Fog Creek Software
Discussion Board




Intel's Itanium: bloated and behind schedule

<quote>
It has taken an entire decade, an estimated $5 billion and teams of hundreds of engineers from the two companies to bring the first Itanium chip to market. As the struggles and costs mount for the companies, skeptical technologists say Itanium now has the hallmarks of a bloated project in deep trouble. It is already four years behind schedule, emerging just as companies are in no mood to spend money on technology.

"Every big computing disaster has come from taking too many ideas and putting them in one place, and the Itanium is exactly that," said Gordon Bell, a veteran computer designer and a Microsoft researcher.

INCREASINGLY, Intel is facing the risk that it has chosen the wrong path to high-performance computing.
</quote>


http://www.nytimes.com/2002/09/29/technology/circuits/29CHIP.html

J. D. Trollinger
Sunday, September 29, 2002

"Young man, don't believe everything you read in the paper ... "

Agnus Moorehead
Sunday, September 29, 2002

I agree with J.D.  The Itanic is an urban legend.  Intel is not going to lead the way into 64 bit computing.  SPARC and IBM are already there, not to mention TRU-64, which by the way Intel acquired midway through the Itanics design.

Intel is sorely lacking in the 64 bit department.  Another strike against them if the do actually build a working chip,  their chips use more electricity and run hotter than most.  Data centers don't like that.  If you want a look into what might be the future of high end computing, think Transmeta, SPARC and IBM. 

It's just too crazy to believe that the company that puts its chips into 90+ percent of the worlds desktop computers is going to design the best high end chip.  Intel lives on Mhz and hype.

Maybe I am wrong.  On the other hand how long has Intel been slipping the time line and saying "next year", "in the next version", etc...  They are peddling vapor.

It would be nice if we got a kick ass 64 bit chip for Windows though, maybe then vldb on Windows would be a possibility.  Well it would, if Windows didn't have a single threaded page file and could multi-task disk i/o.

Still crippled in Wintel Land.

ryan ware
Monday, September 30, 2002

It looks like Intel is going to follow AMD's lead and produce a chip that is a simple 64-bit expansion of the x86 architecture (in the same sense that the 386 had been a 32 bit expansion of the 8086).

Cheap, efficient 64 bit for the masses a-la SPARC / IBM / Alpha, with the advantage (compared to the Itanium) that it runs old 32-bit software FASTER than a comparatively clocked 32-bit CPU, and runs cooler.

The Itanium, if it survives at all, is probably going to become a niche product. Given the investment so far, I'd put my bets on being cancelled altogether within two years.

Some of you may not be aware of this, but this is <shirley bassy>"just another case of history  repeating"</shirley bassy>. Intel placed a huge bet on an OO processor back in the early eightes (I think it was called iAPX 342, but I'm not sure); It failed miserably, and was mostly the reason that Intel's answer to Motorolla's formidable 32-bit 68000 chip was the horrible 286 hack. (The 286's weird protection  model, btw, is some leftover from the 342).

Intel, however, got to its senses quickly and the 386 was mostly comparable to the 68020, taking performance lead back from motorolla with the introduction of the Pentium. Intel's market dominance has less to do with the x86's capabilities and more with general PC takeover -- but I think it's probable that if Intel didn't come back to being competitive with Motorolla, 68000 based machines (Mac, Commodore Amiga, Older Suns, and a few others) would have been much more popular - perhaps more so than the PC.

The sense in which history is not repeating is that in the past, Intel had to protect their market (PCs) from Motorolla - which is much easier than converting their market (PCs) to Itanium, which AMD does not need to do.

But we'll just have to wait to see what the future holds.

Ori Berger
Monday, September 30, 2002

And in a related development...

<quote>
The forecast for the United States remains cloudy, and signs of consolidation in the PC industry are everywhere. Earlier this year, for example, Hewlett-Packard and Compaq Computer, the top makers of personal computers behind No. 1 Dell Computer, merged largely in response to the slowing growth of the industry.

