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.
Yourbankruptcy -
Opteron has first to storm the blade rack-server market, trash Itanium and Xeon, then slowly, slowly start to compete with P4 on desktop. That will be a challenge.
Yes but right now on this benchmark it competes with a less than top of the line PIII. We were promised the highest performance processor in the world but delivering it is a bit more difficult than promising it.
Athlon64 3DMark 2001 SE Score
3dmark 2001SE scored 7913. That is not a lot as you should easily get over 10000 on a currnetly shipping 2700+. Here are the results in detail:
While this is anything but official it is just another bit of anecdotal evidence that says Athlon64 is a disappointment.
dlr_972 -
You see it exactly the same way I see it.
EP
Spokeshave -
Makes sense. Thanks.
Spokeshave -
It can't be interrupt driven. The notification is sent out by radio. If the device is not listening to the radio, it cannot get an interrupt.
Does the entire device need to be powered up to listen to the radio? My PC can powerdown and wake on LAN. Only the LAN needs to be powered up. What am I missing here?
Spokeshave -
The way I understand it, these products spend the vast majority of their time in "sleep" mode, waking occasionally to listen to the Base beacon to determine if there are pending data. If nothing is pending, then it goes back to sleep. If data are pending, then the device requests a send, and the time between the request and the receive is idle. In other words, idle is when the device is "awake" and listening, but not transmitting or receiving. I would expect that, of all the states, these devices spend the least amount of time in "idle".
A couple of days ago someone quoted an Intel spokesman as saying that power was saved by doing less polling, or some such thing. I don't know why they would poll at all. It should be interrupt driven.
Dew
No need for the condescension. The fact that I don't wet my pants doesn't mean I don't understand.
Dew -
Elmer: you are totally missing the point. Any cash spent buying shares on the open market to try to offset options dilution is REAL MONEY out the door that does not get expensed under GAAP. Cash so spent is not available for dividends, capital investments, or any other corporate purposes – it is gone.
So what? Companies buy back their shares all the time, especially when they are cheap. It may be money that's gone, temporarily but it raises the value of the other outstanding shares, so it didn't go for nothing.
If you evaluate a company’s financial condition over time based on a FIXED NUMBER OF SHARES OUTSTANDING (which assures an apples-to-apples comparison between different time periods), you must incorporate into your analysis the cash outlays needed to keep the share count fixed.
So then evaluate a company’s financial condition over time based on the NUMBER OF SHARES OUTSTANDING.
Growth Spurt: Intel-based Servers Scale Up
Not your grandfather’s commodity hardware platform
by Stephen Swoyer
http://www.esj.com/news/article.asp?EditorialsID=453
March 13, 2003
Intel servers are growing up in a hurry, as some major vendors are now shipping 16-way systems. At the same time, 32- and 64-way Intel servers from a handful of industry heavyweights are poised to challenge entrenched RISC-Unix systems in some markets, analysts say.
For years, Unisys Corp.—which markets its ES7000 line of Intel-based servers—was the only game in town, at least when it came to greater-than-16-way systems. But in the past 12 months, NEC has demonstrated a shipping 32-way Itanium-based system and Silicon Graphics Inc. is now selling a 64-way Itanium-based supercomputer. Meanwhile, Fujitsu Ltd., Hewlett-Packard Co. (HP), and—as of last week—IBM Corp. have all made noises about marketing, respectively 128-way, 64-way and 32-way Intel-based servers.
More ....
Dew -
The problem is that money does change hands – and lots of it. Companies like INTC spend billions of dollars each year buying back shares in the open market to offset the dilution caused by stock- option exercises. THIS IS A REAL CASH OUTLAY and is the true economic cost of option-based largesse.
I disagree. The shares bought today on the open market are to provide for the options granted today at essentially the same price as the ones just purchased. There is no real cost because if the shares are eventually exercised, they are exercised at more or less the same price as they were purchased for. It's a push.
Yourbankruptcy -
That is an outrage.
CJ -
What type of organic, low-k is Intel going to use for 90nm?
I really don't know. Not my field but I think the term should be "using". <G>
Sgolds -
Smooth asked about process, not processor speed
Yes, he did, you're right. Should we quote gate delay?
don't know the answer off the top of my head, but IBM typically is the leader in advance process design.
That used to be the case. Today Intel and IBM are considered head to head. Additionally, Intel has to target High Volume Manufacturing while IBM does not so. IBM doesn't need to reach the same level of manufacturability.
Sgolds -
He's probably talking about his 401K
Paul -
Considering the pipeline length, sounds like they're doing OK.
