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.
This story is... OLD.
That too is a positive, as those customers contact their (surely 3rd tier) suppliers, and demand replacements, to be sold by AMD.
LOL.
Well, I'm sure the K7 will continue forward through 2005, but branded as the K7-based desktop Sempron, not Athlon XP.
So we still have eMachines AMD64 notebooks and Gateway AMD64 notebooks?
AthlonXP as a brand is being phased out right now, so it's not really a surprise that AthlonXP and XP-M stuff would be starting to wind down...
Now, will we see some Sempron stuff to replace it in the near future?
I'm not sure I'd call it a "heavy loss", especially if it's all AthlonXP stuff.
Oh! Allowed to take a vacation. I see.
Just checked where?
Then what did you mean by "released" ?
What was supposed to happen was a launch of new Opteron (90nm) systems from several vendors. There was no announcement, hence my question. If you see kpf's post, apparently the announcement was delayed, but kpf's source's company has released their new product(s).
Really? So some new Opteron systems are shipping, and just the announcements were delayed?
Did he have any thoughts on the new announcement date?
Have you had a chance to ask your source about his Dec 28 launch info?
The 800M USD is incorrect. Around ~40M USD. But regardless, catching them is a positive going forward, not a negative.
256K L2 is active, but it is likely that 512K L2 is on die.
It is a socket 754 part. Probably the 90nm, 1/2-cache disabled, socket 754 Winchester variant:
http://www.xbitlabs.com/news/cpu/display/20040923051438.html
The referenced CPUID in the above article (00010FC0h) decodes as 512K L2 on die.
But of course, this part has its AMD64 enabled.
It seems early for it to be an AMD64-enabled 90nm (true) 256K L2 part (palermo), but not impossible.
Regardless, it would appear that 2.4GHz 90nm parts are now shipping, with 2.6GHz 90nm Opteron parts rumored for February 14.
Gigabyte nForce4 ETA 12/30 (today) $129:
http://www.newegg.com/app/viewProductDesc.asp?description=13-128-268&depa=1
Asus nForce4 SLI-Deluxe, in stock! $275:
http://www.newegg.com/app/viewProductDesc.asp?description=13-131-517&depa=1
The 2005 date looks like a "shipping for revenue" thing. Per the official Intel slides, presented recently, Yonah + Napa was listed as "Q1 06".
It's the best play for AMD. They alone should have them for ~6 months before Intel. It will help them continue to take server share.
Intel's dual core 65nm mobile part, launching in Q1 06 with the Napa platform. It probably lacks iAMD64, however.
Price?
Not really. Except for the 1P server crowd, the price-sensitive would be better off halving the sockets and doubling the cores.
Yes, there will be dual core A64s. The CTO mentioned that dual core Opterons would debut in mid 05, with dual "client" (desktop) parts following closely thereafter in H2 05.
Here is the current AMD roadmap:
http://www.amd.com/us-en/Processors/ProductInformation/0,,30_118_608,00.html
"Toledo" is the code name for the dual core desktop part.
I don't know when they plan dual core notebook parts. Perhaps in H1 06.
No, there was an update in Q3, but that is not typical. The previous 3 quarters (Q2, Q1, Q4 '03) featured no such update prior to the earnings release itself.
There is little reason, once dual-core Opterons are available, to buy single-core Opterons. Especially if they are available at 2.4GHz X 2 as a drop-in replacement.
Dual 2.4GHz Opterons coming?
http://www.theinquirer.net/?article=20420
That was kpf's source. I suggest you ask him.
We may see a trickle of dual core from AMD/Intel in 2005, but basically it'll be a 2006 and beyond product.
Not in the server segment from AMD. They have every reason to make hay through Q1 06, when Intel launches a dual core x86 server product.
As far as high expectations, the analysts are only predicting some 70-odd cents of earnings from AMD for 2005. I would call that rather low expectations.
The other risk of sitting on the sidelines is a Dell-type event. You may find out about it in between trading sessions, and awake to the price up 30%.
And, of course, depending on your state, and your style of investment, there are the tax implications of short term versus long term gains.
Why would you expect that?
Monarch sells the 90nm 3500+ for $259.
http://www.monarchcomputer.com/Merchant2/merchant.mv?Screen=PROD&Store_Code=M&Product_Code=1...
The 130nm 3500+ equivalent is $272 from Monarch. (Both are OEM parts.)
I would guess that neither yields nor bin splits are an issue. Demand plus the transition from D to E revisions is a more likely explanation, to me.
The Soitec product is a globally strained wafer, to be used at the 45nm node. And the AmberWave technology is another form of strained silicon, to be sure, but not "DSL".
Everybody and their dog is using a form of strain, but not everyone is using the DSL method.
"Foreigners"? This board is multinational and not American.
It seemed a misplaced reprimand to me, but hardly a big deal, in any event.
Socket 939 parts currently allow 800MHz or 1000MHz HT, and the latter is achieved via 5 x 200MHz. Furthermore a change of the base clock to 250MHz would be problematic for the 1.2GHz and 1.4GHz HT frequencies which also make up the HT 2.0 spec.
