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.
Spokeshave -
Check Table 2-6. Multiply Vcc_max (1.420) by Icc_max (67.4). Clearly, fact checking is not one of *your* strong points.
Nice to see that you finally understand how to calculate power.
Wbmw -
I was looking at the SGI benchmarks you posted on the AMD thread and I was pretty amazed by some of those scores.
http://www.sgi.com/newsroom/press_releases/2003/june/altix_benchmarks.pdf
The SPECfp scores are simply out of this world and if your speculation about HP's SPECint scores proves out on Monday then that will be outstanding as well.
Unfortunately there were no 8-way Opteron scores or 4-way Madison scores but we can still draw some comparisons:
The SPECint_rate scores for an 8-way Madison were 2.13X a 4-way Opteron. 98.3 vrs 46.1
The SPECfp_rate scores for an 8-way Madison were 3.7X a 4-way Opteron. 164 vrs 44.2
When looking at the 32-way and above systems Madison shows amazing scalability.
drjohn -
Thanks for the link.
drjohn -
Even a 533 b steping will get you close to 3 GHZ with a v0ltage increase to 1.7v.
You have no way of knowing this. It could very well be a slow part.
Winners of Clustering Technology Excellence Awards Announced at ClusterWorld Conference & Expo
AMD Grabs Best of Show; Other Winners Include DataDirect Networks, IBM, Intel,
Linux Networx, Microway, Sun, SuSE, Topspin Communications
SAN JOSE, Calif., June 26 /PRNewswire/ -- ClusterWorld Conference & Expo,
the world's premier event on clustered solutions in commercial, research and
academic computing, has announced the winners of the first Excellence in
Cluster Technology awards at ClusterWorld Conference & Expo, held this week at
the San Jose Convention Center. The awards, presented in conjunction with
Linux Clusters Institute, recognize clustering products and technology
innovations by ClusterWorld exhibitors.
The awards program is the first industry effort to recognize the many
breakthroughs in clustering technology that have reshaped the IT landscape in
recent years and brought supercomputing power to all new markets.
Judging this year's entries were: Frank Gilfeather, Interim Associate Vice
Provost for Research and Professor of Mathematics and Statistics at the
University of New Mexico; M. Patrick Goda of Los Alamos National Laboratory
and architect of the Space Simulator and LOKI clusters; John A Goebel, a
Senior Developer at Stanford Linear Accelerator Center (SLAC); David Morton,
Technical Director of the Maui High Performance Computing Center; and Henry
Neeman, Director of the OU Supercomputing Center for Education & Research
(OSCER).
Divided into nine categories, the Excellence in Cluster Technology awards
represent major areas of innovation in the clustering market for GRID, HPC and
HA computing. The award winners are:
-- Best of Show: AMD -- Opteron Models 240, 242, 244
-- Cluster Management/Monitoring Solutions: Linux Networx -- Integrated
Cluster Environment (ICE)
-- Interconnects: Topspin Communications -- Topspin 90 Switched Computing
System
-- GRID Solutions: Sun Microsystems -- Sun ONE Grid Engine
-- Storage Solutions: DataDirect Networks -- SwiftCluster version 1.0
-- Developer Tools: Intel -- C++ and Fortran Compilers for Linux, version
7.1
-- Software Stack/Distribution: SuSE -- SuSE Linux Enterprise Server 8
-- 32-Bit Turnkey Solution: IBM -- IBM eServer Cluster 1350, version A
-- 64-Bit Turnkey Solution: Microway -- Microway Navion
-- Judges were so impressed by entries in this category that they
elected to award Honorable Mentions as well to Aspen Systems for
its Aspen Systems Glacier 12 Beowulf Cluster and to Racksaver for
its BladeRack.
"The Excellence in Cluster Technology awards are the new industry standard
for innovation in clustered systems," said Adam Goodman, CEO of QuarterPower
Media, publisher of Linux Magazine and producer of ClusterWorld Conference &
Expo. "These products and companies represent the very best solutions
available to customers who require clustered systems. The developers should be
applauded for their accomplishments."
About ClusterWorld Conference & Expo
ClusterWorld Conference & Expo is the world's premier event on clustered
systems for commercial, research and academic computing. The first conference
and expo will be held June 23 - 26 at the San Jose Convention Center in
California. For more information about how to attend, exhibit or become a
sponsor, please visit us on the Web at http://www.clusterworldexpo.com.
wbmw -
You bought the 533MHz FSB version, did you not? It's quite possible that your part could have been left over from B-Stepping inventory. The 2.4C GHz Pentium 4 chips tend to have a lot of headroom, given that they are based on a much newer stepping.