Yes, they've just about caught up with a .18u Aluminum Itanium in SPEC...
Smooth -
Anyone know IBM's highest process speed?
So far, I think the fastest processor they've posted SPEC scores for is 1450MGz. This is the great SOI process that is going to save the world from the evil empire.
http://www.spec.org/osg/cpu2000/results/cpu2000.html
Neye -
a spell checker isn't available until we pony up the $'s to become a paid subscriber to iHub
Ahhh. That explains a lot. My apologies.
Paul -
Elmer, give him a break! English is not his native language. I think he communicates pretty effectively, if grammatically imperfectly. Besides, "The spell checker is there to help yo" - perhaps you forgot to use the spell checker? ;)
I think he does just fine, considering, but awailable had to show up 3 times as a misspelling.
BTW, "yo" must be an acceptable spelling for Rocky Balboa types. Yes I did the same thing didn't I!
EP
Yourbankruptcy -
You can think whatever you want, but awailability of Athlon 2400 is better than any P4 model. Even the most popular P4 2.53 Ghz is less awailable. They are definitely taking back marketshare. It seems true that regular XP was adjusted for better output sacrificing binsplits.
It "available" not "awailable". The spell checker is there to help yo. Don't ignore it.
You offer no basis for your belief nor definition. Available on pricewatch? Where? How do you know? Maybe there's plenty around because it's not selling?
Pentium M die size:
I would estimate it to be about 90mm2.
Paul
earlier development units.
The point was that if AMD is dependent on a 64bit version of windows that still isn't here in 2003, what would have happened if they met their original schedule of EOY 2001? I think the answer is that AMD never expected to meet that schedule and they were just blowing smoke up the butts of those who were willing to listen.
YourBankruptcy -
Tell your broker to authorize you for put writing or you'll take your business elsewhere. That should work.
EP
Paul -
One reason I can think of, Elmer, is that they need to push development of the software base for the processor, which is what most of the systems now are involved in. Sure it's great with 32-bit software, but it's even better with 64-bit.
So what would have happened if they hit their original schedule and released Hammer end of 2001?
EP
Sgolds -
Today I am working on my web site. My domain is up and running well, and I have some supplier contracts
Awesome! Let us know when you go public!
EP
Sgolds -
do you still think that Hammer manufacturing is unproven? With this many vendors at Cbit and with so many systems exhibited, I am convinced. Hammer manufacturing is proven - vendors don't show off so much product if they can't get enough to fill the channel.
Manufacturability is unproven by definition as they have not yet demonstrated the ability to produce this device. Why are there no production Hammers, only demo systems, if it is manufacturable? Why has it slipped almost 1.5 years if it is manufacturable? Why was Athlon64 pushed out to September (at least) if it is manufacturable? Another poster here claimed AMD's 1,000 demo systems were fully functional and full speed. If they were manufacturable then why weren't those production systems? It seems obvious to me AMD hasn't as yet produced them because AMD, as yet, can't produce them.
This doesn't mean that Hammer can not be manufacturable tomorrow, but forgive me if I say "show me the production Hammers". So far AMD hasn't.
Note: If anyone is going to quote me on this please don't chop out the operative words and claim that I said something I didn't.
EP
Yourbankruptcy -
it's very hard to play options on $5 stock. You can't even establish a spread. Your puts look great. Your calls most likely will be assigned, so you already know your profit. Are you comfortable with that profit? If yes, then your position is optimal.
No it's easy. You can write many more contracts for the same margin as you can for a $25 stock. BTW, I may roll those calls rather than allow assignment.
For conservative players I strongly suggest to go long and write July $5 calls for $1.30. That will make your entering point at $4. But profit will be very modest. I feel that Opteron is for gambling, not for investing. Right now I'm plain long.
You could write July $5 puts and pick up $0.85 without spending any cash. That would put your entry at $4.15 if assigned.
Thanks for sharing your thoughts. We don't all root for AMD but we all like making money.
EP
Spokeshave -
It really dedends on what systems you look at.
Sure, it always does but I figured I would compare best with best. Would you do otherwise?
EP
Economaniac -
What I would like to see is a little more discussion on how people are investing. In AMD or the market in general. That seems quite on topic for an investment forum.
I'm currently long AMD shares, short $5 March Calls, $5 April Calls and $5 July Puts.
I'm trying to decide if I want to roll my March Calls or allow assignment. I'm inclined to roll but haven't decided yet.
Anyone else actually investing or just reading the press releases?
