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.

Technology Stocks : CYRIX / NSM -- Ignore unavailable to you. Want to Upgrade?

To: Joe NYC who wrote (24108)	2/26/1998 6:01:00 PM
From: Scott Sterling	Read Replies (2) \| Respond to of 33344

> I don't think it was a conscious effort on the part of Intel to come > up with a design that does a lot less work per cycle than M1, M2 or > K6. They would have loved to have more efficient core. But they came > a little short, but the design team was bailed out by other groups > within Intel. That's just not true. A company such as Intel with so much to lose does not screw up in such a fashion. The chips are designed well in advance, many approaches are considered, and and of those many are developed simultaneously. Whereas an AMD might get caught at the last minute with a design that falls short such as the K5 (which was supposed to be 30% faster than equiv clocked pentium and ended up debuting at 0% faster), Intel would be making last minute tweaks to a designed already selected. Intel certainly had the option of sticking with the Pentium design and adding elements such as larger caches and better branch prediction like Cyrix did which would result in a "more efficient" core. It would be very difficult to believe such approaches were not carefully evaluated. But instead they chose a complete redesign of the chip which involved large pipelines, thus leading to not only higher Mhz but also larger penalties which lead to your incorrect perceptions on the "efficiency" of the PII (this decision was announced well in advance of the PII's release...they boasted 30% increase of Mhz given an equiv process). They didn't have to do this you know....it would actually be easier to increase the size of caches and tables. But as far as did Intel intentionally pass up a slightly superior low Mhz implementation for their high Mhz implementation.... well, I do not know which of their approaches in the end had the best performance/cost. All I know is that if I were in charge of selecting Intel's final design, and the low Mhz implementation team announced that their design had 5% better performance/cost than that of the high Mhz team, I would without blinking choose the high Mhz one. --Scott