Intel Again Delays 'Tukwila' Itanium Release

HardOCP News

[H] News
Joined
Dec 31, 1969
Messages
0
Intel says that, during final system-level testing of its next-gen Itanium, it has found an opportunity to “enhance application scalability best optimized for high-end systems.” Unfortunately that means that the processor launch has been pushed back to Q1 2010.

In addition to better meeting the needs of our current Itanium customers, we believe this change will allow Tukwila systems a greater opportunity to gain share versus proprietary RISC solutions including Sparc and IBM Power. Tukwila is tracking to 2x performance vs its predecessor chip. This change is about delivering even further application scalability for mission critical workloads.
 
what i think they mean is they dont see any companies buying that this year so they cant make as much money off of it so they are delaying the release
 
Tukwila is dangerously close to being canceled. Intel only has one Itanium customer - HP - who does almost all of the development work at this point. HP only offers Itanium in two arenas - Integrity (formerly HP9000) and NonStop. Unisys earlier this year dumped Itanium processors completely, and moved to supporting Xeons on their Cell based systems, citing the Xeons offering better than twice the performance of the top end Itaniums. Even doubling performance with Tukwila, it just won't be competitive with POWER6.
 
I thought Itanium was a total failure when it came out... I didn't know they still made it... It sounds like Intel is trying to find a way out of it with the "delay"...
 
ug, please no more itaniums. that are a pain in the ass to work with, 2 minutes to post 2 minutes to get through the EFI boot, slow ass reading from the dvdrom and even with 15k scsi its just slow as hell overall.....but man they can crunch some numbers.
 
I think intel would be better off just scrapping the itanium line and redistributing the resources to other departments.
 
I think it would be a shame if Itanium got killed off. The underlying design is quite good, just underdeveloped and underfunded compared to Intel's x86 offerings. Also it's using a comparatively ancient process.
 
what are/was sorry....
It's not like regular server processors. It's made as a RISC replacement in "big iron" (mainframe applications) and highly scalable systems. In a nutshell from Intel's page:
Itanium-based servers are incredibly scalable, allowing configuration in systems of as many as 512 processors and a full petabyte (1024TB) of RAM.

That said, it's beyond ridiculous how horribly late Tukwila is.
---

For the people calling on Intel to scrap Itanium, Intel sells about 2x more in dollars of Itanium processors than AMD sells of Opterons, and sales dollars of Itanium servers also eclipse Opteron servers. Should AMD also scrap Opteron? :p
 
For the people calling on Intel to scrap Itanium, Intel sells about 2x more in dollars of Itanium processors than AMD sells of Opterons, and sales dollars of Itanium servers also eclipse Opteron servers. Should AMD also scrap Opteron? :p

while this is may be true, microsoft owns something like 75% of the worlds itanium systems. i oh so hate dealing with them.
 
Is Itanium actually profitable for Intel? Didn't know that. I thought they almost always lost money on that, only doing it because the research and difference with the other architectures and designs.
 
Is Itanium actually profitable for Intel? Didn't know that. I thought they almost always lost money on that, only doing it because the research and difference with the other architectures and designs.

From people I have spoken to they typically break even on it.
 
Is Itanium actually profitable for Intel? Didn't know that. I thought they almost always lost money on that, only doing it because the research and difference with the other architectures and designs.
Yes, very profitable now. Intel spends ~$500 million a year on design/testing/manufacturing (probably more going forward now that it has 2 design teams), and it brings in $1.5-$2.5 billion/year in revenue (est. by analyst Gordon Haff).

The earlier years were subsidized by x86 profits of course. I can't find cumulative revenues and costs, so it's hard to tell if/when it has broke even. Many people like to believe that it hasn't. ;)
 
Back
Top