I think its more a PR thing. No idiot is gonna say "we're unsure of 65nm yields" to already shaky investors. Right now, its more a crapshoot, because I believe AMD has not begun sampling yet. Traditionally, AMD is a full year behind Intel in process technology.
I think its more a PR thing. No idiot is gonna say "we're unsure of 65nm yields" to already shaky investors. Right now, its more a crapshoot, because I believe AMD has not begun sampling yet. Traditionally, AMD is a full year behind Intel in process technology.
No, but I think upper management is supposed to be held accountable for what they say at conference calls, they can't blatantly lie to their investors can they?
I am aware of AMD and Intel's timelines on process changes. But I am not discussing that at the current time.
Not to mention as well, that processors aren't Intel's only sources of revenue.
Remember, 8 Billion accounts, for Core Duo, Pentium M, Pentium 4/Pentium D, Celeron D, Itanium 2, Xeon MP, NIC, chipsets etc etc...
Something interesting I picked up from another forum:
A Plea Not To Stampede . . .
Ed Stroligo - 7/16/06
Here's a few numbers that might quell the savage beast in you if you feel panicked about getting a Conroe in the next couple months.
This quarter, Intel will make around four million of them. That's a little less than a third of AMD's total current production.
Christmas quarter, Intel will make something close to fifteen million of them. That will be around all of AMD's total production.
The quarter after that, the number of Conroes built will exceed twenty million.
These aren't going to exactly be rare collectibles. This is not like socket 939 Opterons, where maybe a hundred thousand were made. Paying a premium for these chips is more like paying a premium to eat one of the first new meals at Mickie D's.
True, we'll have to share with all the big OEMs, and they might want so many in the beginning as to lead to spot shortages in the retail market, but any shortages will only be around for a short period of time.
But millions of these things are going to be made, unlike, say, socket 939 Opterons. There's no reason to panic, or pay a ton for the first of what eventually will be one of a couple hundred million of them.
A Plea Not To Stampede . . .
Ed Stroligo - 7/16/06
Here's a few numbers that might quell the savage beast in you if you feel panicked about getting a Conroe in the next couple months.
This quarter, Intel will make around four million of them. That's a little less than a third of AMD's total current production.
Christmas quarter, Intel will make something close to fifteen million of them. That will be around all of AMD's total production.
The quarter after that, the number of Conroes built will exceed twenty million.
These aren't going to exactly be rare collectibles. This is not like socket 939 Opterons, where maybe a hundred thousand were made. Paying a premium for these chips is more like paying a premium to eat one of the first new meals at Mickie D's.
True, we'll have to share with all the big OEMs, and they might want so many in the beginning as to lead to spot shortages in the retail market, but any shortages will only be around for a short period of time.
But millions of these things are going to be made, unlike, say, socket 939 Opterons. There's no reason to panic, or pay a ton for the first of what eventually will be one of a couple hundred million of them.
DrMrLordX
Lifer
- Apr 27, 2000
- 22,894
- 12,953
- 136
33:36
Start there. Fab 36 ahead of schedule, 65nm production in the second half of 2006 at mature yields. Aggressive transition to 45nm production within 18 months of the 65nm transition.
Derek Myer is the speaker, I think? He's introduced at 31:32
Keep in mind that my player starts at 29:09 for some odd reason, so if yours starts at 0, adjust times as necessary.
Thank you, but I actually found the actual chat transcripts so I don't have to listen to these, to find it, once I found the transcripts it was not too hard to find the time index.
I just wanted to know if the word "fully" was there and it isn't, so this doesn't tell me anything new because I already remember from Conference Call in Q1 that they said they will be introduceing 65nm at mature yeilds.
Intel Q2 2006 Conference Call Chat Transcript
http://seekingalpha.com/article/13895
AMD Q2 2006 Conference Call Chat Transcript
http://seekingalpha.com/article/14001
From what I have listend so afar AMD is basically on schedule as I knoew AMD was aiming for a faster then normal aka "less then 2 years" transition to 45nm, but by the looks of things, Intel is looking like it will achieve less then 2 years transition to 45nm as well.
1. to Coldpower...you were absolutely correct about "fully" not being in the CC (apologies). Unfortunately I had read a "transcript" with the word in there, and it stuck in my mind...thanks for the correction, and I have corrected the poster that wrote the "transcript".
2. Henny
Intel (because of their large size) use "Copy Everything", which is more efficient when you have a large number of Fabs to convert. In essence, they tweak their process in a single Fab first (I believe it's currently D1D or F12), then exactly copy the process to other Fabs (in this case F24 and D1C which they are currently doing).
The net result of these differences is that AMD is much quicker to produce mature yields, but Intel is much better at very large volume mature yields.
3. keys
4. PCSurgeon
Remember also that 4MU is the total number (possibly) by the end of Q3 (July, Aug, and Sept combined) and that it includes ALL C2D chips...that said, bringing 15-16 MU online by Q1 07 is a screamingly fast ramp!
