That is the thing, AMD is currently on a multi-core/multi-threaded direction and the logical step is not to pursue single thread performance. Rather to aim for the single thread sweet spot and hit Intel with simply a much more efficient design. Now there may be prefetch and prediction enhancements but besides that; I really don't logically see anything different beyond the sharing in the front end and in the FPU. I admit I'm probably wrong but it is certainly something to ponder.




Reply With Quote

Bookmarks