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.
NVIDIA Business Platform: A competitive desktop alternative to Intel?
http://www.anandtech.com/IT/showdoc.aspx?i=2730&p=1
Joe
SmallPops,
I seem to recall several reports of numerous senior IPF engineers leaving Intel over the past few. Do you think the net effect has begun to impact the roadmap?
I can be the other way around: It may be that the Itanium roadmap is causing the employees to leave.
Joe
chipguy,
So that leaves what, over 340 IPF engineers at Intel FTC.
Hardly the turn the lights out scenario painted by bob10.
The fact that only 8 guys would follow a senior architect
says a lot about Sam, AMD, and the attractiveness of
working on x86.
What it tells me is that the first 8 guys used up all of that fancy resume paper in the office, and the remaining 340 are waiting for the next shipment.
Joe
chipguy,
Hmmm? I remember Red Storm as a development project
paid for by Sandia Labs who chose the processor. Alpha
had just been killed so there was a lot of misplaced blame
on Intel and IPF in some of the national labs.
I am sure you understand that these people didn't want to bet their careers on another processor that is facing risk of cancellation.
Joe
AMD_Dude,
Zowie, so each "cabinet" will house 128 quad core cpus, so on a 8 socket motherboard, will house 16 motherboards per cabinet. Man, what a system
32 (2U) systems with 4 sockets each is also a possibility.
Joe
smallpops,
CJ, it is very interesting that the article says "will start to take delivery on". Is it possible that AMD already has QC running at 2.6 GHz that it can start supplying for this system?
It is an interesting wording, but I would not jump to conclusion on current availability of QC - availability of anything that can be installed. Possibly some kind of pilot chassis or component. The processor appears to be K8L, and it may just be in process of getting the first silicon.
Joe
combjelly,
Impossible. Didn't wbmw prove to us that quad core couldn't break 2.4GHz?
I guess Cray must be wrong then...
wbmw,
2.6 GHz QC - likely to available in 2007 - sounds pretty good too. It may even be K8L based.
Joe
Energy lab to run petascale computer
By Joab Jackson, GCN Staff
NEWPORT, R.I.—The Energy Department’s Oak Ridge National Laboratory will start to take delivery on a computer at least as three times more powerful than any now in operation.
The system should be operational by 2008, according to Thomas Zacharia, ORNL associate lab director for computing and computational sciences. Zacharia spoke at the High Performance Computing and Communications Conference this week in Newport, R.I.
Cray Inc. of Seattle will supply the system, named Baker. It will run approximately 24,000 2.6 Ghz quad-core Opteron processors made by Advanced Micro Devices Inc. The nodes will be housed in 187 liquid-cooled cabinets. The system, which is still in early stages of design, will have either 187 or 400 terabytes of working memory (depending on the cost of memory modules) and from one to 11 petabytes of storage.
Such a computer, if operational today, would be considered the world's fastest; it may even be the first to break the 1 petaflop limit. A PFLOP is 1 quadrllion floating-point operations per second.
According to the Top500.org list, today's most powerful computer is the Blue Gene/L, run by Lawrence Livermore National Laboratory, another Energy Department outfit. That system currently runs at about 350 teraflops (350 trillion floating-point operations per second), though the lab and IBM Corp. plan to expand its capability to 1 PFLOP, also by 2008.
The laboratory will rely on the Tennessee Valley Authority for the power needed to run this behemoth. The power company is building a 170 megawatt substation to support this and other ORNL projects.
Like current ORNL systems, this one will be used by some of the country's best scientists to further explore the limits of their fields. Researchers must demonstrate the need for using the computer, and the lab will look for those jobs that are too large for other systems.
"We're entering new realm of scientific discovery, where we are looking at things at a much smaller scale. Computing expands upon our understanding of the natural world," Zacharia said. "Greater computer capacity allows researchers tackle more complex problems."
http://www.gcn.com/online/vol1_no1/40250-1.html
Kinda like harnessing nuclear fusion, eh?
