Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
CJ
Changed their mind? They were StrongArmed...
Good one!
Greg -
This is good news for Intel but CJ was correct that Palm did originally say they were going with a TI version of strongarm IIRC. Nice to hear they changed their mind.
Greg -
Be it rightly or wrongly, the customer was shopping for MHz and QH depended on fooling him into believing that was what he was getting. AMD's argument seems to be that they were tricking the customer into making the decision he should have made, even if it wasn't the one he wanted to make. Sort of "the end justifies the means" thing.
EP
Doug -
AMD formulated QS, it was cheap, accepted, and worked.
And depended on the customer being deceived in order to work, if it even did work.
Economaniac -
unfortunately that isn't quite what Hector said. [etc etc]
I'm sure you're right but I thought the board would get a kick out of it anyway..
Hector Ruiz: AMD should break even in Q2 2003
http://www.geek.com/news/geeknews/2003Apr/bch20030502019843.htm
Doug -
The results are outstanding. Our Fab 30 performance is just outstanding. The yields, the quality are outstanding and we're very happy with that.")
I don't know how long you have ben following AMD but I remember Jerry saying much the same thing in the past, only to have to sheepishly admit later on that the yields were terrible all along and he had lied to shareholders.
The lawsuits are still pending in court. As Shareholders, lets hope Hector has more integrity than Jerry did.
FNF -
Sure the margins are lower but there is still profit there to be had. Profit is profit.
Spokeshave -
So, in summary, AMD has suffered the worst losses in their history since I got a new dog. Coincidence? Oh, by the way, Intel strongarmed me into getting a new dog.
The mind boggles as to what they did to the dog to accept you!
More seriously, you have a point but the real point was that it's hard to claim a marketing program was a resounding success under the circumstances of massive losses. You should have pointed that out to him, but you chose to criticize me for pointing it out. You are 100% predictable.
Doug -
I think the fact that QuantiSpeed so upsets the Intel fans pretty much confirms it was an unmitigated success. And you're right, incredibly low cost, at that.
However you want to qualify it, AMD has suffered the worst losses in their history since instituting it. Coincidence?
Haddock -
Considering they have been limited by how many they could sell, I don't think you can conclude anything with certainty. You could, however, spread some phud.
What I am presenting are facts. You can run the numbers for yourself if you like but you won't like what you see. As discussed yesterday, in Q4 of '01 Intel was capacity constrained, the market was hot and AMD didn't deliver. As you point out, we can't conclude with certainty but there were several explanations as to why. The least damning to their manufacturing capability was that they simply didn't realize they should make more for the busiest quarter of the year. Sort of an "duh... oops"! The other explanations are even less flattering...
YB -
FYI - I run my work PC from home with VNC too. Windows is now accessable remotely. Not as clean as Unix but it can be done.
Chipguy -
I don't disagree with anybody's definition here because the hardware capability has grown enormously over just the last couple of years. Some systems are kept in a back room because they are too old and slow to be of much use anymore but kept around just for software compatability. I know many testers run on SunOS Sparc 10s & 20s and many "workstations" are kept in a backroom just for offline vector compiles. Nobody would have one on their desk anymore because they're such dogs, while Intel Linux boxes are the fastest simulation engines in the world and under most engineers desks. The ubiquity of graphics display with Unix means it simply makes no difference where a box is. The only thing that matters is the graphics capability of your local machine and the speed of the ethernet connection between the two. For me, a waveform display on my home PC at 1600x1200 on a 20 inch monitor is just as good as being there.
Haddock -
Opteron is 193mm2, so they can do more than 4 million of those per quarter without exceeding the capacity that they were using for Thunderbird.
Don't be so sure. The defect density for SOI may not be the same as for bulk SI. It may be several times higher. If AMD had world class yields they'd get about 75 good Opterons per wafer while Thunderbird would get about 161. With a 3x defect density (not unlikely for SOI) it would drop to 33 for Opteron. Considering that AMD has never met the output potential for Fab30 one must question the yields there for even bulk SI.
chipguy -
You can run some jobs remotely on *other* people's workstations too and direct the remotely running application to use the Xserver logical display for your machine. But you still need your workstation at your desk.
I'm working from home running multiple xapps on both Linux and HP machines and viewing the graphics on my home PC using VNC over my VPN on cable. It's exactly like being there except about half speed on the screen. The beauty of Unix is that it makes no difference where the process is running. It can be displayed anywhere. Windows will never be a big time player without this capability.
Joe -
How does that work?
