To: Charles Gryba who wrote (71767 ) 2/17/2002 2:57:51 PM From: pgerassi Read Replies (1) | Respond to of 275872 Dear Charles: THere will be more Hammers sold in the first 3 months after introduction than all IA-64 CPUs combined. Second, given the low sell rates of IA-64, Intel can not use allocation to restrict where Yamhills go as OEMs would switch to Hammer and thumb their noses at Intel. Any allocation even would sell more Athlons and improve ASPs for AMD to boot. Third, Yamhill can not be incompatible from x86-64 because such would delay availablity by a year at least beyond where it could be done now. Fourth, even with full compatibility, Yamhill will not be available till H2 03 at the earliest since a FSB change would be necessary. Fifth, without those FSB changes, 64 bit mode has too few enhancements to justify the increase in code size required with 32 bit physical addressing and that would remove any hope of full compatability with x86-64. Seventh, with an integrated NB, motherboards with Hammer will be available much sooner than you think as many state that Hammer MBs would be sampling Q2-3/02 and in production Q3-4/02. Finally, Yamhill must be in the open at least 1 year before the target market could accept it and that would virtually end IA-64 at that point which Intel is loath to do even now. All in all, Intel has painted itself into another corner and that to jump out of it requires either a large loss in prestige, a large loss in money and power or both. Delay only makes it harder for Intel to jump (a year delay with RDRAM cost it plenty already (12-15% market share and $4-5 billion a year in profits)). In either case, AMD only needs to release Hammer for the fireworks to begin. Enough precursors of that will occur with the showing of functional silicon samples alone. And any missteps by Intel like more allocations, and AMD will be in the black this quarter and much higher ever after. Pete