To: Ali Chen who wrote (40613 ) 11/1/1998 10:22:00 PM From: Elmer Read Replies (1) | Respond to of 1573682
Re: "EP, <K7 is stuck with very high, expensive pincount,> Could you enlighten us exactly how high is the K7 pinout? " I don't know what it is, they didn't say, but I can tell you this, it's at least 64 + address&control lines or 128 + A&C lines more than it should be if it had the L2 on die. That's either very expensive or VERY VERY expensive. Re: "Of course, it is highly economical to build single-CPU systems with a highly-redundant SMP-capable (and therefore slower) "universal" chipset, very true :[ " You keep missing the point. For Intel, it isn't the chipset but rather the bus that makes it all possible. Re: "<The Intel P6 bus has none of these problems.> There were many other busses without problems except one - they ran at 1-10MHz..." If I knew what you were talking about I could respond :-( Re: "Wow, what a revelation! You must be in strategic marketing, bud. You only forgot who has a better technology to built dense caches for smaller silicon footprint..." That IS the question, isn't it. If AMD doesn't have a problem with LI then why did they cause themselves all the problems with a BSB? Why did they give up the performance and increase the pincount if they really can yield large SRAM arrays? Why did they avoid the on die L2 and instead create a whole host of other problems instead? If AMD really has a clean process then why be so shy with the die size when the competition is already doing huge L2? If they can't run with the big dogs they should stay out of the hunt. Re: "what else can a screwdriver do against highly acquainted individuals like youself? " Ali, my modesty prevents me from commenting further..... EP