This does pose problems doesn't it?
If a process is set to not use iHT, why can't the OS still use iHT when running the first with a second unrelated process?
Joe -
Well you probably have the chance now to see what AMD can deliver. With Opteron's performance potential I'm sure AMD can sell a bunch. Let's sit back and watch and see what happens.
EP
Joe -
But let me ask you a related question. When did Intel go from one core to entirely different core within 1 quarter? P4 ramp ramp took 4 quarters.
I'm sure the answer is never, but that would require converting about 5 fabs almost overnight. Not going to happen. It's a lot easier to convert 1 than 5. The point is that regardless of the excus.... I mean reason, AMD has never produced at anywheres near capacity, even when the market was willing to buy.
Joe
The same way Intel was capacity constrained on Willamette, AMD might have been constrained on desktop Palimino (there a grand total of about 1 reference to the fact from The Inq). You may recall that this was the first quarter of sales of desktop Palomino, so capacity constrained on a brand new CPU could be understandable.
Sorry but it didn't add up. Simply taking the die size showed that AM could have made a killing if the simply produced at anywheres near capacity, assuming normal yield. AMD didn't deliver and the excuses were legend. The only one that didn't require yield problems was that AMD simply didn't think to start enough wafers.
Joe -
Don't be so sure. The last time I recall AMD to be production constrained was Q3 of 2000.
In Q4 of '01 Intel was capacity constrained by the oversized Willamette die. AMD was in a position to sell anything they could make. AMD delivered about 1/2 what their capacity should have been able to deliver. There was much debate at the time as to why, some claiming that AMD just got caught flatfooted and didn't start enough wafers and other's (myself included) saying nonsense, AMD's yields must be in the toilet. Since then, AMD has never been able to deliver anywheres near what they theoretically should be able to deliver.
wbmw -
Ridiculous. I know quite a few people who run Windows XP on 300-500MHz Pentium II/III systems, and problems are non-existant.
I have a few old machines running seti. One 300MHz Celeron overclocked to 450MHz and another 466MHz Celeron. Both run under Windows2000 and they run flawlessly.
Joe -
Another datapoint in the power consumption discussion is the head Newisys guy, who said that the actual power consumption of Opteron is in around 40 Watts (or 40s).
Time to comment here on datasheet vrs actual power numbers.
The fact is that the vast majority of units are well under the datasheet power spec. Additionally, as time progresses the power distribution at a given speed goes down significantly as fabs refine the process. OEMs know this because they characterize parts and practically nobody designs to the datasheet spec because not only will hardly any parts run that high during early introduction but essentially none will as time progresses. Companies such as Intel, and probably AMD as well are plagued by customers who don't design to the datasheet but still return parts that fail for timings or other reasons. It's a pain in the butt and it happens all the time.
Smooth -
Ummm, and personally, my guess is that there's not a lot of "strongarming" that Intel can do to IBM. Pleading, maybe.
It was a joke.
Intel replaces exec leaving to head Applied Materials
Intel Corp. Wednesday named Jason Chun Shen Chen vice president and co-director of its sales and marketing group, responsible for all Intel sales operations worldwide. Chen replaces Michael Splinter, executive vice president, who is leaving Intel to become chief executive officer of Applied Materials Inc., a Santa Clara, Calif.-based semiconductor wafer fabricator that, like Intel, has operations in Hillsboro.
http://portland.bizjournals.com/portland/stories/2003/04/28/daily32.html
sgolds -
This is all a sideshow, however, because Xeon is the target. Any marketing of Opteron that included Itanium comparisons would only be to build and reinforce the perception that Xeon is a has-been.
While it's true that Xeon doesn't beat or match Opteron on many benchmarks, it still does on quite a few and we haven't seen a new Xeon speed bin released in quite some time. I expect newer versions to make a much better showing. Perhaps not closing all the gaps but a substantial improvement is reasonable to expect.
Andy Grave -
You quoted a hack from Aberdeen group.
You're right. That wasn't an IBM quote.
Thanks for the correction.
EP
Doug -
Elmer, why did it take IBM so long to offer an Itanium-based product, do you think? Why now? Why not 9 months ago?
I really don't know but perhaps we should quote IBM:
"IBM had to have a strong Intel 32-bit and 64-bit product line in order not to confuse customers into thinking that there's some unannounced transition to AMD in the cards," Kastner said. And IBM's Itanium support trails that of rival Hewlett-Packard, "which is betting the company on Itanium," he added.