That's my guess too.
Paul -
First off, I am not a process expert. I just know a bit about it because my work interacts with it.
From a manufacturing perspective it would make perfect sense to offer two products from the same die. What do you think Intel does? They offer 3. P4, Xeon and Celeron are all more or less the same die now or just different steppings of the same one. I'm not revealing any secret here because anyone can buy them, crack them open and see for themselves. I'm sure the competition has... It makes production much easier. One product through the fab. You can differentiate at wafer sort on a lot by lot basis (just Thorton), or save those with half the cache bad (Thorton and Barton). It gives you great flexability. If they are different die then you have to guess right about demand 3 months in advance. With the same die you could wait until the wafers are finished to make your decision.
wbmw -
Would you normally expect a 1.6GHz chip to have such small headroom unless it was hand sorted or rescreened?
All parts are individually screened, fast ones and slow ones, and they have to have a speed guardband to compensate for the inevitable degradation that results from some long term effects to silicon. There is always speed distribution. I recently bought an Intel 2.4GHz P4 and it has very little overclockability, I was unlucky. I would think 2.4GHz is very near the bottom of the speed distribution, if not below it. Maybe it's the cheap memory I bought? You have to look at a large distribution to get a good feel for the overall process however this one single sample point is not inconsistent with the other data that says AMD has yet to get a good handle on manufacturing Hammer. Now if you got a 1.6GHz Hammer and it easily overclocked, that would be unexpected!
Haddock -
Amazing what you can do when the entire benchmark fits in on-die cache!
I'm sure there will be plenty of other benchmarks on Monday.
wbmw -
So it looks like 1.68GHz is the limit for full AGP compatibility (5%), while 1.76GHz is the highest possible in Windows without OC tricks (10%). With a boost to DDR voltage and a lower CAS latency, 1.8GHz was the highest achievable overclock (12.5%). It seems pretty clear, then, that Opteron has very limited headroom. A lot of work will have to be done before Athlon 64 can launch with full AGP compatibility at >2.0GHz.
We should see the results of more than 1 part before we draw any conclusions.
wbmw -
If HP gets the same percentages, they will have 1275 SPECint and 2076 SPECfp, higher than any other SPEC scores to date.
WOW! Can't wait to see those scores posted on Monday!
Joe -
Has Intel released the SPEC scores of Madison (and others in the family)?
Wait till Monday.
They are announcing now because Intel asked them to.
Sure Joe. They're announcing before Intel announces because Intel asked them too.
Brilliant!
Get over it.
Come Monday Performance will be redefined.
Neye
I think I recall reading a post or two at Aces in which it seemed like it was conventional wisdom that AMD simply couldn't engineer the same cache density that INTC could - even before the SOI issue happened.
For years AMD used a process technique called local interconnects which made their sram cells much smaller than Intel's. They licensed the technology from IBM. Intel didn't use it because it was hard to control and they didn't want the yield problems (so I am told). Something has changed and Intel now has the smallest reported sram cells in the industry. This accounts for the extremely large caches seen on modern Intel processors. I don't know what changed but obviously something did.
drjohn
Tbird was competitive its not sure A64 desktop will be
when introduced in September.
You can be sure it'll have a competitive QS+ rating!
wbmw -
Wow! A whole single [1 GB] DIMM experience.
Yes, but you get 64 bit addressing for free!
subzero -
re: Dell Introduces 64-Bit Server for High-Performance Computing Market; Dell Again No. 1 Provider of Intel-based Supercomputing Clusters According to New IDC Data
Great news! The momentum is growing.
JoeP
I'm not going to say that Hector is doing a bad job, but Jerry was running the company for about 33 years. He must've been doing something right.
Have you looked at the share price over the last 20 years? It's gone absolutely nowhere. That's what Jerry did for AMD.
Spokeshave -
In my opinion AMD did the right thing is using the QS strategy. While their position that MHz alone doesn't tell the whole performance story is valid, there was money to be made by fooling the Techonignorant with inflated comparisons. That's the choice they made and we'll never know how they would have faired if they had taken a more conservative route. We can only speculate. As you know, pricing is intended to maximize profits and not shear volume alone. The variable is the time frame you are optimizing. Short term? Long term? How did that line up with the expected Hammer schedule? The market looked like it would keep growing forever. Intel had been screwing up constantly. What did AMD expect Intel to do during that time frame? How did the joint venture with UMC play in here? It was all a house of cards built by Jerry and it had to work absolutely perfectly to keep from collapsing. It was far too risky and it didn't work perfectly. In perfect character Jerry had a fall guy waiting in the wings. You should respect Hector for the work he's done with what he was given.
