Elmer, >Daniel I am simply baffled by Intel's RMBS strategy and I have no clue as to why they are continuing to support it. I don't have any answers on this one. Just the same questions everybody else has.
Maybe, as Scumbria is beginning to allude to, the high bandwidth at low pincount gets significant at higher (than 1 GHz) clock speed CPUs. Maybe Intel was too early committing so much to Rambus, or they felt a need to get cozy before anyone else? Nah, on the latter, they were probably just too early.
I'd like to see, as you said on the Intel thread (or was that me), some Intel System Performance Architecture predictions on performance with RDRAM, as a function of clock frequency, memory and cache sizes, I/O data rates, etc., with the various benchmarks. They do SPA predictions, don't they?
One of the great mysteries of the 90's , 00's here.
Tony |
