It's the same old story: Chipzilla shows a new fastest-ever CPU to a bunch of analysts or developers and a few weeks later, AMD cobbles together something using 20 gallons of liquid nitrogen and 5Kw of fans (and sacrifices a couple of virgins to be on the safe side) in a bid to match Intel’s performance. This fixation with raw MHz is as meaningless as the MIP rating beloved of old mainframers, yet both Chipzilla and the Great Satan of Taperecorders persist in playing the "mine’s bigger than yours" game. So, rather predictably, Jerry Sander’s swansong at the AMD annual shareholders' meeting yesterday was a demonstration of the KryoTech Super-G computer running at 1 GHz. KryoTech expects to deliver the Super-G to customers at speeds up to and including 1 GHz. Not above 1 GHz, you’ll notice. Now, Intel is targeting Coppermine to launch in September at 600MHz, but early reports from inside the chip behemoth indicate that yields of up to 733MHz are easily achievable, right now. So we have a situation where Intel has a choice: does it launch Coppermine at 600MHz and then phase in the faster 667MHz and 733MHz parts, or should it go for 733MHz from September? Going for 733MHz from day one would put Intel so far ahead it would leave AMD in a very tricky position in matching Chipzilla for raw speed – so tricky it just might be the final straw that breaks the AMD camel’s back. Your call, Chipzilla… ®