2. Henny
AMD has been going through the learning curve on 65nm for some time now. They began turning out production parts on a small scale in October last year. The biggest reason for the delay is that they can't afford to release to volume production (like Intel can) without mature yields as their Fab space is more limited. They also have an advantage over Intel in that their APM system allows MUCH greater granularity in adjusting manufacturing...they are able to adjust and analyze dynamically down to the individual die during any stage of production. The system is completely automated and makes corrections 24/7.
Intel (because of their large size) use "Copy Everything", which is more efficient when you have a large number of Fabs to convert. In essence, they tweak their process in a single Fab first (I believe it's currently D1D or F12), then exactly copy the process to other Fabs (in this case F24 and D1C which they are currently doing).
The net result of these differences is that AMD is much quicker to produce mature yields, but Intel is much better at very large volume mature yields.
3. keys
Actually, if they were expecting less than mature yields I would expect them to say nothing at all and avoid any questions about it...why would you open yourself up to a lawsuit on purpose??
4. PCSurgeon
I hate to dispute "Often-Wrong" Stroglio, but Intel themselves have said 10% of mid-high end desktop in Q3 and 20% in Q4...going from 4MU to 15MU would seem to disagree with Intel's own numbers. The 15MU number is close to what Intel is projecting for Q1 07.
Remember also that 4MU is the total number (possibly) by the end of Q3 (July, Aug, and Sept combined) and that it includes ALL C2D chips...that said, bringing 15-16 MU online by Q1 07 is a screamingly fast ramp!
Here guys, what do you think about this transition data?
http://www.hkepc.com/bbs/attachments_dir/ext_jpg/A202767TH-3_ipPkKuUjjzf6.jpg
http://www.hkepc.com/bbs/attachments_dir/ext_jpg/A202767TH-3_ipPkKuUjjzf6.jpg
Wow! That is VASTLY different than all the previous guidance! What this shows is that C2D will be 10% of ALL desktop rather than the mid-high-end desktop they previously guided for...where did you get that and do you have an idea as to how accurate it might be?
From the site listed in the link of course, and it is newer information then the roadmaps posted by Dailytech, so I have no reason to misbelieve it, until proven otherwise.
Let me rephrase then...since the link is to a BBS and not the main site, did you find an article about this as well?
It comes in the article talking about the E4300. Which isn't too surprising, since that would help accelerate the Core 2 Duo transition.
What is most interesting though is that Celeron D isn't really Intel's largest % sold processor, both the Pentium 4 and Pentium D outweigh it. This isn't surprising to me as most buisness systems even though they may not "need" a Pentium 4 do keep buying those processors. because they are Intel's flagship brand, or once was.
Sorry to be a pain, but I can't find the article...could you post a link for that as well (I'd like to read it...)
It looks to me, after seeing that. That OEM's during Q3 will be offering C2D for the high end. Thats my guess anyways. 10% may be alot, but far and away from a full scale desktop blitz. This to me shows what I've been trying to say all along. There will be plenty, and I mean plenty of C2D's to go around. The only worry I have, is people rushing out and buying it at whatever price to satisfy thier e-penis upgrade dreams....
code65536
Golden Member
- Mar 7, 2006
- 1,006
- 0
- 76
Yes, I know. Even being conservative and estimating that the desktop segment accounts for only a quarter of revenue, this still means that this conroe stock will represent only 10% of their their stock in dollar terms. In quantity, this number would be lower since Conroe chips are priced higher than older generation chips. Of course, this is certainly in step with Intel's roadmap projects where a mere quarter of chips will be Conroe by year's end, which is disappointing, as it means Conroes will carry a premium price for some time.
All that I'm saying is that this $200 million inventory stock, in the greater scheme of things, is not that big and is more or less directed at people who say "that's a lot of chips". Yes, it's a lot, from the perspective of an individual or a retailer, but it is a relatively low number considering that these chips are going to be dispersed to all corners of a big world market...
Originally posted by: code65536
Yes, I know. Even being conservative and estimating that the desktop segment accounts for only a quarter of revenue, this still means that this conroe stock will represent only 10% of their their stock in dollar terms. In quantity, this number would be lower since Conroe chips are priced higher than older generation chips. Of course, this is certainly in step with Intel's roadmap projects where a mere quarter of chips will be Conroe by year's end, which is disappointing, as it means Conroes will carry a premium price for some time.
All that I'm saying is that this $200 million inventory stock, in the greater scheme of things, is not that big and is more or less directed at people who say "that's a lot of chips". Yes, it's a lot, from the perspective of an individual or a retailer, but it is a relatively low number considering that these chips are going to be dispersed to all corners of a big world market...
After all that...there will STILL be plenty to go around
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes + WCL Discussion Threads
- Started by Tigerick
- Replies: 22K
-
News NVIDIA and Intel to Develop AI Infrastructure and Personal Computing Products- Started by poke01
- Replies: 366
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
Facebook
Twitter