Replies to post #3978 on Advanced Micro Devices Inc (AMD)

[Haddock]

Programmers absolutely hated the 16 bit x86 segmentation model
and many gushed over 68k computers like Amiga and Next computers. But where are Amiga and Next now?

Guess what: Amiga and Next weren't capable of running 16 bit x86 segmentation software with zero performance penalty. (Actually, Next hasn't disappeared. Their ideas, people, software etc. are what makes up the backbone of MacOS X). The 386 was capable of doing that, and who wanted a 286 after that came out?

About $16B of Xeon servers sold last year. That is a absolutely huge incentive for ISVs to code for PAE when server class application needs more than 4 GB of memory.

Yup, and Opteron runs all that software too.

But while the pain of using PAE may be worth it for database people, who are used to handling their own block IO and where there isn't a huge number of different database apps to port the question arises what to do with new software that hasn't yet been hacked around to make it PAE-digestible.

And yes, I think management will listen to the software grunts if they say that a PAE port will cost lots of time, money, testing, added complexity (=bugs) and buy no performance advantage at all over just porting to real 64 bit platforms like Power4, IA64, Sparc or Opteron.

FFS, even ARM is getting 64 bit extensions. SPARC, Power, MIPS got them years ago, x86 has them now. Intel has delusions that by ignoring the 64 bit x64 market they can make it go away. All they will achieve is to disadvantage themselves going forward in the only market they ever made money in: x86-based CPUs.

But what do I care, I don't have INTC shares. The only reason I keep an eye on it is because it's such a huge boon to AMD that Intel is doing this, so I need to know when they are going to throw in the towel and admit they were wrong.