Tuesday, July 18, 2006

INQ and I are in agreement on Itanium

I agree with their insight quite often, Charlie at INQ wrote his analysis on Itanium, which I agreed in March 2006 more eloquently in a post titled "Intel execs are brain dead". But unlike Charlie, I gave Intel the benefit of doubt and assumed that they sell Itanium at $5000 with a cost of $4000. Let me repost my old article below:

Intel execs are brain dead
SUN and IBM make all the money on Sparc and Power, they are system solutions providers, they do the whole hardware and software stack, CPU, server, networking, storage..., operating systems, enterprise software stack and services, the gross margins are very high.

What does INTEL get for Itanium?

Even at $5000 per CPU, Intel is only getting $1 million bucks for the 1000 CPUs it is selling. Right now, HP is making all the money on Itanium.

With its primitive copy-exact methodoly, Intel is tying a FAB line for making a few hundred chips a quarter. This is bad business.

However, without Itanium, Intel becomes a pure AMD64 clone maker. Intel has to continue Itanium for the last bit of pride it has left.

Posted by: sharikou Posted on: 03/23/06


Anonymous Anonymous said...

The Itanic has been flogged to death.

The purpose of Itanic is not to somehow sell millions of chips.

Rather, the purpose of Itanic was to kill Alpha and the other promising developments in the RISC market.

And guess what? Itanic succeeded for the most part.

No more MIPS for example. And no more SGI (which spent hundreds of millions on switching to Intel/Itanic).

For today, Itanic is just a testbed for Intel to see how to make giant chips. So whenever Intel needs to, it can apply that learning to the Core platform. For example, look at the giant cache that is available on the new Xeon MP.

Journalists like Charlie that have to write about it because they can't do it... need to move onto interesting topics. Find something new to flog.

3:07 PM, July 18, 2006  
Anonymous Anonymous said...

Charlie@inq is a big joke, and you are a much bigger one Sharikou

3:08 PM, July 18, 2006  
Anonymous Anonymous said...

Itanic is the joke, not sharikou.

3:34 PM, July 18, 2006  
Anonymous Anonymous said...

"With its primitive copy-exact methodoly, Intel is tying a FAB line for making a few hundred chips a quarter. This is bad business."

Do you really think Intel has a "fab line" dedicated to making Itaniumm chips - it is simply a different set of lithography masks and alterations to some process steps; not much more different than running a Athlon vs Opteron chip or a Core vs Core 2 vs P4. Are you really that ignorant on how things work in a wafer fab?

6:20 PM, July 18, 2006  
Anonymous Anonymous said...


you and the INQ remind me of Dumm and Dummer..

Good for you

7:50 PM, July 18, 2006  
Anonymous Anonymous said...

Some other thoughts..

Try for one moment to be unbaised.. I know for Sharikou its very very hard if not impossible as I've been watching this thread recently and have not once seen a level headed objective post.

But give this a thought.

You have worked hard with some partners and invested billions and years into a new product line. This new product line isn't making money, may in fact be losing money, lots of it. But your customer still has a reputation and commitments on this product line.

Most important this customer also buys other products from you and you make LOTS of money on this other line.

What do you do?

Drop the money losing line like a rock and screw your customer?

Keep development at the minimum required?

For you Sharikou lets pretend its AMD we're talking about... lets say there old AM29000 line

The Doctor

8:25 PM, July 18, 2006  
Blogger netrama said...

Just one quick glance at that INQ report and it is clear that Pat is full of plain White lies. I am unsure who he is trying to fool with his comments. Not the average Joe. And neither the so called "Japanese" customers. The only folks remain is the crooks and analysts who perhaps buy such BS crap.

8:51 PM, July 18, 2006  
Anonymous Anonymous said...

hey sharikou,
have you seen the reviews of montecito?
Each of the reviewers below have great things to say. I think its the beginning of something great for Intel.

Montecito starts pumping "big iron"(zdnet)
The beginning of a new era (pcblitz)
Game changing chip (sako)
Intel Vroom, Sun Doom (64bitworld)
Montecito juggles 40X performace over Opertons (Jdreview)

Some people in the blog refer as itan"ic" but it may actually be fantast"ic"

11:26 PM, July 18, 2006  
Anonymous Igor said...

The problem with both yours and INQ analysis is wrong Itanium positioning.

Take SPARC for example. We all know Sun revenues for systems, but we do not know TI revenues for chips. However, I suppose revenues and profits are good for TI.

Same here. HP and Intel. And non-disclosure agreement between them.

All the rest is pure marketing, which you don't understand and therefore hate.

11:27 PM, July 18, 2006  
Anonymous Anonymous said...

Intel probably has in its labs a 24MB cache Woodcrest Xeon with an Itanium "compatibility unit" on the chip as well.

As soon as Itanium and Xeon are done moving to a common socket, that will be when Intel can phase-out Itanium at the same time "saving face" for the various Japanese management types whose only option at this point is ritual suicide.