So far the response of the personal computer industry to its worst decline in history has largely been one of denial.

"People are walking around like members of the cargo cult after World War II," said Mark Resch, a partner at Onomy Labs, a Palo Alto, Calif., technology consulting firm. "They're just hoping the planes come back."

But some in the computing industry believe that the planes will never come back, at least in desktop computing.
</quote>

http://www.nytimes.com/2002/09/30/technology/30SPEE.html?pagewanted=all&position=top

J. D. Trollinger
Monday, September 30, 2002

No, Itanium won't kill them, it will just be a costly delay until they panic and get Yamhill up and running (the "lightweight" 64-bit chip). Once they get that into the channel AMD will once again be perceived as the low-end by corporate buyers.

AMD can't get no respect (in the corporate market) but at least they force Intel to be more competitive. If not for Opteron, do you think Intel and HP would bother with Yamhill, or would they just force the market to accept Itanium?

By the way, this seems to happen a lot at Intel -- they make a major strategic decision that the market doesn't accept and then backtrack. A recent example was when they tried to get everyone to switch to RDRAM. But this never seems to hurt them in the long run.

ns
Monday, September 30, 2002

"But some in the computing industry believe that the planes will never come back"

It's true that people have stopped buying computers but the reason is not as the article says that everybody's computer is now fast enough and so they have everything they need.

In fact, people I talk to and interact want something better than what they have now. Something a lot better. Because what they have now is frustrating, badly designed, unreliable and just plain doesn't work.

Here are questions I get asked all the time since people think I should know the answers since I'm an engineer:

Why can't the stereo play back music from the computer?

Why can't the computer record and play back TV shows?

Why can't the computer run for more than a few hours without crashing?

How come I have to spend 90 minutes downloading their 9,500 spam messages in my email just to see if any of their grandkids wrote me a letter?

How come people are sending me pictures of people having sex with animals and when I click on 'remove me from this list', I get ten times more pictures.

What is the deal with all these ad windows popping up everywhere and I can't get them to go away.

How come if I upgrade their computer, I have to replace my scanner and printer because XP drivers are not and will not be available.

How come if I upgrade their computer all my software stops working until I pony up several hundred dollars to 'upgrade' each app -- if it's even still made, and if its not how come I have to lose all my data as a price of upgrading?

--

It all starts looking like a scam to them and that it is.

The manufacturers have made a terrible mistake thinking the American Consumers are dumbsh*ts who will keep forking over their hard-earned cash for cr*pola. Well, here's a clue -- people have woken up to the con.

Computers don't do what they promise. They are lemons.

When computers provide value and not frustration to people, they will start buying. You can bet on that.

Sarain H.
Monday, September 30, 2002

Gee, I don’t think the problem here is that computers don’t deliver. In fact they deliver on so many fronts to day, it is incredible.

A computer today running windows XP is *way* more reliable then a old computer run windows 3.1.

Back then, things like drivers, and using a scanner were unbelievably UN-reliable. Today, I have a USB scanner, a digital camera, a portable cd-burner and even two printers plugged into my computer (one is USB). All of this stuff worked perfect right out of the box. In fact, I even use this same computer to record all my TV shows (I have a all in wonder TV video card from ATI). My palm also syncs into this computer.

Heck, I was recently traveling over seas and a friend of mine need to download some pictures from his digital camera. I plugged in a USB cable from my notebook I had with me right into his camera, the thing just poof...appeared as a external hard disk. No mess, no fuss, and it was fast. I copied a 128 megs of pictures in no time. Unplugged the cable and we parted.  I needed NO drivers, and did not even have to load, or find something specific  to that camera......it just worked...wow!

Users find new computers *way* more reliable then the computers they are replacing of just a few years ago. In addition, they plug in way more devices and options and these also work much better, and with MUCH less trouble.

My USB scanner runs circles around the old parallel port scanner. It works with more software, and it the install was so easy.  It also works while I running other software. The old scanner could only run by its self. Thus 3rd party OCR software for scanning in book articles etc. also just works right out of the box!

