Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
2H06 for the start of the 65nm transition, although this will probably start with server & mobile (IMO), so if Anand is only referring to desktop (this is in the AM2 section), Q107 might be correct.
Looks like your track record is about to go down the tubes.
Says someone standing in the sewer looking up...
That's pretty weak reasoning. The model number isn't high enough, so it won't offer better performance???
I think I'll take the Inq's multiple sources averaging to better than 10% clock/clock on the desktop.
How nice for you.
And Hector today saying that worldwide, AMD has 40-50% of retail share.
Your "data" is such a small snapshot of a piece of the market that it doesn't speak to much of anything. I'll take the Inq plus the CEO saying they will be close to capacity this entire year over your data.
Keith, it was a Q&A roundtable led by a moderator, not a "presentation". A hype-filled marketing spiel is not what this audience wants.
You might want to glance at the larger markets today before ascribing any move in AMD's stock price to a lack of a glitzy, sales-y performance by Hector.
My predictions have a rather better history than yours.
Yup. You got it. Again, they could do it whenever they need to. If Woodcrest is as late as Keith thinks (late Q3 / early Q4), they might not need it.
It was a PREDICTION.
Repeated claims that AMD is sold out for Q1 from the Inq.
Keith, #2 is not "should", but "will". He was asked to reiterate, and he did, that Fab36 will ship for revenue in this Q.
Also: Fab36 yields mature, ramp is aggressive.
Regarding #4, also: expect to be done with conversion to 65nm by summer 2007.
Regarding #6, he did also say: "AMD sees nothing of concern regarding market demand."
AMD is sold out. OEMs get priority over the channel.
Core Duo rebates, too? I guess 32-bit may not be selling...
I don't see any Woodcrest spec submissions in your links. :)
Well, let's see:
May is 2 months away. Intel's NGA is 4-8 months away.
The degree of hype is different.
This is more of a new offering, not really a replacement, like Intel NGA.
I'm afraid you're going to be wrong YET AGAIN. :)
You realize Intel's compiler does this, right?
Woodcrest is supposed to score 2800/2500 in SPEC_CPU2k
Two words: Parallelizing compiler.
It hasn't swung-on-a-dime for the first 2/3rd's of Q1...
Not gonna happen, unless AMD splits 3-for-1.
wbmw must be shocked. After all, this is a 90nm part, with a 95W TDP. Wait until the Rev F 2.8GHz 90nm DC parts arrive, with the additional ~10% clock/clock performance improvement... implying the performance equivalent of a 3.1-3.2GHz Rev E DC part.
Yes, but the question is-- an Intel demand shortage, or an overall demand shortage.
With Intel being a LARGE percentage of the market, it can be hard to tell until the numbers come in.
I think Intel lacks competitive products right now, PLUS, they are osbourning their current products with talk of Q3/Q4 products, PLUS the market may be typically seasonal this Q1 vs. last year, where it wasn't down as much in Q1 as normal, IIRC.
Also, I think Intel is losing their premium pricing ability, now that AMD is a perfectly acceptable (and for a growing number, preferred) alternative. So Intel can no longer command premium ASPs... so Intel's ASPs can fall quite a bit, even as AMD's could remain stable, as Intel was starting out with a 50% premium.
Hey wbmw, looks like once again it was *you* having trouble understanding the concept. :)
-----------------------
Re: Sounds like Intel is headed for the low end of Q1 guidance.
Sounds like you're having trouble understanding the concept here. Let me help:
Intel is getting grey market activity on their 800 series Pentium D, but their 900 series is selling at a premium to list price. That can only mean that vendors are trying to offload their 800 series chips for 900 series replacements. Since Intel has already sold the 800 chips at full price, they are not losing out from the repurchasing. And if anything, it's a sign that vendors are making room for the new products.
But if you want to interpret that as a sign that they will underperform their Q1 guidance, let me know so that I we can revisit your prediction later.
Sure ya are. Your dreams of INTC at $37 must be dashed by now, right?
Imagine what Intel would say about it, for one thing.
"Partially broken AMD server processors. Do you trust your datacenter to salvaged CPUs?"
They don't have the issue as long as they use two DC's on a cracker.
Uh huh. I suppose you can't say who that was, right? :)
. People haven't had any qualms about buying chips with lower levels of cache, which have often been chips with parts of larger cache disabled due to defects.
And how many of those people do you think REALIZE that's what they're buying?
The TC problem is that *all* of them would be QC parts with one core defective/disabled, and so they are all tarred with the "salvaged" label.
Do you have any facts from AMD about their *selling* a TC part at any point in the future? If not, stop assuming they think remotely like you. And you still could do with easing up on the exclamation points, and the "I am the (only) seeker of the truth" routine.
I'm using X3 to mean *any* tri-core part-- perhaps not the best notation. I'll switch to 'TC'.
I said that CTOs don't make marketing decisions, because they aren't very good at it.
You have absolutely no facts or information on which to base that on.
You must be thinking of your unscheduled, unannounced, and probably imaginary "X3"...
How exactly was that "speaking for AMD marketing"? I think you need a chill pill.
Jules, no that was using 1 socket out of 4 for a special processor, linked by cHT.
I don't believe we'll see AMD selling any X3s, but if you can show me one for sale in the future, I'll be happy to admit I was wrong.
CTOs don't make those sorts of decisions. :)
The fact that you can make a legitimate, logical, technical argument for why a partially disabled QC part is perfectly acceptable is insufficient. (And this is one reason that engineers tend to suck at marketing.)
Yeah, the ex-CTO. CTO's aren't generally known for their marketing prowess. :)
From what I've seen, they plan to do exactly that. Replacing what might be the fourth core with a more complicated cache structure.
Huh? Where have you seen this?
Doesn't work that way. Every 3-core part will be a defective 4-core part, and that word *will* get out. It's when you have some of each (meaning 3-core dies working fully + 4-core dies partially disabled/defective) that you can get away with it.
Finally, the X3 boosters might note that AMD has made no mention at all of such a part on any roadmap.