Intel-AMD Seesaw Battle - What's Your Take?

[Guest]

Pentium beat K5
K6 beat Pentium
Pentium II beat K6
K6-2 beat Pentium II(price/performance)
Pentium III beat K6-2
Athlon beat Pentium III
Pentium 4 beats Athlon (I predict)

The most telling battle, however, which will eventually decide the outcome of this chip war is:

Itanium vs. Hammer.

While all the aforementioned battles pitted different two brands fighting over the same x87 architecture, the 64bit battle pits two different brands, two different architectures: x86-64 vs IA-64.

AMD seems to me to have a better strategy by bridging the 32bit gap. They have released their x86-64 spec and tools and are putting serious resources into the development of this chip... In two years we'll see if AMD can truly lead the market, or will they have to suck it up and follow intel as in the past.

I always root for the underdog... GO AMD!

 

[Killerkris]

Is AMD truely the underdog tho. They are playing with the big boys now with a reasonable market share, ok not up to Intels but i hope that AMD win too. I also hope that all these 64bit processors will be cheap (i wish) and able to play my current games........

Please dont hurt me.....please

 

[Lucol]

On paper, Intel's new architecture is revolutionary, while AMD's is evolutionary. I'll have to go with AMD in that all software now, will be faster with Hammer, while with Intel, it'll be slower (since it has to basically emulate 32bit processing). But what Intel supposedly has, is a very good compiler for it's Itanium, that discovers parallelism in code and optimizes it for better execution. On paper, Intel's architecture sounds far superior in terms of parallel execution of code and avoiding pipeline flushes, but like so many things, what's on paper is hardly ever translated into reality.
I'm an underdog fan myself, so AMD get's my vote.

 

[Guest]

What will happen(no degree of error) as long as the current time tables are followed:
1)Hammer will take market share form intel in mid/high range servers as well as consumer pc's(as long as they are cheap enough for consumer pc's). There performance will rock.
2)intel will counter with their plan b and release a chip similiar to hammer(extended x86 instruction set). Probably similiar performance to hammer.
3)itanium will not be scrapped but will evolve into intels main money maker once enough software is written for it probably mid to late 2004(for mid/high end servers).
4)by beginning of 2002 amd's market share will hit a plateau
due to limited production capabilities. By mid 2002 their market share will drop because overall demand will increase while their production remains stagnant. So intel will pickup market share by default.

And the struggle continues....

 

[Guest]

What is likely to happen is that hammer is a fine processor that competes well with Pentium 4 in strictly 32bit environment ,and M$ will not give a damn about doing an X86-64 optimised windoze (or would give us another unstable 64bit cheap hack of a 32bit OS just like the hellish win9x).

Things would change if Intel decided to extend their x86 processors to 64bit too (and no, it will not be AMD's X86-64. It'd be incompatible in a SSE vs. 3DNOW! way). This is when M$ and intel's DELL will be "excited" about 64bit x86 processors ,and we might start to see M$ working on a true 64bit x86 OS for Intel's instruction set.

K9 will then support Intel's 64bit X86 technology and abandon AMD's own x86-64 coz whoever got the support wins.

about the future of IA64? I really have no idea yet....=/

Just my opinion of course

 

[Guest]

yeah, Intel may have a good compiler, but I am still doubtful EPIC is as powerful as they say, just look at the crusoe... low power, lower performances. Just my thoughts.

Seoman. Newbie at last!

 

[Guest]

What's your take on the reported complexity of the Itanium.

People have said that its "complex instruction word" achitecture is hopelessly difficult to implement and very slow.

Perhaps they went in the wrong direction... Or maybe it just appears that way on the surface - like P4 and RAMBUS... That unholy union may work out well for Intel in the long run.

There are already noise issues cropping up with Athlon and DDR on Micron's Millenia MAX XP. It may be very difficult for AMD to implement DDR300 or QDR400 to keep up with RAMBUS.

RUDIE