Doug Siebert (foo@bar.bar) on 2/24/09 wrote:
---------------------------
>anon (anon@anon.com) on 2/24/09 wrote:
>---------------------------
>>>If software compatibility was not important new ISAs would be introduced and dumped
>>>frequently. But it is, and it means there's less sense for Itanium to exist than some people imagine.
>>
>>And yet it continues to grow, despite years and years of disadvantages versus x86
>>(and against continual predictions of naysayers -- how can you continue claiming
>>you are right when you predicted it would die several years ago?).
>
>
>But see, the thing is, Itanium is not truly growing and never has been. PA-RISC
>and Alpha stepped aside to make room for it, and IPF has been slowly filling the
>vacuum they are leaving behind.

If you redefine "growth" so that it fits your argument, I can't see much hope for a good debate! But I'll try anyway :)

> Once the migration is done (and amazingly its not
>quite finished just yet, given the fairly impressive for Q1-Q3 2008 sales of PA-RISC
>powered Superdomes) that's pretty much it for growth, its all downhill from there.

Downhill as in shrinking market share? Maybe. I don't know I'm not trying to predict the future. But unlike naysayers have been forecasting for years, IPF right now has been growing, now has a very significant market share, and appears to have a pretty good medium-term commitment from Intel.

And I don't think it is at all guaranteed that market share will be relegated to PA-RISC replacement. What about SPARC? How long will POWER compete in the high end? Sure many of them will want to try migrating to x86, but as long as Intel segregates IPF into the high end, some will look there.

>It is nothing more than the next generation of PA-RISC, just as it was originally
>designed by HP to be. Just like PA-RISC migrated all the HP-UX customers off a
>different architecture while it absorbed the Apollo customers, Itanium likewise
>migrated all the HP-UX customers off a different architecture while it absorbed
>the Alpha customers. Itanium will never end up being any more dominant than PA-RISC
>was, because its life cycle is identical.

I'm not very knowledgeable about the market, but aren't a significant amount of IPF sales windows and linux based?

>I always predicted Itanium would plateau when PA-RISC was discontinued, then begin
>a long slow decline and it would suffer the same fate as Alpha (I'd say between
>2013-2015) Its all a matter of perspective, as we both see the same facts but draw
>different conclusions from them that so far lead both of us to believe our scenarios
>are unfolding exactly as we have foreseen :)

No I'm not trying to look into a crystal ball (I'm not a fanatical IPF proponent OR opponent). I'm just saying that the naysayers' crystal balls of ~5 years ago have turned out to be sadly mistaken.

I would never try to make a claim like "IPF will take over the high end indefinitely". But the fact is that it seems to be seen as profitable and worthwhile by Intel, and there is absolutely nothing wrong with going after a very significant market, even if it is slowly shrinking. Especially if the *eventual* outcome appears like your lower end products will grow up and take over completely, that does not automatically make the effort in the high end market a waste.

If you want a wild guess from me, I think it will depend on how well Intel executes the next couple of generations of IPF; where (if anywhere) AMD moves x86 to; how well POWER goes etc.

In the worst (for IPF) case, maybe they won't get another design out, and just respin things (maybe adding cores) and let it slowly die out at the hands of x86, along with the other high end RISCs. In the best case, it could eventually carve out a bigger chunk of high end market. It could even push on x86 in some areas too (eg. HPC. even mid-servers if the CPUs end up being *really* good).

I don't know. But the thing is, the "just wait for the next one" guys are not worse than the "there will not be a next one" guys, as far as I can tell.