SI
SI
discoversearch

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 : Advanced Micro Devices - Moderated (AMD) -- Ignore unavailable to you. Want to Upgrade?


To: kash johal who wrote (15885)10/25/2000 12:17:45 PM
From: PetzRead Replies (4) | Respond to of 275872
 
kash, Gateway considers P4 a "niche product." So, I repeat my question, How many of the top 10 OEM's will have a P4 machine available before 1/1/2001?

1. Intel is gonna stuff half their parts into a dustbin so they can meet a 1.4Ghz launch.

Doesn't seem very credible to me.


With a normal bell shaped speed distribution curve, there should be at least a 20% spread between the bottom and top speed bins. We have 7% here.

Either they are not selling the 1.6 and faster P4's or there aren't any at 1.6 or faster. If you think they exist, but Intel is not selling them, ask yourself this question:
When in the last year did Intel decide NOT to sell at the highest MHz possible?

If there aren't any 1.6's and faster, then there definitely are some 1.2's and 1.3's which Intel is throwing out. No CPU line produces CPU's over only a 7% speed range.

In other words, the very narrow speed distribution is proof that there are lower MHz P4's which will not be sold.

Petz



To: kash johal who wrote (15885)10/25/2000 12:21:31 PM
From: AK2004Read Replies (1) | Respond to of 275872
 
Kash
I am sorta amdroid and I never claimed any of the points but I think that paper launch is not out of the question (100s of K and not mils).
Bell shaped distribution with low at 1.4GHz and std 25% implies sweet spot at around 1.4*1.3 = ~ 1.8Ghz. If 25% is 2*std then the implied sweet spot is ~1.6Ghz.
So the question is if the sweet spot at at least 1.6GHz why intel is not releasing anything above 1.5Ghz yet.
Regards
-Albert



To: kash johal who wrote (15885)10/25/2000 12:32:28 PM
From: dougSF30Respond to of 275872
 
Kash, recall that the Register also claimed 1.7 GHz parts from the Israeli fab would arrive in Q3 2000; presumably this was from some 'roadmap' or other, as well.

Now they are saying Q1 2001.

Just something to keep in mind,

Doug

p.s. regarding Intel's Rambus 'daze': Oh, I think they're in a bind [-ing contract] for another couple years, even if they have finally faced up to reality.



To: kash johal who wrote (15885)10/25/2000 12:53:42 PM
From: hmalyRespond to of 275872
 
Kash Re..<<<1. Intel is gonna stuff half their parts into a dustbin so they can meet a 1.4Ghz launch.

Doesn't seem very credible to me.<<<


Kash, with a normal launch, it probably wouldn't happen, but with a shortage of Rambus, it just might make perfect sense. Why waste Rambus supplies on chips which make you less money and could have problems matching up to a T-bird. In addition P4 + Rambus will have serious problems on a price performance basis. So why not save the Rambus for higher mhz models which command a premium in price anyway. As you have explained, chips normally have a 30% variation in yield. P4 should yield between 1.15 -1.5 if 1.5 is high end of curve, and 1.4- 1.75 if P4 is the low end of curve. Because of afformentioned rambus problems, I would chose the first scenario, as opposed the second, because I believe if Intel could go faster, they would.

3. Mhz doesn't matter anyway.<

From what I gather, we are saying mhz isn't the only thing that matters. ; IPC does also. In addition, Intel used to have 100% of the top end sales. P4 should increase Intel's percentage , or at least stop the slide, but to recapture 100% again would be highly unlikely. To even get back to 80% of high end x86 pc sales would be difficult.



To: kash johal who wrote (15885)10/25/2000 2:15:40 PM
From: Gopher BrokeRespond to of 275872
 
Now they are gonna start at 1.4Ghz and are publishing roadmaps 1.7Ghz for Q1 and 2Ghz for Q2 2001.
What this CLEARLY means is that sweetspot is NORTH of 1.4Ghz and some parts will yield well up to 1.7Ghz.


What would be the effect of shipping a 1.3 GHz P4 that ran slower than a 1 GHz P3 on most benchmarks? Intel has no choice about the low-end speed they launch at.

What I find significant is that with a low end speed of 1.4 they can only achieve a high-end speed of 1.5. If 1.4 was truly the low end wouldn't you expect there to be 1.6 as well? And no, I don't think that Willy will be so well received that they can justify downbinning high-end chips to have something in reserve.