Yeah, except that constantly remains 20 years in the future (IIRC).
Joe
chipguy,
Let's go with your number for the sake of argument. Intel
has been selling the 8870 chipset for about 14 quarters
now, give or take. That is $5.6m total. You really think
three 250 nm or 180 nm logic process masksets cost
over $5m?
If it took them 10 tries to get it right, maybe.
BTW, mask sets are only a small fraction of the cost of designing, manufacturing and supporting a chipset.
Joe
wbmw,
interesting collection of parts. But regarding:
Just how many design teams do you think AMD has?
I think you may be slightly off base. It does not take a 2 design team to come up with full cache and half cache version of the same chip. Nor does it take 2 design teams to come up with dual core and single core version of a chip. It took a design team to come up with Rev E which enabled DC.
It does take a design team to come up with Rev F, which has various capabilies / options for implementation. The actual implementation is just handed off to another team that implements the mask sets.
As far as design teams are concerned, there is one for K8 refinementL G step (dumb shrink?), H step (K8L?), another one for next gen notebook, and another one for next gen server. Those I would call design teams.
There may be a similarity between Dempsey -> Woodcrest and 90nm -> 65nm Opteron QC, in the time on the market, but Intel has more resources, is more desperate, and has all eggs in one basket as far a NGA, so there is a stronger need for a plan B. 90nm QC Opteron would be a plan A+, which is something AMD is not in a desperate need of.
Joe
cruzbay,
OK, let's see: ~35K Itaniums sold last year, at (you tell me) an average of $1500 each, comes to a grand total of $52 million dollars. If the chipsets were $400 each, that adds a whopping $14 million, total revenues (skipping your nebulous "toolchains and libraries") coming to $66M.
Most of Itaniums are sold by HP and HP has its own chipset. I think some 80% are sold by HP. If another 10% are sold by SGI, you have 10%, or 3,500 Itanium sold by others. If the average # of CPUs in these boxes is 4, there are less than 1,000 Itanium chipsets sold by Intel. With $400,000 revenue per quarter, Intel is losing a bundle. It is doubtful that Intel makes enough money for the mask sets. Intel may be losing $10 for each $1 of chipset revenue.
Joe
wbmw,
They don't have the resources to do three quad core projects (90nm, 65nm shrink, and 65nm K8L), but they might have the ability to do two if it means competing with Intel in the "quad core race". The question is which two will it be?
Unlike some of this thread, I don't think there will be a 90nm quad core CPU from AMD. While I think it would be feasible, the payoff for such a project in terms of resources expended is not great.
Regarding resources, it looks like more are coming onboard:
http://www.realworldtech.com/page.cfm?NewsID=359&date=03-28-2006#359
It seems that some are jumping the Itanic even before the inevitable collision with the iceberg.
Joe
mmoy,
that 4200 offer you mentioned was a great deal.
BTW, I just checked out Dell site recently (for some reason the USB hub on my Dell 2405 monitor was acting a bit flakey) looking for drivers or other possible solution. I noticed that they sell these 24", 1920x1200 LCD panels for $799 now.
That's an incredible price, if you asked me. I wonder what Dell's cost is on these, or if Dell is trying to chase Top Line figure by selling components at cost...
Joe
Bobs,
Good points. But, BTW, I don't see any white nights on the horizon, the least likely one being something wireless in order to provide a medium for very high bandwidth applications.
Joe
morrowinder,
Your just ranting and insulting now and knocking down your "it's the same install as win 32" strawman. Have fun playing with your doll
No, I am commenting on something I have done - installing 64 bit Windows. You are claiming to be authority on something you have not done. I will leave it up to others to judge who has more credibility.
Joe
mas,
25W Turion X2s
This can not possibly be true. I know because wbmw told me.
Joe
mas,
The reason for the delay was given as "problems with a copy protection format" for its Blu-ray high-definition DVD player, preventing a planned release this spring.