Just one man's humble opinion.
Spokeshave -
You're as entitled to an opinion as I am but I disagree. The points you make are correct but in perspective they either make little impact in the great scheme of things (#1 & #2) or they blame Hector for the events that were set in motion long before he showed up. Sure Opteron should have been introduced earlier and at a faster speed if AMD was to compete effectively but the problem was promising too much in the first place. The technology just wasn't up to the task and AMD didn't have the $$$ for the needed R&D. Everybody in the industry could see that except for Jerry. AMD was committed when Hector took the helm. You make it sound as though it's Hector's fault that a bad plan laid out by an egomaniac didn't work out as promised. Hector didn't make those promises, Jerry did. You want to see some of Jerry's promises of the highest performance in the world and a 80mm2 die?
You can argue about pricing or QS ratings but it's equivalent to the old saying about rearranging the deck chairs on the Titanic. Hector inherited the ship when it was already sinking. Don't blame him for the gaping hole Jerry left.
Just one man's opinion.
Edgar --
I continue to believe it is wrong to blame Hector for the course that was set by Jerry. The decision to bet the farm on a risky process was not Hector's. The decision to insist on going head to head with Intel was not Hector's. Hector didn't sell off the PLD business to fund a vendetta. Hector didn't kill the uController business to use all the money to battle Intel. Hector didn't sell off the headquarters building in Sunnyvalle. Hector is trying his best to keep AMD off the rocks. Jerry sent the ship into troubled waters.
IBM unveils Opteron server
By Stephen Shankland
Staff Writer, CNET News.com
June 24, 2003, 3:43 PM PT
SOMERS, N.Y.--IBM on Tuesday gave supercomputing aficionados a glimpse of an Opteron chip-based system that is geared for high-performance tasks.
As earlier reported, the system, unveiled at the ClusterWorld conference in San Jose, Calif., is a dual-processor, rack-mounted server measuring 1.75 inches thick. It is expected to ship in the second half of 2003. The machine, which IBM first discussed in April, is designed to be clustered in large numbers to form a single, powerful computational engine.
Clustered systems are increasingly popular with customers that have intense computing requirements, such as pharmaceutical designers, automakers and banks. Intel's Xeon processor is widely used for these systems, but rival Advanced Micro Devices is working to find a place for its new Opteron chip as well.
"For those segments that have already embraced AMD, the enthusiasm for Opteron is very, very high," said Dave Turek, leader of IBM's new "Deep Computing" team, pointing to government laboratories in particular. "We have substantial pull from customers."
AMD's Opteron processor runs the same software as do Intel's Xeon and Pentium products, but it also has 64-bit extensions that allow it to speed some types of calculations and to accommodate more memory. To take advantage of those extensions, software must be rebuilt for the chip--an Opteron version of Linux from SuSE is available, for example.
One difference between IBM's high-performance computing business and that of its competitors is its program to rent out its own supercomputing centers. This program enables customers to tap into the centers when they have more number-crunching to do than their in-house machines can handle. The first customer for the rental service is Petroleum Geo-Services. On Tuesday, IBM announced another customer: GX Technology of Houston.
Turek said IBM's Opteron systems will arrive in early fall. They'll cost somewhat more than Intel Xeon systems but much less than Itanium systems, he said.
In addition, IBM will release Opteron versions of some of its software for high-performance computing, including its Global Parallel File System and its systems management product, Turek said.
IBM expects interest in the Opteron systems from customers in government agencies and universities, as well as in the life sciences, petroleum and automotive design industries, Turek said. To accommodate some of the demand, IBM this year will put some Opteron servers in its supercomputing center in Poughkeepsie, N.Y., which customers will be able to tap into as needed instead of buying systems on their own.
Big Blue's supercomputing work has led the company to trail niche leader Hewlett-Packard by a single system on the "Top500" list of the world's 500 fastest machines. To accelerate its work, the Armonk, N.Y.-based company formed an expanded supercomputing team--named Deep Computing--in April and last week hired longtime IDC supercomputing analyst Debra Goldfarb to be vice president of strategy and products.
