Thanks for the complement, cdrakejr.
I read your article. It was pretty interesting. Of course it all sounds great on paper, and the author certainly puts a spin on everything so as to put AMD's future in the best light possible. I think if you were to read a similar article on Intel's future, you would have an equally good outlook in that direction. Fortunately or unfortunateky, however, Intel usually likes to keep secret the plans in developement.
AMD certainly promisses a lot, but as the article admits, they have made many mistakes in the past. Even Intel has shown the public that it makes mistakes. Therefore, AMD's new record of success isn't guarenteed to follow them forever. To put things in all fairness, AMD can certainly do everything they say they will. To put things more realistically though, they have been very lucky to ride their flagship product into excellence, but the number of achievements they promis will be in store may be more difficult for them than they think.
In terms of their 760 chipset, I have no doubt that the UP version is on time. It won't be hard for AMD to catch up in terms of features like AGP4x and ATA/100. Being able to support DDR 266 is going to be a very difficult task however. I have heard of difficulties with DDR 200 electric properties. I immagine DDR 266 will be even harder. Not to mention that AMD is planning on creating 2 point-to-point interfaces for their MP version of the chipset, which requires some very advanced techniques to keep cache coherancy between processors.
Intel tried putting too many features into one chipset once. It was called the i820, and we know the issues with that board. The performance of Rambus on the i820 was stifled because of electrical limitations on the board (Rambus requires a certain amount logic to implement itself efficiently; the i820 had to compromise because of electrical limitations). Perhaps the 760 will follow the same fate, since AMD is trying to accomplish so much in so little a time.
The artical mentioned a lot about fab allocation, and it seems that the Austin fab creates processors below 900MHz (including Duron), while the Dresden fab creates the 900MHz and above speed grades with copper technology.
The Muastang and Palomino chips will certainly be using copper, which means they will have to be manufactured in Dresden. The artical suggests the fab still has room to grow, and AMD certainly wouldn't want to sacrifice their other processors being manufactured. Therefore, the volume for Mustang and Palomino will depend on how much room AMD can create in their #1 fab. Personally, I am scepticle that high volumes can be produced with limited space, but maybe AMD will surprise me.
I figured that Microsoft would eventually announce plans for an x86-64 based OS (they wouldn't miss a business opportunity like that), but I have a hard time believing that they will have the resources to finish the OS in less than a year. Granted, they could have begun developement long ago, but since AMD recently released the x86-64 spec, I somehow doubt that Microsoft had exclusive information for too long before anyone else.
Besides, Microsofts long record of release pushbacks would suggest that a brand new operating system would take more than a year to finish (though early builds will probably be ready for the Hammer tape out). Microsoft spent years creating the IA-64 version of Windows... granted that IA-64 is far more complex. But, Intel gave Microsoft a lot of help, especially in the compiler arena. AMD does not have the compiler gurus that Intel has, so Microsoft will have to do a lot more work themselves. This could impact the release date.
I also find the rumor that Apple would use the Athlons extremely hard to believe. First, OS X is far into developement, and it isn't likely that an x86 port could or would be done at this time. Second, the architectual implications of an x86 processor in a mac system would require a major redesign of the entire chipset. The validation would be tremendous, since compatibility would have to be maintained. I don't see this happening in the next couple of years.
But, the article did provide another angle to the future of the industry. Thanks again for the link, cdrakejr.
Facebook
Twitter
Instagram