Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Politics : Formerly About Advanced Micro Devices -- Ignore unavailable to you. Want to Upgrade?

To: Elmer who wrote (33182)	6/8/1998 1:06:00 PM
From: Steve Porter	Read Replies (1) \| Respond to of 1573891

Elmer: Slot1/2/? can also scale very well with the GTL+ specification. There is no reason why the data path can't go from 64 bit to 128bit (or more) other than added cost. Not a problem on a highend Merced system. Frequencies should go to 133mhz and then 266mhz without a basic change in the bus definition (other than timing plus double pumped clock?). This could raise bandwidth dramaticly with or without increasing pincount significantly. This then raises questions about the value of large L2 in a single processor system. Generally I agree.. while there are implemntation details to be ironed out in widening the buses, it is nothing that I would consider show stopping by any stretch. Question for you: Does SlotA use some kind of a split transaction protocal? I assume it must. I believe that I read that it does use a split transaction protocal. However I don't have the documents in front of me now to dig up the details. I'll see if I can find my papers in the office tomorrow (I decided to take a day off and relax ;-) ). I don't really see any big difference between the two buses in terms of potential. Any differences that exist now can be quickly fixed/removed by widening the Intel bus out to 128bits. I do seem to recall that the EV6 functions at 333Mhz on the memory controller side of things, but 266 to 333 isn't _too_ big of a jump. Ofcourse when you are talking about 333Mhz memory controlers and support chips you are adding cost becuase you have more defects (in terms of making the damned things.. I mean 333Mhz is a top of the line CPU clock speed these days ;-) ). So cost goes up. The same with 266. Timings are a lot tighter than at 66 or 100Mhz. All in all everyone will eventually be in the same place, even though that may take differnt roads to get there. I think Intel could do themselves a HUGE favor by publishing the P6 bus specs in their entirety. Let AMD, CYRX, IDTI all make Slot 1 and Slot 2 chips if they want. I mean it's not like Intel will loose anything. One the low-end Slot 1 can't compete price wise with socket 7. However if everyone was on Slot 1 Intel may pick up a couple more $$ in upgrade business. Steve