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: kash johal who wrote (37957)	10/2/1998 7:27:00 PM
From: Scumbria	Read Replies (1) \| Respond to of 1571220

With the higher clock speed, bus speeds and architecture improvements a performance improvement of 30-50% in SPEC numbers is anticipated by me. Kash, Spec should fit nicely into the K7 L1 cache, so bus speed will have little impact on the numbers. It will improve Winstone numbers somewhat however. I am very interested to see what impact K7 has on Intel's high end pricing. Scumbria

To: kash johal who wrote (37957)	10/2/1998 7:59:00 PM
From: Tenchusatsu	Read Replies (2) \| Respond to of 1571220

When it comes out it will be on the Alpha bus and there will no board chip set issues as these will have been debugged by Alpha. The Xeon disaster that Intel is having is unlikely to occur. Considering that one of the bugs that held up Xeon was a bug in the processor which only appears when the processor is used in 4-way 450NX configurations, and not a chipset bug, I think that AMD will have quite a huge job on their hands validating the processor and chipset together. Remember that K7 is still x86 architecture, and an x86 processor running on an Alpha platform is bound to reveal some nasty, obscure bugs which could crop up at the worst of times. As you know the bus speed will 200 Mhz, approx 2X Intels current best. Intel is talking about 133 for next year but no 200 Mhz that I know of. Next year, Intel's Camino chipset will handle Rambus' RDRAM technology. My understanding is that RDRAM runs at 800 MHz over a 16-bit datapath. That would translate into a 200 MHz 64-bit datapath, which can be achieved if Intel uses a 100 MHz front-side bus which double-pumps data. As for Slot A, I heard once that the point-to-point connection will run at 333 MHz, but I can't imagine why the K7 will need such an incredibly fast front-side bus. I also wonder about the multiple point-to-point processor buses and how the characteristics would affect a multiprocessor K7 system, both in terms of price and performance. In addition the CPU will be 500Mhz at least and we know that AMD is promising substantially higher performance per clock than with K6-2. I personally think that the K7 will have similar performance per clock with the K6 family, but the K7 will have deeper pipelines allowing it to run at much higher frequencies than the K6 could ever dream of. This is the advantage I see K7 having over K6, not that it can squeeze more performance out of each clock, but that it can just run faster. My prediction stems from AMD's own prediction that the K7 will reach speeds of 1 GHz by the year 2000. AMD can't get there if it sticks with the current K6 design. It needs deeper pipelining, which I think the K7 will have. Intel is architecturally bankrupt as the Xeon, PII and Celeron A devices use the same PII core. They plan on selling KNI extensions in 1999 for hundreds of dollars more than PII solutions and we will see how the market reacts when AMD starts selling higher performance CPU's than Intel much CHEAPER than Intel. Maybe that same P6 core is one of the secrets to Intel's current success. AMD would love to be able to push their K6 family into all market segments, but currently they still have a tough time breaking out of the low-end Celeron segment. Meanwhile, Intel's strategy is analogous to a car manufacturer pumping out a $10,000 econo-car, a $25,000 family car, and a $60,000 luxury car, all from the same assembly plant. As for AMD selling higher performance CPU's than Intel at lower price points, I'll let Paul answer that one. AMD's K7 is a stroke of genius. I guess we'll have to wait until the uP Forum to be sure. I sure can't wait to see what sort of architecture the K7 will feature. Whether AMD can execute well on the K7 or not, do people on this thread think that the stock price will jump a few points after the K7 announcement on mere hype alone? If so, we may have a short-term money-making opportunity here. Tenchusatsu

To: kash johal who wrote (37957)	10/4/1998 12:34:00 AM
From: Paul Engel	Read Replies (1) \| Respond to of 1571220

krash - Re: " and new kick ass architectures they will be the performance leader PERIOD. " There you go again, Krash - MORE HYPE - more KICK ASS HYPE - and no data ! You spout back rumors of clock speeds and bus speeds - but have no facts. You claim Intel has a bankrupt architecture. Well, since Intel put at least one of your companies out of business, no wonder you are so emotionally in need of trashing a company that you personally could not compete with. Trash away, kash. Keep polluting the thread with your HYPER Hype. Looks like you are really nervous about reality and are attempting to create your own. Paul