someone (someone@somewhere.com) on 2/24/09 wrote:
---------------------------
>Linus Torvalds (torvalds@linux-foundation.org) on 2/24/09 wrote:
>---------------------------
>[..]
>>
>>I'm just not seeing the market accepting Itanium.
>
>Maybe that's why people pay Gartner and IDC big bucks
>for their data instead of taking your opinion for free.
>
>
>>I'd like to point out, that as a doubter of Itanium, I'm
>>the one who has been right historically.
>
>Delusional as well as megalomaniacal, what a
>winning combination!
>
>Linus: Itanium sucks and will die any day now.
>
>Buyers of big computers: Itanium is great and we
>will buy a billion dollars worth more of them per
>year every year for five straight years.
>
>Great call there genius. I'll go with the people who vote
>with their wallet over the guy who only runs his mouth.


Your mindless IPF cheerleading is pretty funny to watch, your reality distortion field must have been manufactured by the same company that made the one Steve Jobs uses. Not to sound all megalomaniacal like Linus or anything :) but the IPF rollout is going just like I have been telling you for years it would go. Its been growing by pretty much the same amount as it has been displacing PA-RISC (and to a lesser extent Alpha, for those sticking with VMS)

Any switching from Sun or POWER to IPF is probably about matched by switching in the other direction (well at least on the POWER side) as has always been normal in the RISC world. But once the switch goes in the direction of x86, it is permanently a lost sale for the RISC/IPF market, thus the relative size of that market continues to shrink versus the ever-stronger x86 market.

PA-RISC displacement is nearly complete by now - I would have said it is complete but for the surprising fact that 15% or so of new Superdomes sold in 2008 Q1-Q3 were PA-RISC! So maybe there's $1 billion more growth left between the last dregs of PA-RISC market uptake along with the pent-up demand those who have been delaying new purchases waiting the long wait for Tukwila.

After that last $1 billion of growth I see the IPF market as flat to slowly declining as x86 continues to eat away. I'm sure you completely disagree with my assessment here. You used to crow about doubling YoY, then nearly doubling, but that's long over so now you talk about adding $1 billion a year. Once that's gone and the growth is flat, are you going to be reduced to bragging about IPF's growth beating that of POWER and Sparc by a few percentage points? Or even worse for you, to claim IPF is winning because its negative growth is less than that of POWER and Sparc - should that occur, no doubt you'd use the economy as an excuse and claim the recession will somehow shake things out in a way that lands IPF on top and relegates the others to the dustbin! (I'd argue if that happens it is x86 that would be the winner)

You used to talk about all the non-HP OEMs using IPF (despite many selling only a handful of systems) but as they abandoned IPF one by one you made excuses. At first you said IBM was just waiting for the next generation of Itanium, then you said it was due to the POWER people wanting to kill it because it made their CPU look bad. Dell was just a box shifter and couldn't compete against HP's custom boards (no argument there, they can't even compete against HP selling x86 hardware anymore) Most recently you said it didn't make sense for Unisys to sell IPF (where was that belief for the last 5+ years you have been proudly listing them amongst the "many" IPF OEMs?

I look forward to you twisting and torturing yourself in even more astounding ways in the next few years as you continue to try to save your indefensible position with a your favorite tools: a combination of cherry picked statistics, wishful thinking, and a complete abandonment of all logical thought.