   |
|||
|
|||
|
:
News
: Archives
|
|
AMD and 65 nm |
|
|
The Bittersweet Taste of Modest Success |
|||
|
|
|||
|
by Josh Walrath |
|||
|
AMD was finally able to deliver its first 65 nm parts this month to OEM’s, with retail availability supposedly appearing this January. While there are those on both sides praising and criticizing AMD’s timing of this release, we all should take a deeper look at the issues surrounding any process change and how AMD has and is handling it. We must also take a look at processor design, the manpower behind such endeavors, and the timeframes which AMD has to work with. If designing a processor was an easy task, then we would have seen far more companies doing it. As it is, this highly complex dance of design, fabrication, and technology has effectively left only two major players in the x86 arena. AMD’s Step Transitions Process node changes are invariably complex and time consuming, with each generation bringing up a plethora of technical challenges that must be overcome in order to succeed. Compared to Intel, AMD is a small company with a limited amount of fabrication engineers and Fabs. Considering that AMD has net revenues in the same league as Intel’s net profit per quarter, we can see that AMD just does not have the cash or resources to go head to head with Intel when it comes to process technology. This does not mean that AMD has thrown in the towel against its bigger rival, but it certainly has to do things a lot smarter to keep competitive. AMD currently has two Fabs in Dresden, Germany. These are Fab 30 and Fab 36, both located right next door to the other. It was only this year that Fab 36 opened up for production, thereby easing production constraints for AMD. Also by opening Fab 36, AMD was able to have more space to implement their 65 nm line, all the while continuing to produce 90 nm parts at both Fabs. This allows AMD a greater measure of flexibility than it had experienced before with its transitions from 180 nm to 130 nm to 90 nm, where the only Fab producing processor parts and implementing new process nodes was Fab 30. So how exactly is AMD able to compete with Intel from a manufacturing standpoint when Intel started transitioning to 65 nm over a year ago, and currently have three full Fabs running at 65 nm? AMD has pursued a surprisingly simple yet far flung plan to leverage its experience with others to hopefully produce a world class 65 nm process. The “virtual gorilla” concept that AMD uses is nothing new. AMD has almost always been behind Intel in most technology changes, and they simply do not have the money or manpower to compete against Intel in a one on one situation. Instead AMD has partnered with other industry leaders in process tech. This partnership leverages each company’s R&D resources, engineers, and experiences so that the effort each company puts into the endeavor is multiplied when it comes to the final product. Currently AMD is partnered with IBM, Infineon, Chartered, and Samsung for developing their advanced processes. The resources these companies can pool together make it very competitive with Intel. While in the past partnerships like this were few, the technical challenges of deep sub-micron development have become so overwhelming and expensive that the idea of going at it alone is simply not an option for most companies. Intel of course has the engineering talent and monetary resources to do such a thing, but they are in a nearly unique position.
If you have found this article interesting or a great help, please donate to this site.
Copyright 1999-2006 PenStar Systems, LLC. |
|||