Also, there are some amazing reports of reduce support costs with windows XP. The driver support for mother boards etc. is amazing.

So, while the computer market is not perfect, it certainly has improved a lot, and continues to do so.

The real problem is not that computers are crappy...the problem is that the industry is maturing...and we don’t need to purchase a new pc every 3 years like we used to.

Any box with a 800 MHZ or more processor is *MORE* then enough for the average consumer. Doubling the performance of a 800 MHZ pc is not compelling enough for me to go out and purchase a new pc.

The real problem here is that industry is heading towards a replacement rate of PC’s In other words, it will be like the auto industry. That is still a healthy industry but it ant going to grow at 20% a year any more.

When you take a old world war 1 BI plane, and double the speed, increase the comfort level then the consumer will most certainly upgrade to the new air plane. Now that the plane has all these features, then is not much to go when the next version comes around. The next version will probably reduce maintenance costs (just like new air planes do, but they really are not faster then the old ones!).

Simply put...the industry has matured. The next model that comes out really does not offer compelling reasons to upgrade. It has nothing to do with crap.

Computers are amazing in terms of value, speed, and yes reliability.

Albert D. Kallal
Edmonton, Alberta Canada
Kallal@msn.com

Albert D. Kallal
Monday, September 30, 2002

Computer are still WAY too difficult to use.

I recently helped an elderly couple set-up their first computer.  It was an Apple Macintosh.  You don't realize how hard it is to use a computer until you try to teach a 70-year-old "computer virgin" how to use it.

Part of the problem was the fact that the computer came packed with about ten pieces of paper -- manuals, errata, cardboard CD holders, etc.  One single-page note said that the first thing you (the new owner) should do is install OS X.  So, that's the first thing I did.  Later, I found another packing slip that said that OS X is already installed on the computer, though System 9 comes up by default when you turn on the Mac.

Finally, it turned out that the computer was unstable running OS X because there wasn't enough memory.  Of course, by this time I had wasted several hours fussing with the thing (it takes forever to install OS X on top of System 9).

After I finally got the computer running, I had a lot of trouble connecting to the Internet.  The Mac was throwing error messages about an unrecognized modem, or some such thing.  Apple tech support was totally useless.  They kept telling me that it was Earthlink's problem, even though I had not even gotten to the point yet where I could choose an e-mail provider.  Ultimately, the computer spontaneously stopped generating the error messages, and I was able to get on the 'Net.

All in all, it was a huge pain in the ass to set up the computer.  It took me all afternoon, and involved several (pointless) calls to tech support.

My acquaintances would never have been able to set up the computer by themselves.  Ultimately, though, they *did* learn how to send e-mail to their son who works overseas.

Alex Chernavsky
Monday, September 30, 2002

Thanks Alex.

Albert,

You gotta realize that the experience for you as a software developer is a lot different from that of a non-expert.

And it is NOT because the non-experts are stupid, either.

Sure, it's easy for you as a developer to use your computer.

But the fact is that most people out there are NOT satisfied with their computers and DO NOT find they work reliably and DO NOT find the computer meets their needs.

Sure, with a lot of support they can get it to do email and word processing and even pull up a spreadsheet and you can do this all an a 800MHz machine, even though it crashes every hour or so.

But there is a lot they want to do that it can't do but should be able to do.

When/if the manufacturers and developers get a clue instead of going around thinking they know what's best for everyone, then sales will pick back up and people will upgrade. I figure they'll get a clue in well -- about never.

Sarain H.
Monday, September 30, 2002

Perhaps an example will suffice.

I notice that Joel sells a product to manage web sites.

It's major features:

ftp only is needed
don't need to learn a programming language

So, are people who use it too dumb to learn Perl or Python or PHP or ASP or whatever the latest thing is?

Jeez, what's wrong with people, there are at least 100 packages out there that do what Joel's does and they are free! Must be really stoopid people who need Joel's program as a crutch cos they don't know how to program.