Personally my guess is that Intel strongarmed IBM into doing it...
wbmw -
Elmer, what do you make of this comment?
You said it best -
"Releasing a system now is a great testimony of how far Itanium 2 has already come."
Doug -
So you mean the first CPU in the Itanium family didn't cut it at all for IBM?
You are correct. The first version didn't cut it. Not even close. The second version does though and the third version, due shortly, is far superior to anything on the market, including Opteron. The point is that IBM is offering an Itanium now and thinking about an Opteron tomorrow. The momentum is building.
http://news.com.com/2100-1010-998878.html?tag=cd_mh
IBM's Itanium server goes on sale
By Stephen Shankland
Staff Writer, CNET News.com
April 29, 2003, 9:00 PM PT
IBM has begun selling its first major Itanium server, the four-processor x450, the company plans to announce Wednesday. And a larger sibling that will accommodate as many as 16 processors is due by the end of the year.
The x450 uses the second-generation of Intel's Itanium chip family, the Itanium 2, linked to memory and with a variant of IBM's Enterprise X Architecture chipset. It's the first major system from IBM to support Intel's strategy of selling 64-bit Itanium processors for high-end servers requiring large amounts of memory, instead of the more conventional 32-bit Xeon chips.
However, IBM initially will sell the system not for general use but rather for certain programs such as housing large databases or extracting useful information from sales data, said Deepak Advani, vice president of high-end xSeries servers for IBM.
"We're picking the applications we believe will deliver a lot of value and customer benefit," Advani said. "The Itanium ecosystem is still being developed."
The x450's 16-processor comrade will use newer processors and a newer chipset, though. It will employ the third-generation Itanium 2 6M, code-named "Madison," and the second-generation version of the EXA. The two-phase introduction parallels how IBM debuted the EXA on Xeon servers, first with the four-processor x360, then with the 16-processor x440.
"Our plan starts with a four-way (server) now, introducing the scalable line later in this year, right after Madison comes out," Advani said.
The second-generation EXA chip also will be used in the forthcoming x445 server with as many as 32 Xeon processors.
The arrival of Itanium in IBM's servers means a substantial expansion of the number of processor families in Big Blue's total of four server lines. Its xSeries servers use Intel's Xeon and Itanium processors, its pSeries and iSeries use the Power4 processor and its predecessors, while its zSeries mainframe line uses a different IBM-designed processor.
Advani acknowledges that Itanium means additional complexity, but he argues that too much simplicity also can be a problem when "one-size-fits-all" product lines fail to cover all customer requirements. "Digital Equipment in the 1980s and Sun (Microsystems) in the 1990s showed that model is not sustainable," Advani said.
The acceptance of Itanium into the fold means IBM must contend not only with more server designs but also with more higher-level software. IBM is working on Itanium versions of server software including its DB2 database program and its WebSphere business software, each of which will be available in versions for Linux and for Windows.
But some believe it's worth the extra trouble.
"There's no doubt that Itanium machines will begin to take material market share away from other architectures going forward this year," said Aberdeen Group analyst Peter Kastner, though Xeon-based systems likely will bring in more revenue for at least the next five years.
Timing is everything
IBM had to release its Itanium product soon, given the plans it announced last week to sell some servers using AMD's Opteron processor, Kastner said.
"IBM had to have a strong Intel 32-bit and 64-bit product line in order not to confuse customers into thinking that there's some unannounced transition to AMD in the cards," Kastner said. And IBM's Itanium support trails that of rival Hewlett-Packard, "which is betting the company on Itanium," he added.
IBM had expected to put x450 on sale "early next year," but shipping the product earlier would have had limited success given that Big Blue expects most of the servers to use Windows. Microsoft only last week released Windows Server 2003, the first Microsoft operating system to support Itanium.
Prices for the x450 begin at $25,999 for a single-processor model, but IBM expects most customers to pay about $38,000 for a setup that includes two 1GHz Itanium 2 processors, each with 3MB of high-speed cache memory.
IBM for a time sold a server using the first member of the Itanium family--the "Merced" product that arrived late and performed poorly. But that system was geared not for real-world use but for developers who wanted to begin the process of rebuilding their software for the new chip, Advani said.
Intel's Itanium debut has been hampered by the fact that the processors can't effectively run the older software written for computers with Intel Pentium or Xeon chips. Intel is trying to address the issue with new software that lets the Itanium emulate the other chips.
IBM waited until this stage to release its Itanium server because not enough software was available, Advani said. "Our customers told us early on that hardware without accompanying software is nothing more than a heater," Advani said.
