new itanium coming july 18

[jhu]

these are based on the montecito cores. dual-core, 26 mib of l3 cache. apple should really consider these things for the xserve line.

and to preempt anyone's prior ideas of the "itanic", these chips are still intel's fastest. about the only processor that beats it is ibm's power5.

[AlBDamned]

Originally Posted by jhu

these are based on the montecito cores. dual-core, 26 mib of l3 cache. apple should really consider these things for the xserve line.

and to preempt anyone's prior ideas of the "itanic", these chips are still intel's fastest. about the only processor that beats it is ibm's power5.

Just saw a story on this in Macworld UK.

1.72Billion transistors. I have zilch knowledge about these things but that sounds like a hella lot of silicon. Especially as the Yonah has a pithy 150 million transistors.

Link

P.S. Montecito - another cool sounding intel codename, but Itanium sounds pretty smart too...

[bigandy]

wouldn't mind seeing something like that in a top end xserve... they are beasts, they really are...

[AlBDamned]

Originally Posted by bigandy

wouldn't mind seeing something like that in a top end xserve... they are beasts, they really are...

Would make sense wouldn't it - and with rumors of a new Xserve soon and Montecito being released on a Tuesday, we could be in for them to launch tomorrow. Maybe.

[MisterMe]

Originally Posted by AlBDamned

Would make sense wouldn't it - and with rumors of a new Xserve soon and Montecito being released on a Tuesday, we could be in for them to launch tomorrow. Maybe.

Apple has been very clear on this issue. Macintosh computers will be based on x86, not Itanium.

[AlBDamned]

Originally Posted by MisterMe

Apple has been very clear on this issue. Macintosh computers will be based on x86, not Itanium.

Have they? What's the difference? Like i said, I don't know much on this stuff - but a Mac with 1.72 billion transistors and 26MB L3 Cache would be quite impressive and sometimes it's amusing to speculate wildly.

And also, when, recently, has Apple been clear about anything they'll be using in their products?

[MisterMe]

Originally Posted by AlBDamned

Have they? What's the difference? ...

This is reasonably good summary of the Itanic [sic]. If you want to read more, then you can find more on Factbites.com.

[poppe]

Through it in the Towers haha ha....

[jhu]

Originally Posted by AlBDamned

Have they? What's the difference? Like i said, I don't know much on this stuff - but a Mac with 1.72 billion transistors and 26MB L3 Cache would be quite impressive and sometimes it's amusing to speculate wildly.

And also, when, recently, has Apple been clear about anything they'll be using in their products?

i'm not really sure what apple's plans for xserve really are. as rack-mount servers, they're not really all that special. on the other hand, if they did stuff an itanium in there, they could have bragging rights as a high performance computing platform.

[MisterMe]

Seeing as how only two Itanium-based supercomputer is in the Top 10, three are in the Top 20, and seven are in the Top 50, exactly what will Apple have to brag about?

[jhu]

Originally Posted by MisterMe

Seeing as how only one Itanium-based supercomputer is in the Top 10, two are in the Top 20, and three are in the Top 50, exactly what will Apple have to brag about?

look closer. there are 2 itanium systems in the top five (#4 and #5), and they have the least number of processors. certainly if you made a cluster of itaniums with 131000 processors, it'd be right at the top.

[MisterMe]

My previous post has been corrected, but the point remains: As the basis of only seven of the Top 50 supercomputers, the Itanium 2 is not exactly the standard for high-performance computing.

[jhu]

Originally Posted by MisterMe

My previous post has been corrected, but the point remains: As the basis of only seven of the Top 50 supercomputers, the Itanium 2 is not exactly the standard for high-performance computing.

by that notion, neither is the power5 since there's only one power5 system in the top 50. the point is that apple could start marketing the xserve as an hpc-type platform, and it'd be more attractive with an itanium given its raw processing power. plus, there's no way that apple would be crawling back to ibm.

besides, apple is probably the best at blowing its own horn about such things.

[MisterMe]

Originally Posted by jhu

by that notion, neither is the power5 since there's only one power5 system in the top 50. ...

The POWER5 is not the only IBM processor. Of the Top 10 supercomputers, four are IBM eServer-based systems. Of the Top 50, twenty-four are eServer-based systems. This doesn't count the Xserves and other PPC 970-based systems. When these are included, IBM processors account for more than half of the Top 50. Itanium is not in the same league. I am anxious to see how Apple satisfies demand for MacOS X-based high-performance computing. You can rest assured that Itanium is not a consideration.

[jhu]

Originally Posted by MisterMe

The POWER5 is not the only IBM processor. Of the Top 10 supercomputers, four are IBM eServer-based systems. Of the Top 50, twenty-four are eServer-based systems. This doesn't count the Xserves and other PPC 970-based systems. When these are included, IBM processors account for more than half of the Top 50. Itanium is not in the same league. I am anxious to see how Apple satisfies demand for MacOS X-based high-performance computing. You can rest assured that Itanium is not a consideration.

you were talking about processor architectures. if you want to change the comparison to processor manufacturers, intel takes more than 54% of all of the supercomputers on the list (this includes ia-32 and emt64). this only shows that practically any architecture is a viable hpc platform. on the other hand, itanium is at least comparable to, if not better than, anything ibm has to offer. just because there are less itanium systems on the list doesn't necessarily mean it is not a viable hpc platform. by the same logic, alpha, pa-risc, and sparc aren't viable hpc platforms.