As I read I mused as follows...

Aside from massive core count MC is simply the wrong chip on the wrong infrastructure for OC. If you are curious about what you would achieve, go OC a Phenom II X6 and know that those setting are better than you would ever get for MC. The cost to get an enthusiast-class MC infrastructure on par with, say, a Crossfire 4 are basically insane. An MC customer is already getting great consolidation and performance, and adding a second box or second processor is astonishingly affordable and a lot safer than OC.

A dual socket Phenom II board would be much more interesting. Easier to achieve stable results and potentially finding a home in the visualization space (and a couple others). It's a completely different world when Quadfather walked the earth and I think such a board would be interesting. It's not going to happen but it would be interesting.

The fundamental problem in client is multi-core scaling and right now there are a few application categories that love cores, and with most other categories completely saturated, GPU bound, unable to figure out what to do with more. The ultimate use for the PC in entertainment - the Holodeck (and it's predecessors, more on this soon) - needs CPU and GPU balanced performance. To achieve this in a reasonable TDP does not involve a dual headed monster, rather, a powerful CPU and a powerful GPU - those create plenty of noise and can pack plenty of cyberpower, thank you very much.

In server land there is infinite need for consolidated cloud compute enabled by virtualization. For mobility, there is infinite pull right now for sub-10W compute performance. So many reasons why I would love to see an exciting low power platform - I want smartphone DIY.

And I want a holodeck.