While IBM talks about supporting Opteron they launch Itanium -
http://news.com.com/2100-1010-998878.html?tag=cd_mh
IBM's Itanium server goes on sale
By Stephen Shankland
Staff Writer, CNET News.com
April 29, 2003, 9:00 PM PT
IBM has begun selling its first major Itanium server, the four-processor x450, the company plans to announce Wednesday. And a larger sibling that will accommodate as many as 16 processors is due by the end of the year.
<much more>
Thread - The news just keeps coming
http://www.nwfusion.com/news/2003/0429ibmhp.html
IBM, HP to release Itanium 2 systems
By Tom Krazit
IDG News Service, 04/29/03
IBM is planning to announce its first Itanium 2 server this week, ending speculation about the strength of the company's commitment to Intel's processor, according to sources familiar with IBM's plan.
IBM's x450 server was shown last year at Linuxworld, but the company had not released any other details prior to this week. At last year's show, IBM said the server would feature between four and eight Itanium 2 processors.
An IBM spokeswoman declined to comment on the unannounced products.
IBM has its own chip for 64-bit servers, the Power4, which had led many observers to question IBM's desire to release a system based on a competing chip. The company did build a cluster of Itanium 2-based servers for The National Center of Supercomputing Applications last year, but had been otherwise quiet about its plans for the Itanium 2 processor.
Separately on Tuesday, HP said it would ship two Itanium 2-based workstations for the high-performance computing market that run Microsoft's new Windows XP 64-Bit Edition Version 2003 operating system. The workstations are the first HP products to make use of the new operating system that is tailored for Itanium products. A 64-bit version for Advanced Micro Devices' Opteron processor will be released later this year by Microsoft. The zx2000 and zx6000 will be available on May 1, and cost $3,298 and $4,896, respectively.
Spokeshave -
I feel quite certain that you are the one and only person who reads this board who actually took my post seriously.
And you took mine seriously??? For shame....
Spokeshave -
Intel conspiracies: It's all right here in black and white. For shame....
A perfect example of unsubstantiated accusations. You couldn't find a more perfect example. And you buy it hook line an sinker. For shame.....
Doug -
If you don't believe that Intel threatens and bribes customers considering AMD products, you're incredibly naive.
As a matter of fact I don't believe it and I think you are in desperate need to explain away AMD's failures. What better than a big bad evil mysterious Intel who threatens customers but never seems to get caught despite investigation after investigation?
Let's leave this subject be, though, as I'm not in a position to make public what I know. Yes, I know you'll at least claim that I don't know anything of the sort. That's your choice. I'll not post anything more on this subject.
Perfect!! The slight hint of hidden knowledge. The carefully planted innuendo. Hint! hint! Well I know a few things too that I don't post. If you can't back it up then don't play silly games.
Are people laughing at you behind your back? Have you been abducted too?
chipguy -
Don't take my word for it, download the design rules for any foundry logic process and I bet dollars to donuts that they have specific rules for laying out laser fuses.
I believe you! I believe you!
My only point was that Intel doesn't use them.
Doug -
Remember that there's a difference between knowing something and being able to prove it in court. Intel is skilled at operating carefully in this regard. They ask "Is it really important for you to be there?" instead of threatening outright. The point still gets across.
I guess this is something all companies in a position of power have to endure. Regardless of whether or not they are doing this there will always be people like you who will say they are and offer no proof whatsoever. There's no way to prove you aren't doing something. Rumors are embraced by those who have a need to make excuses and this is one of the most cherished.
Maybe we should start a list of rumors believed to be facts by the AMD crowd.
#1 Intel threatens customers who consider AMD products.
#2 Intel's compilers are designed only to produce good SPEC scores.
Somebody help me out here. I know there are plenty more...
chipguy -
I'll take your word for it but I've never seen one in production at any company. Intel has on die fuses, and many of them, that are programmable under test using standard prober/handler and tester. Details are not public.
chipguy -
One-time programmable with laser fuses. Can do that in any process without any changes or cost adders.
Laser fuses? Are you talking about an external laser to blast a fuse?
chipguy -
Consider the infinitesimal benefit of storing config data in flash. Now consider the cost of adding flash capability to a performance logic process. Intel doesn't even want the trouble of thick oxide devices for I/Os in its MPU processes. Flash? forget about it.
Sorry to break the news to you but that is how it's done. I probably can't go into the specifics as they are proprietary but a lot of information is programmable and stored on the device.