Replies to post #23378 on Advanced Micro Devices Inc (AMD)

[SemiconEng]

For SemiconEng's benefit: just my humble opinion.

You must have me mixed up with someone else. FYI, I can pretty easily tell the difference between what is data, and what is opinion, but thanks for singling me out. I really "appreciate" it.

BTW, Humble? Ummmmmm......... Never mind.

:-)

Regards,
Semi

[yourbankruptcy]

Actually I love the fact that K9 is on 90 nm. I recall some people complained. Why?

That's a good strategy, if Intel's 65 nm line is head-to-head performance wise with AMD's 90 nm line, why not to use mature process to compete against Intel's non-mature process?

Speed advantage is a good thing. It let you pocket it almost directly.

[sgolds]

UpNDown, I'm not sure what will be in K9. Based on what we know, Intel does not have to offer a product superior to AMD's K8 in 2005 - it just has to be 'good enough'. Remember, Intel has quite an impressive manufacturing capability plus a name that remains golden in the mind of many buyers. They can make quite a good business with a slightly inferior x86-64 solution, throw a lot of cache at it, and crank up the manufacturing.

Since AMD will be manufacturing limited at least until 2006, and because people will still want to purchase PCs, Intel can use this as a stopgap strategy while they continue to pursue Itanium. Alternatively, they can work on a better x86-64 in that time frame. Don't count them out so easily!

However, under this scenerio, AMD does make a lot of money (and that is a good thing). Also, AMD earns a reputation for having the best product available (also a good thing).

[chipguy]

Intel is now too far behind AMD64 technology to catch up on 64-bit x86. By the time Intel would be able to introduce a competitor to K8, K9 will be out.

???

Perhaps you need to learn the difference between ISA
architecture and CPU implementation. After how many years
and how many revisions of the ARM ISA did Intel step in
with XScale and grab most of the high end mindshare and
sockets? If you think there is something magical about a
64 bit x86 instruction set extension that would prevent
Intel from coming straight off the starting blocks with a
definitive implementation then you know little about Intel
or processor design.


The transition to 65nm. is probably the "tipping point" here. If AMD is able to pull that off -- K9 on 65nm. (I know K9 is supposed to be on 90nm. first, I'm talking about the first shrink), the microprocessor world as we know it will be turned upside down.

ROFLMAO. That paragraph belongs in a letter to Penthouse
magazine, all geek edition.