if all the more that's different between HD2000 and HD3000 is the number of EUs, then that would imply that something else in the design is gating the HD3000 performance on a fair number of games. Why? Because there are some games where we see the expected 70%-80% performance delta going from 6 to 12 EUs at the same clock, whereas in others it's less than 50%, dropping all the way down to 20% in a few. That something else that's gating the design could be memory bandwidth, or maybe some of the other graphics logic. We'll have a better idea of which exactly once we get to see how Llano compares, since it should have similar memory bandwidth constraints.

Edit: Forgot that I'd wanted to include the source for the sandy bridge graphics die size estimate - http://forums.anandtech.com/showpost.php?p=30459768&postcount=6