EP
KPF -
On your link I find CPUs from AMD, IBM, Fujii, Sun and the likes - however not for those of Intel. Could you provide a link for the lower-grade CPUs as well pls?
http://www.spec.org/osg/cpu2000/results/cpu2000.html
AMD posts SPEC scores for XP3000+
As expected they come in a little behind a 2.66GHz P4 in INT and about equal to a 2.0GHz P4A in FP.
http://www.spec.org/osg/cpu2000/results/res2003q1/
Economaniac -
I would not like to give that up unless it proves to be a serious problem. If that is not possible I for one will return to RB.
I'm inclined to agree with you. Let the members define what is appropriate. Obviously vulgarity and personal attacks are out as well as spam but so far I don't see anyone complaining but the self anointed moderator.
EP
wbmw -
Clearly, if Intel is losing market share, it is to these two bulls, not to AMD! Any theories as to how that is?
Sorry, no idea.
Flash Sales estimates for 2002
This was posted on the Intel board showing Flash sales for 2002. Column 1 is 2001 position and column 2 is 2002 position. This data seems to indicate AMD lost significant market share in 2002.
Intel's revenue grew 2.6% as their market share increased slightly while AMD's revenue dropped by 35%. AMD's market share dropped also.
http://www.siliconstrategies.com/story/OEG20030311S0031
2001 rank 2002 rank Company 2002 sales 2001 sales % change % of market
1 1 Intel $2,058 million $2,005 million 2.6% 26.1%
8 2 Samsung $1,213 million $374 million 224.3% 15.4%
5 3 Toshiba $842 million $547 million 53.9% 10.7%
2 4 AMD $727 million $1,120 million -35.1% 9.2%
3 5 Fujitsu $657 million $881 million -25.4% 8.3%
4 6 STMicro $609 million $682 million -10.7% 7.7%
6 7 Sharp $420 million $540 million -22.2% 5.3%
7 8 Mistubishi $368 million $468 million -21.4% 4.7%
10 9 Hitachi $304 million $297 million 2.4% 3.9%
11 10 SST $275 million $294 million -6.5% 3.5%
9 11 Atmel $163 million $340 million -52.1% 2.1%
12 12 Macronix $70 million $88 million -20.5% 0.9%
13 13 Winbond $60 million $55 million 9.1% 0.8%
15 14 Micron $37 million $40 million -7.5% 0.5%
14 15 Hynix $35 million $55 million -36.4% 0.4%
19 16 NEC $26 million $4 million 550.0% 0.3%
Spokeshave - OT
This heat can cause great destruction well outside of the destructive range of the shock wave and/or Mach front. However, it can be very subjective to quantify and is not accounted for in the "tonnage" of a nuclear weapon.
This makes sense. The heat radiation would be much more destructive in an area of wooden buildings compared to masonry ones when outside the direct blast zone.
EP
Greg -
jeez Louise, do you have to turn everything into a debate? <VBG>
No and I agree that he has done a fine job.
EP
Spokeshave -
On bombs- How do you factor in the heat generated by a nuclear bomb? Surely it is part of the destructive power but not the blast power.
Greg -
Au contraire! IH Admin (Matt) absolutely hates spammers and those who try to use multiple aliases. He has his ways to discern those who try to use more than one alias and he's really good at it. He will fry anyone who tries to do that.
I agree that it's a bad thing to use multiple aliases but I don't see how you can completely control it. You can track IP addresses and put cookies on people's machines but that's not conclusive. There are instances where individuals have multiple machines and multiple IP addresses. Additionally I know of cases where multiple people use the same machine so how can they tell for certain who's at the keyboard?
EP
Not a Short
The man simply boiled down your statement to its simplest form.
Only for a simply mind.
Spokeshave -
I find it interesting that the 3.06 Xeon datasheet lists the maximum power as 101W, while the P4 at the same speed lists only the Thermal Design Power, but if you multiply max current by max voltage, you still get 101W. The two processors have the same core. There was some heated discussion some time ago about what the max power dissipation of the 3.06 P4 was. It now looks like those who believed it was 101W were correct.
No they were wrong. Do you need me to point you at the datasheets that shows different Vcc and Icc for the 2 products? Why the difference you might ask if you were careful enough to actually read them. The Xeon probably needs stronger buffers on the FSB pins because it drives more capacitive loading.
Yourbankruptcy -
explain - that still doesn't fit into 50 Watt blade servers or into 100 Watt dual-cpu blade servers, or I'm missing something?
No you're not missing anything. It won't work in the application you describe.