Was haltet ihr davon?
Scheinbar geht das Gerücht um, Intel versuche das Itanium Projekt los zu werden. Offenbar ist man dermaßen von der Performance des Woodcrest überrascht, dass Itanium nun völlig auf Eis gelegt werden könnte. Derzeit noch ein Gerücht, wie gesagt.
vgl. http://theinquirer.net/Default.aspx?article=30894
Japanese buying Itanic rumour is just that
In the land of the rising speculation
By Nebojsa Novakovic: Monday 10 April 2006, 14:51
THE RECENT "strong rumour" about Japanese consortium (whatever that may mean) "buying" over the Itanium platform from Intel (whatever that may mean) sounds quite interesting - if not intriguing, as it is hot on the heels of the more official announcement recently about a 10 billion dollar Itanium investment fund by a group of vendors.
Of course, that group seem(ed) to be anyway investing an aggregate total amount of that size in the recent times on this platform.
But why a Japanese consortium, that is, one in which there is no HP, the co-author of Itanium and the party who sacrificed the most for it, and SGI, the nearly bankrupt Itanium system technology leader?
One source pointed out that, if the newly refreshed Intel leadership treats the Itanic as possible financial fiasco, the easiest way out is to entice someone as far away as possible from Santa Clara (and the prying eyes of English-speaking press) to take in the hot potato.
That someone should be seen as "saviour" and "committed partner", i.e. able to continue to at least support, if not co-develop, the product for a while, at least from technology capability standpoint, and yet have a remote chance of limited local success based on heavily proprietary and localised software market - so, Japan seems to be the ideal place for this. But, why would such a consortium even buy the IA64?
Firstly, the meaning of a word "buy" can mean different things - for instance, you can buy something for one dollar, and still receive a one billion dollar "marketing fund" cheque thrown in as a "bonus gift" by the seller. Rarely happens to us ordinary mortals, but I guess still can happen.
Secondly, even if for some reason that consortium doesn't continue to sail the ship Itanic after awhile, and sends a kamikaze to sink it down silently, some of Itanic-related technologies and IP, like a very capable FP unit, for instance, could be useful for Japan's continuing niche processor design efforts, including vector processors and alike. After all, Fujitsu designed general-purpose SPARCs so much better than the Sun ones that, at the end, Sun gave up and decided to use Fujitsu ones anyway.
Thirdly, as mentioned, in Japan's highly proprietary and unique, and most importantly self-sufficient, IT market, a line of Itanic-based systems could continue to live nicely for sometime, without dependence on the whims of, say, Microsoft and Oracle.
Now, why isn't HP and SGI in that rumoured (remember, it is still only a RUMOUR!) consortium? Well, for SGI, the company situation itself tells you the answer: chances are, SGI as a company is unfortunately counting its last days, with the last remains of graphics floating towards Nvidia, and whatever useful other patents maybe in the hands of Microsoft. And for the system engineering division? Well, it may end up in HP or one of these Japanese vendors.
And for HP? Hmm, I'm not privy to all the internal corporate shenanigans and mood swings when it comes to the strategic decisions, but another (even stronger) rumour abounds that, like many other vendors, HP is very positively suprised with the performance of Woodcrest Xeon. The thing not only beats the current and near-term Opterons in the mainstream 2-socket arena but, more importantly, seems to beat the Itanic in more benchmarks than ever before. But, hey, these are only small systems?
Well, due to the pressure from Opterons' success in the lucrative 4-socket and 8-socket systems, Intel has no choice but to let Woodcrest and its successors scale up in one way or another. Whether it is via complicated quad-FSB chipsets, the delayed CSI interconnect or, as I suggested, taking in HyperTransport (just like they took in iAMD64 and made good money) and solving the problem once for all, only Intel knows. But that move, combined with the new platform's performance, could spell a death knell for Itanium as a major platform - at least within the Intel portfolio. So, maybe you'll soon find new, dual-core and dual-use Itaniums in the Land of Rising Sun, running as both CPUs and fashionable teppanyaki grills...
|
Thread abonnieren
ETF-Sparplan