Hm, but then how come so many of Joel's users have PhDs and/or are developers, etc? Maybe because they are smart enough to know its not worth it wastin gtheir time screwing around with something that is more complex than it needs to be.

It's the same with the computers -- there is indeed a market  opportunity to sell sots of reliable, intuitive, powerful computers to people who don't have unlimited time and energy to mess around trying to get their PoS 'puters to work. They aren't buying because this product is not available -- NOT because what they have right now is satisfactory.

Sarain H.
Monday, September 30, 2002

To paraphrase General Motors...

If kitchen refrigerators performed as well as computers and software do today the human race would have died out from food poisoning.

Simon P. Lucy
Tuesday, October 01, 2002

If computers and software worked as wel as refrigerators today, all they'd do is cool a can of softdrink ... and then some users would forget to close the case.

Just me (Sir to you)
Tuesday, October 01, 2002

The industry is finding out that these micro-incremental upgrades in software don't require hardware upgrades.  We have lots of spare cycles.  Come on developers, fill it up with useless features ;)  Then we'll buy hardware.  Once exception here is the peripheral manufacturers.  A lot of scanners and printers are version of Windows specific and do require a new one when you go to XP.

I think the next big hardware push will be Palladium.  I don't think Palladium will be about security as touted.  It will really be about DRM and the PC vendors making a killing on new "secure" computers.  Look for a flat market until Palladium.  Palladium will surely suck and necessitate moving to Palladium 2 on albeit incompatible hardware, thus necessitating a phone call to Dell et al for the second round of hardware upgrades. 

At least this is the track I believe will happen, as the software updates aren't doing enough to bolster machine sales.

What could be really interesting and screw up typical (current style pc's) is if the average home pc becomes something of an embedded device that "just works" on the web and doing simple things like word processing and maybe a spreadsheet and the usual mp3/music stuff. 

Maybe these will be disposable with some sort of removable media that houses all your files that you just insert into your new one.  No more "where are my data files stored?"  Just pop out the removable storage and put it in the new one.  That, in my opinion, is about the current capability and desire of the average home users concept of easy computing.

Here's to hoping that in the future when someone goes to the store and buys a new computer it doesn't necessitate a day of driver updates and two days of downloading security fixes to be functional and safe.  I am not sure we can get to that point with the old guard in charge.  We need innovation, lip service just doesn't cut it anymore.

ryan ware
Tuesday, October 01, 2002

>that "just works" on the web and doing simple things like word processing and maybe a spreadsheet and the usual mp3/music stuff

hmmmm...

Sounds like a PC just without the ability to add anything to it.

SM
Tuesday, October 01, 2002

"hmmmm...
Sounds like a PC just without the ability to add anything to it."

Basically.  You wouldn't have .dlls crapping on one another. 

Features would be email, word processing, rather than Outlook Express or Word per se.  These things are commodities and should be treated as such.  Word processing should be no different than TCP/IP.  It should be part of the "system"

ryan ware
Tuesday, October 01, 2002

[We need innovation, lip service just doesn't cut it anymore]

Nonsense, lip service is the backbone of this industry! We need to pay homage to those that payed homage to those before us with lip service. We can't go rocking the boat with innovation because developers will just poo poo it anyway saying stuff like "Why do we need X when we already have it!". Innovation is all around us it just takes time and money (not exactly a lot of this floating around) to change computing. How are you gonna sell them your shiny new snake oil when it really is just the same old snake oil except now you charge for a sip of yours instead of selling the bottle (read: service based applications) and you've been telling them that buying the whole bottle was the way to go fo so long? PCs have not been around for that long so it will take time.

Your idea of the new computer is as old as computing. The network is the machine.

trollbooth
Tuesday, October 01, 2002

Buying a computer for the average home user should follow the same thought process as getting a television, blender, or microwave, but for some reason people give computers mystical qualities.

ryan ware
Tuesday, October 01, 2002

*  Recent Topics

*  Fog Creek Home