While the Top500 list highlights cutting-edge supercomputing designs, most of the market is for less-exotic systems, Goldfarb said in an interview Thursday.
Buyers spent less than $1 billion last year on super-high-end "capability" systems built to accomplish specific goals such as modeling climate change or predicting nuclear weapons' physics, she said. In contrast, about $5.5 billion was spent on more standard systems, typically close relatives of servers sold for ordinary business use.
Cluster supercomputers typically use ordinary lower-end servers that run the Linux operating system. However, a Dell Computer cluster at the Cornell Theory Center that runs Microsoft's Windows OS has appeared on the Top500 list.
One cluster of Opteron-based machines, code-named Red Storm, is slated to be created by supercomputing specialist Cray for Sandia National Laboratories in the United States. On Monday, SuSE announced that Cray had selected its version of Linux to run on the machine. Under the deal, the German Linux seller will provide developer support as well as its operating system.
For its part, chipmaker Intel hopes to use the cluster market to further its comparatively new Itanium family of 64-bit processors, which compete with AMD's Opteron and established 64-bit chip lines from IBM and Sun Microsystems. On Monday, Intel gave details of a new version of its upcoming Itanium 2 6M "Madison" chip. This version will be tailored to dual-processor machines, which are well-suited to clusters.
Compared with the mainstream Madison chip--expected to debut on June 30--the dual-processor Madison will have less high-speed cache memory, according to company spokeswoman Barbara Grimes. The new chip brings to four the number of Madison varieties that Intel plans.
In addition to IBM, specialists in cluster computing such as RLX Technologies, TopSpin Communications and Platform Computing are expected to highlight advances in their own wares at ClusterWorld.
What the matter with you guys? Why hasn't someone posted this before me?
http://www.idg.net/ic_1322761_9677_1-5044.html
IBM readies new Opteron cluster system
Robert McMillan, IDG News Service\San Francisco Bureau
June 24, 2003, 15:20
This fall, IBM Corp will ship its first cluster server based on Advanced Micro Devices Inc.'s 64-bit Opteron microprocessor, according to IBM officials at the ClusterWorld Conference & Expo in San Jose on Tuesday.
The new system, which IBM was demonstrating to press, analysts and customers in a private booth off the ClusterWorld show floor on Tuesday, uses the same chassis as IBM's Intel Corp. Xeon-powered eServer Cluster 1350 system. The demonstration model was a dual-processor rack-mounted system containing two 1.8GHz AMD 244 Opteron processors. IBM executives say that when the system ships it will come in a variety of configurations based on AMD's 200 series Opteron chips.
IBM began work on these systems this spring, around the time the company first announced support for AMD's new processor, said IBM Vice President of Deep Computing Dave Turek.
The new systems will come with as much as 12G bytes of memory and will include 2 SCSI (Small Computer System Interface) disk drives. Like the Cluster 1350, they will ship with Linux and IBM's Cluster Systems Management for Linux software, Turek said.
IBM's Opteron demo comes the week before Intel is expected to launch the third generation in its line of Itanium processors, code-named Madison. Although Turek said his company plans to deliver an Itanium 2-based cluster system, he clearly was more bullish on Opteron.
"The simple fact of the matter is that the uptake on Itanium has been slow," said Turek. He blamed Itanium's slower-than-expected adoption in part on the fact that ISVs and customers have been reluctant to port their applications to Itanium's 64-bit instruction set. "It is a bad thing to do when you introduce a new technology to tell customers, 'by the way, everything's new,'" he said.
Both Opteron and Itanium can run 32-bit Intel applications, but because the Opteron processor appears to outperform Itanium when running 32-bit programs, and because developers can more easily change their applications to take advantage of Opteron's 64-bit instruction set, AMD's new processor is beginning to get traction in the market, according to Giga Information Group Inc. analyst Stacey Quandt. "If you just want raw power, the Itanium is a good selection," said Quandt, "but if you're looking at price performance and the ease of porting an application to a 64-bit architecture, Opteron is the right choice," she added.
"The marketplace tends to embrace technologies that are reasonably easy to digest. And the point of view of our customers is that (Opteron) is easy to digest," Turek said.
Pricing for the new systems has not been determined, but Turek said they will be priced "in the ballpark" of the Cluster 1350.
wbmw -
Apple G5 Power Dissipation 97W
Isn't that on 130nm SOI?
ChipGuy -
I thought it was "My dog ate my homework"?