The WSJ article made the approval claim. FYI, I do not work for AMD, nor consult for them in any way.
I assume that is a typo, unless it is possible to select 1/2-step frequency multiples of 200 MHz, and it is a steping E product with enough IPC increase to make it worthwhile to release a 100MHz faster clocked part.
2.6GHz 90nm Opterons coming in February.
That would seem to be the implication of the Inquirer article referenced here, once they work all the typos out of it. This is good news. It should allay any fears about the 90nm process scaling up to (and then beyond) 130nm frequencies. That they will release server parts at this frequency suggests to me that a 2.8GHz FX part is doable at the same time.
http://www.siliconinvestor.com/readmsg.aspx?msgid=20862152
http://www.theinquirer.net/?article=20252
This article repeats the same error regarding marketshare changes, but this one says the deal will be signed "next month" instead of "next week", as in this morning's article.
That highend is likely to grow smaller and smaller, as Opteron and Xeon solutions encroach upon it. 32 socket (64 core using dual-core parts) Horus-based Opteron single-image solutions should be available in 2005, for example.
If you read closely, you'll see the claim is really for SiS volume shipments, which is in fact a new event so far as I know.
This would appear to be the guy:
About Dave House
Dave House, president, chief executive officer and chairman of Allegro Networks, a privately-held data communications equipment startup headquartered in San Jose. House has over 35 years of experience in the high technology and communications industries. Most recently, House was the president of Nortel Networks, joining the company through Nortel's $8 billion acquisition of Bay Networks in 1998. During his 22 years at Intel, House is credited with creating and launching the "Intel Inside" campaign. Prior to Intel, House held management positions at companies including Microdata, Honeywell and Raytheon. Born in Muskegon, Michigan, House earned a BSEE from Michigan Technological University and an MSEE from Northeastern University in Boston.
It could be that his no longer working for Intel allows him to speak honestly about Itanium.
Elmer
"I think history will record Itanium as a failure," says David House, a former senior Intel executive who approved the original project.
I wasn't referring to the AMD exec trashing Itanium, but to the quote from the very Intel exec who approved the Itanium project trashing it.
I agree. The added latency of the extra hop makes it a non-starter.
Itanium deemed "failure" by Sr. Intel exec who approved it:
http://www.siliconinvestor.com/readmsg.aspx?msgid=20861828
H-P isn't spelling out exactly how much of the $3 billion is new investment and how much already had been allocated to Itanium-based projects. But Mr. Marcello said the company is definitely stepping up spending to encourage the development of more software for such computers. H-P also plans to set up a design center in Singapore to develop lower-cost servers based on the chip.
Intel also has continued to talk up Itanium's successes, including the chip's growing popularity in some of the world's largest supercomputers. Richard Dracott, general manager of Intel's enterprise marketing and planning group, argued that the chip has received "critical mass" in attacking high-end computer markets, while the company continues to promote its Xeon chip for high-volume applications.
Intel and H-P had much broader ambitions for Itanium when the joint effort was announced in June 1994. The chip was expected to first be applied in high-end computing applications, and later to be used in workstations, high-volume servers and even PCs, possibly succeeding the ubiquitous x86 design altogether.
The Itanium didn't become available until 2001, at least two years late, and didn't run popular software applications as well as expected. IDC, a market-research firm, estimates that only about 8,000 Itanium systems were sold for each of the past three quarters.
"I think history will record Itanium as a failure," says David House, a former senior Intel executive who approved the original project.
Itanium also has been affected by the rise of Advanced Micro Devices Inc.'s Opteron chip, which competes with both Itanium and Xeon chips. Though Itanium has some advantages in running high-end software, Sun, H-P and IBM are using Opteron in low-end servers that are likely to get more sophisticated over time.
"As time goes on, we see less and less of those benefits [of Itanium] being relevant," said Ben Williams, vice president of AMD's commercial server and workstation business.
Intel and H-P had much broader ambitions for Itanium when the joint effort was announced in June 1994. The chip was expected to first be applied in high-end computing applications, and later to be used in workstations, high-volume servers and even PCs, possibly succeeding the ubiquitous x86 design altogether.
The Itanium didn't become available until 2001, at least two years late, and didn't run popular software applications as well as expected. IDC, a market-research firm, estimates that only about 8,000 Itanium systems were sold for each of the past three quarters.
"I think history will record Itanium as a failure," says David House, a former senior Intel executive who approved the original project.
Itanium also has been affected by the rise of Advanced Micro Devices Inc.'s Opteron chip, which competes with both Itanium and Xeon chips. Though Itanium has some advantages in running high-end software, Sun, H-P and IBM are using Opteron in low-end servers that are likely to get more sophisticated over time.
"As time goes on, we see less and less of those benefits [of Itanium] being relevant," said Ben Williams, vice president of AMD's commercial server and workstation business.
http://www.siliconinvestor.com/readreplies.aspx?msgid=20861823