Even if you have never really been interested in Intel’s financial reports, you may still have got the feeling that the last few quarters haven’t been the company’s best days. The preparation for the shift to new micro-architecture that has been started very unexpectedly, causes Intel some tangible losses in this transitional period of time. According to the company’s primary spokespeople, Intel hurries to finally switch to the new micro-architecture, and hence they are revising dramatically the launch schedules for the upcoming CPUs in three major market segments.
The Inquirer as well as a few other sources reported some news from Intel’s recent meeting with the company shareholders. First of all, Intel is planning on another restructuring of the company: the losses keep growing and this cannot go on like that any longer. To regain the lost market share, Intel is going to speed up the arrival of the new Core platform. Therefore they rescheduled the launch of the new server Woodcrest processors for the end of June, moved the desktop Conroe CPUs launch to July, and mobile Merom processors will come out in August. As you know, all these three processor families have similar architecture.
Our colleagues from x86-secret site decided to comment on this new launch schedule. After a while of silence they decided to turn to their own sources. So, server Woodcrest processors will be launched this quarter, but it will be a nominal event: the actual launch is planned for June 26 – just 5 days before the end of the quarter. In the meanwhile, Intel is going to be very aggressive about pushing Woodcrest into the server segment, so that the share of these CPUs will reach 75% by Q4 2006.
The launch of the desktop Conroe processors may happen in July, but the date hasn’t been selected yet, and the possibilities vary between the first and third weeks of July. As always, the first independent benchmarks will be conducted with the Conroe XE processors, and the more affordable models will start shipping later. By the way, this is where the frequency potential of the upcoming top-of-the-line Conroe processors gains more importance and actuality. Some sources claim that the currently emerging Conroe samples feature A1 core revision, while by the launch day they may acquire B0 core stepping.
The Merom launching schedule is the least certain of all. These mobile processors will be compatible with the existing chipsets, however the common inertia of the notebook market suggests that the new Merom based platforms will start appearing only in the fall, even though the Merom processor is scheduled to be announced in August. As for the completely new fourth-generation Centrino platform aka Santa Rosa, it should come out in 2007. Merom processors will then undergo some minor modifications (acquire higher bus frequency) and will enter the new platform, too.
Intel representatives have also said that they will continue reducing the production costs of the dual-core processors all the way until the end of next year, and by the end of 2007 we should expect the production cost of dual-core processors to equal that of the single-core ones. This is when low-cost dual-core Intel processors will probably start selling. However, they will continue with the single-core Prescott and CedarMill dies for the Celeron D family for quite sometime.
As for this year’s plans, Intel is going to rapidly get rid of all NetBurst based processor reserves, so distributors should probably expect even greater price drops than Intel promised at first. The shortage of desktop chipsets also needs to be eliminated, and this is when Intel is going to resort to ATI’s help: they have already increased the production of their discrete desktop chipsets by five times.
All in all, Intel seems to have a ton of plans. Hopefully, the end-users will really benefit from them and feel the difference this year already.
Comments currently: 2
Discussion started: 04/29/06 09:56:57 AM
Latest comment: 05/01/06 12:56:28 AM
Expand all threads |
Collapse all threads
[1-1]
1.
IF and only IF Conroe and kin perform like they have led us to believe, then this is a shot in the arm and a direction that Intel needed yesterday. Prescott was a bad idea when it came out. All subsequent iterations of it from dual core to higher cache have all been variations on that bad basic design. How many YEARS have to go by before you can put a new foot forward? Well if you are Intel it takes a lot of them. If AMD hadn't been battling the credibility scale, but had started 2 years ago, where they are now in reputation, Intel would have lost far more market than it has.
But the momentum is at an end. People care less and less which they choose, and want the best performance, the best efficiency they can afford at a given price point. Testing (good and bad kinds) is widespread and it's no secret where things lie. Intel should, if they'd applied resources correctly, have been able to bring Conroe last year or the begining of this one. And yes, I do know that is a hard thing on a scale like they have. But to sit calmly like they do and believe that a failed tech is something they can continue to promote as the "Intel way" is just plain bad business. It's irrespectful of customers and investors investments, and its long term damaging to Intel's reputation. Now, the only place they genuinely seem to care, is where it finally shows, in their wallets. That's embarrassing, imo. And it shows just how badly and how long their pants have been down at their ankles to see even with a design that does perform well, just how long, it takes for them to bring it to market.
Anyone will tell you from the business world that it's about speed, quick adaptability to change, and not leaving a bit of quality out of the picture that governs business today. Customers want it all. And they know they can get it in most regards. Intel had better smarten up, because another launch like this, delayed by this much could land them in the poor house. It's not just about great ideas, or great ideas, quick and agile and solid execution are equally important. And so far in those areas I'm quite dissappointed in Intel.
Checked: room for much improvement...
[Posted by: Anemone | Date: 04/29/06 09:56:58 AM]
+ expand thread (1 answer)
- collapse thread
You obviously have no idea how long it takes to come out with a brand new architecture like the P8 is. That's all your rant says.
There is no if with regards to performance for the P8. The motherboard makers are all concerned about the poor performance of the K8 vis-a-vis the P8, and have concerns about how well they will sell. There have been plenty of benchmarks on the web for it too. The Conroe is much faster than the Athlon 64, much. Better get used to it.
AMD hasn't really changed since 1999, since the K8 is really a K7 in drag. It is even less a change than the Prescott was compared to the Northwood. They are hugely complex designs that take a long time to make, especially something like the P8 which uses a lot of new technology.
The P7 was largely a failed design, but we should not forget that it did beat the K7 for a while with the Northwood core, and probably would still beat the K8 if they had not changed to the dreadful Prescott release.
Intel made serious errors when they made the P7. They were not expecting x86 to go to 64 bits since the Itanium was supposed to replace it. They also did not expect thermal/power problems to limit the clock speeds of the processors. It was a first for that, so I guess it was understandable. So, they had to throw together the Prescott to handle 64-bits, and it was severely limited by power/thermal problems which they did not anticipate.
So, it is not that they are so slow, compared to AMD they are speed kings, it is that they made market and technological predictions that did not happen. They never even expected to need anything after the P7, the Itanium line was supposed to replace it.
So, yes, they could have had a better crystal ball, but they have done a pretty good job designing a fantastic processor in a reasonable timeframe. AMD is still stuck on a modified K7 design, and it is showing its age badly against the modern stuff. It is even older than the P7, and certainly older than the Prescott, yet they are still stuck with this antiquated design. Why is it taking them so much longer? Were they content with good enough, rather than looking in the future? I wish I knew, they should have been able to beat Intel to the punch with a K9, but instead they offer the pathetic AM2 release. Yes, they are real design speed kings. Took them forever to even make an acceptable DDR2 controller. Great job!
My big question is, why did Intel not switch to the Pentium M line, at least to a large extent, when they knew the P7 sucked. It is a better chip than the Athlon 64, which a review on another site showed in a recent test comparing the Pentium M and Yonah against the Turdion and Athlon 64, and way better than the Prescott. They had good processors, but they sold the miserable Prescott instead in vast numbers. It is a mystery to me, but there is some justice that they lost market share with that vile processor. They got what they deserved.
[Posted by: TA152H | Date: 05/01/06 12:56:28 AM]
[1-1]
Make Home Page
Add to Favorites
Print article
Send link to friend
Digg this
Post to del.icio.us
Post to Slashdot