So the common socket Xeon will come with a little sticker "Itanium Inside".

They'll do the sticker for a few years and then Intel will quietly drop it based on lack of demand and desire from their partners "to go in a new direction".

11:59 PM, July 18, 2006  
Blogger Sharikou, Ph. D said...

I suppose revenues and profits are good for TI.

Total crap. You even saw TI execs hyping Sparc making billions?

Intel's retarded execs were lying off their rear. They are trying to distract the fact that Itanium was a total failure.

If TI's CEO and CTO were out there bragging Sparc made billions, they would be laughed at like idiots.

Intel's situation is worse. Beaseide FAB the chip, they have a large groups of engineers designing this chip---and they only sell a few hundred pieces per quarter.

12:17 AM, July 19, 2006  
Anonymous Igor said...


get your logic back, please.
When company introduces new product, especially on crowded market (SPARC, POWER, ALPHA, MIPS) there is always a lot of hype.

What is success and failure ?
10% of server market is success or failure ?

Suppose for AMD you'd say it is HUGE success, but for Intel is TOTAL failure.

Itanium is really big success for Intel in high-end.

12:35 AM, July 19, 2006  
Anonymous Anonymous said...

I think the problem is NOT if Itanic is a good or a bad chip.
Intel tried to use its marketing and financial force in the corporate server market: the problem is x86, and more AMDX86-64, got so much better, in terms of cost/performance ratio, that pushed higher and higher the level where a dedicated architecture is needed.
Ten years ago, if someone told you that Google would use x86 servers, I think it would have been considered mad. x86 is gaining market because you are NOT bound to proprietary architecture, even if you lose something in terms of performance.
You can swap components you could buy anywhere, and that is a great "plus".
Itanium was a big error, son of "it's impossible to apply 64 bits to x86" and "world doesn't need x86 64 bits".

1:07 AM, July 19, 2006  
Anonymous Anonymous said...

"What is success and failure ?
10% of server market is success or failure ?"

In business world, that is simply. If you make money out of it, that is a success. If not, that is failure.

In a business venture, if your partners are making tons of money while you are losing money. DO you really want to continue this relationship? Want to be the next SGI? Think again.

In the business world, co-orporation is the mean, and profit is the end. There is no such thing as business friendship, but business venture.

4:01 AM, July 19, 2006  
Anonymous Anonymous said...



"If we could unwind the clock, I would have just built a RAS version of Xeon to attack the market," he said, using an industry term for "reliable, highly available, and scalable" chips, and referring to Intel's Xeon server chips, which employ the widely used x86 instruction set. Itanium uses a less popular design called EPIC."

Come on guys, how can anyone defend this product?

4:04 AM, July 19, 2006  
Anonymous Anonymous said...

Intel solution for everything:

More cache ($$$$)

6:32 AM, July 19, 2006  
Anonymous Anonymous said...

owever, without Itanium, Intel becomes a pure AMD64 clone maker
Not only that, intel also copied AMD's x86 architecure. Further more it copied MMX, SSE, SSE2 and SSE3 as well. The only thing they didn't copy is 3DNow. Even the SSE4 also copy from AMD future chip. There is no innovation in Intel.

I heard that AMD accidently lost documents contains PCI, USB, LAN, AC97, etc and intel went and declared as if it is Intel's. intel is so evil.

AMD rocks! 4x4 rocks!

8:07 AM, July 19, 2006  
Anonymous Anonymous said...




Lets face it Intel fanboys, x86-64 put a fork in EPIC/IA-64 as there are far more Operton servers sold in a year than Itanic has been sold in it's entire lifespan. It was simply an ISA platform that Intel was using to not only to subvertly eliminate all competition from the server market, but eventually AMD from the desktop as well as they had such illusions of grandure with Itanic as the future of 64-bit computing.

The bottom line was simple between EPIC and AMD64: Why pay exorborant prices for such a niche ISA where you will have to buy all new software(what little you could find that is) for and suffer it's horrible 32-bit performance in an emulation mode? Instead, you can get a seamless, robust, well based 32/64-bit ISA that will already use your existing apps and give you room to expand fully to 64-bit in the future and gain even more performance as the 64-bit software becomes available and mature. The choice was a no-brainer it seems for most IT professionals out there. I won't even touch much the hardware side of Itanium or Itanium 2 as it's still on the 180nm process last I read, read: slow and hot. AMD's upcoming Torrenza co-processor technology with the K8L and beyond will indeed seal Itanium's fate for sure.

You should all be thanking AMD for keeping us away from the horrors of EPIC, all it's properitary BS(software) and for keeping Intel from being a complete monopolist in the CPU market, much less making the average consumer pay $2000 for a entry level Itanium or Itanium 2 chip.(which it's performance is far less than stellar for most tasks versus x86-64, save it's floating point capabilities)

Itanium was and still is a still-born ISA for the mainstream market, DOA.

9:41 AM, July 19, 2006  

Post a Comment

Links to this post:

Create a Link

<< Home