Replies to post #8329 on Advanced Micro Devices Inc (AMD)

[SemiconEng]

I honestly think that when Intel can build Itanium processors + chipsets for a total cost of < $100, they will push hard to replace IA32. IMO, this will be in less than 3 years.
The investment in 12" fabs and tools that Intel made was foolish if they aren't plan to be making mostly IA64 within 4 years.
Petz


Maybe they will someday. I think business will get to 64 bit before Home use. Most likely it will be a combination of cost reduction on IA-64 and demand. 3 years sounds about right, probably less. By then intel will have several fabs on 300mm. I can't recall how many are already in operation, and how many announced. Fab11x in New Mexico, Fab D1c (D1D?) in Oregon, IFO in Ireland, and the recently announced conversion of Fab12 in Arizona come to mind. Seems like intel has to do something with that capacity.

Semi

[Jerry R]

Petz - they will push hard to replace IA32

IMHO, Intel would push hard to replace IA32 if and only if there was a reasonable expectation that this would expand its revenues and profits, not cut them. This would necessarily be dictated not by Intel but by it's customers.

Intel has always had an intense focus on generating revenue and future opportunities for itself. It has made several mistakes along the way, but its business model provides for a margin of error, and it doesn't stop trying. This is one reason why I respect them as a company and willing to invest my money in them as well.

There are some here on this board who believe Intel is too arrogant and has too much pride to either drop a failed product line or adopt another company's standard. Those who believe this are wrong. They are wrong because Intel has demonstrated throughout it's history that it can make the tough choices, and that it doesn't have the kind of pride that blinds it to potential revenue and profits.

For example, Intel invented the semiconductor DRAM chip (the 1103) in 1970. It went on to become their first commercial success. Less than a decade and a half later, competition from Japanese memory makers forced Intel out of the DRAM business. This was absolutely humiliating, as the company that invented the product could no longer make them. But it made the right decision from a business sense, invested in the 386 CPU (as well as a few commercial failures like the i432 and i860), and became profitable once again. More recently, the decision to execute on Rambus memory technology was a failure, and placed the Willamette platform in a precarious position, but Intel subsequently reversed course and supported DDR, which has become the de-facto standard and helped Northwood to achieve almost unquestioned performance leadership.

Obviously, companies that make mistakes and sticks with them are bad and not worthy of any investment attention. Those that make mistakes and learns from them, however, typically have a long and profitable future. Just my two cents.

[kpf]

Petz

I agree. But I cannot imagine more than a small fraction of tens of thousands X-86 applications will be ported to IA64 in less than five years from now.

From the original Itanium-schedule (Workstations end of last millenium) this transition possibly could have been pulled by Intel. From today's viewpoint, I think it could only be done if Intel would pay up massively for porting.

K.

[Tenchu]

Petz, When do YOU think the transition to 64-bits will take place?

I honestly don't think such a transition will take place before 2010. Even the transition point for 32-bit software didn't take place until years after the introduction of the 386, and the jump from 16-bit to 32-bit was much more useful than the jump from 32-bit to 64-bit. (Multitasking, you know.)

Tenchu