Dr John
Not too surprisingly INTC is up today. With Intel holding to guidance and AMD missing I would think that Intel is gaining market share at a good clip and ASPs will hold.
Haddock -
Sorry, I didn't realise you were also short puts, thought you were talking about calls.
I'm short those too.
belgiangenius -
Oh, I'll blame Hector, as he fills his pockets with millions as shareholder equity goes up in smoke.
Hector didn't chart this course. Jerry did.
Belgiangenius -
Oh I have a small position in AMD; not enough for me to care much. But, it amazes me how much one management entity can screw things up so consistently and royally.
This course was set by Jerry years ago when he committed all of AMD's resources to a extremely risky plan. It was a personal vendetta and the shareholders pay the price. Don't blame Hector for this. It was handed to him and he's done the best he can with what he was given.
Haddock -
Re: NiceGuy -
The eternal optimist. I can't believe that guy.
Haddock -
I'm a little surprised at that. Surely your hopes are much higher now than they were a day ago?
Huh?
There's a long time to October, did you get a good price for those 3 months - 13 cents/month?
While I admit, it's not a killing, it still come to about 23% annualized.
wbmw -
Now, here's another prediction: watch for multiple press releases from AMD saying how they've gained market share in particular segments, and ones that list their design wins and make it seem like everything is hunky-dory. That's what happens every time AMD is down. The question is, will the same people buy the same spin again?
I've just got to go over to SI and see how NiceGuy is taking the news...
YB -
I'd like to check how Nokia and Motorola are doing. If they don't blame asia and sars wery well this time, then AMD is lieing.
I had hoped that with Jerry out of the picture AMD would adopt a new set of ethics. However you have to take their overall picture into consideration. They desperately need to appear viable in order to have any hopes of establishing Hammer. I can't blame them for waiting for the last minute to announce.
Once again the buyout speculation seems appropriate. Again I must say that whoever hopes to sell Hammer will need IBM process technology. IBM won't give it away for free.
wbmw -
you'll have to explain what it means to "roll" your calls or puts. I would love to hear more about that strategy.
Sure. I was short July $7.50 Puts. My hopes of them expiring worthless are not looking good so I bought them back to close @ $1.45 and sold Oct $7.50 Puts to open for $1.85. Net $0.40. I am still short the same number of contracts but I bought myself some time, or maybe I should say I sold some time.
wbmw -
I rolled some July $7.50 Puts out to Oct. Picked up $0.40.
wbmw -
SARS, indeed...
I'm waiting for the Israelis and Palestinians to blame their troubles on SARS...
Dew -
Zillions of posts about trivia and minutia. Then AMD lowers sales guidance by $100M (!) and there is a deafening silence.
I think we're stunned. I didn't see this coming.
ChipGuy -
That's what, next week? Good, SGI is a company needing a hot product right now and it looks like they got it.
The rumor is June 30th.
LOL, anyone care to make predictions about the composition of next year's TOP 500 supercomputer list?
Just a wild guess here but the company supplying the processors will start with an 'I' and will have more than 3 letters.
AMD Cuts 2Q Sales Outlook Amid Weak Asian Market
SUNNYVALE, Calif. -- Advanced Micro Devices Inc. (AMD) Tuesday lowered its second-quarter sales outlook, citing declines in personal-computer and handset sales in Asian markets, primarily related to the SARS epidemic.
The supplier of integrated circuits said it now expects sales of about $615 million for the second quarter, compared to a previous estimate of at least $715 million.
"The anticipated global sales improvement in the month of June did not materialize as we had anticipated," said Robert J. Rivet, the company's chief financial officer, in a prepared statement. "In particular, the decline in personal computer and handset sell-through in China and other Asian markets, largely related to the SARS epidemic, significantly affected AMD's sales in the second quarter."
AMD, the perennial underdog to Intel Corp. (INTC), has struggled under the weight of the tech industry's downturn, but has taken some dramatic restructuring measures, including several rounds of layoffs, and moved to a higher-priced mix of products to help it survive the slump.
The company plans to report its second-quarter results July 16.
-Yolanda E. McBride, Dow Jones Newswires; 609-520-7861
(The Wall Street Journal Online contributed to this report.)
(END) Dow Jones Newswires
June 24, 2003 10:24 ET (14:24 GMT)
SGI Altix 3000 Supercluster Extends Seismic Processing Capabilities At French Oil and Gas Giant Total
http://www.prnewswire.com/cgi-bin/stories.pl?ACCT=104&STORY=/www/story/06-24-2003/0001970821&...