There would be some poetic justice in Sony losing the PS franchise (to MSFT XBox) because of their warship at the DRM altar...
Joe
morrowinder,
You said:
I applaud MMOY for actually getting it to work, it is not easy.
In fact, Windows x64 installs / loads exactly the same way Windows XP does. The installer appears to be the same, certainly from the user's perspective.
Then, you posted:
I have a 64 bit capable Pentium D I recently built but I find no credible reason to even think about loading such an OS till VISTA ships.
So you clearly loaded the OS (32 bit version) with identical installation. So why are you applauding MMoy something that is "not easy", while you, a Dell gamer have managed to do? MMoy has clearly shown his competence in the computer area, and installing an OS is not something he can do in the background while he does other productive work.
I am sorry, but this is not the first disingenuous post from you, and I am sure it is not the last.
Anyway, next one of your twists is that it is not easy because my mother could not do it. It's like you are piling BS on top of BS.
Finally, what possible purpose would there be in installing it for 95% of the users out there? Bragging rights lol.
The progression of 32 bit OSs in the NT code base was NT Server, Workstation -> 2000 -> XP. So surely, those who jumped early on the 32 bit bandwagon did it for the bragging rights as the early adopters of 64 bit - NOT.
Anyway, maybe I give you more credit than you deserve, but I think that you are just playing dumb, that you are not really dumb. Something is telling me that your posts on this board are just part of some sick game you are playing...
Joe
Pete,
oops, thanks for the correction.
Joe
morrowinder,
It doesn't have anything to do with how hard win 64 is to install. I think the marketshare numbers(<5%) and microsoft's own extremely modest expectations and marketing speak volumes on that topic. I applaud MMOY for actually getting it to work, it is not easy.
Complete nonsense. You can load the entire x64 version of Windows, and not even notice that you are installing 64 bit, rather than 32 bit.
Have you ever tried it (loading x64) if you claim it is not easy, or are you BSing (as usual)?
Joe
jackthex,
I've been lurking for around 8 years and I know that perhaps while they appear logically challenged it is obviously intentional. I have kept migrating from SI, to the moderated SI, to Ihub, to moderated ihub, to the strictly moderated ihub and it is just getting pathetic. Why Keith puts up with this crap I don't know.
The moderators have a limited set of weapons to keep the thread from degenerating. You can remove post for following reasons:
- Personal Attack
- Duplicate (double post)
- Spam, Advertising, Promotion
- Vulgarity
- Author asked to remove
- violation of privacy
- Threat
Being plain stupid, intentionally flooding the thread with nonsense doesn't clearly fit under any category.
Joe
Alan,
I think one reson with the delay is that you need to have an appropriate "vehicle" a core revision that makes enough / big enough changes in the same area as the new feature, so that you can add the feature. In this particular case, AMD has not done any (or enough) changes in area where micro-op fusion would go to, so it didn't get implemented.
If K8L only touches FP / SSE, it may not be implemented again.
Joe
mmoy,
SQL Server trial is even better. It is 180 days, and you get a full 64 bit or 32 bit Enterprise version. After my 2000 version went down the drain with a hard drive on a development machine, I loaded the trial, which gives me a plenty of time before I have to shell out the bucks for the real thing.
Joe
mmoy,
BTW, CompUSA has an HP X2 4200, 250 GB HD, 1 GB memory -
expandable to 4 GB, TV Tuner and a few other goodies for $699/
Wow! Great price for a great computer.
BTW, (I posted this only on SI) a friend of mine got a great looking HP dv8xxx, 17" laptop for $994. 64 bit Turion version (future proof).
Joe
Alan,
That said, the time from finding something like macro-ops fusion to implementation will be several years.
It is entirely possible that AMD has been working on it for several years already. I remember posting on aces at the time Banias was released that it would be a good idea for AMD to copy. I am sure that AMD has people who thing of those things before i can think of it.
Having said that, I doubt it will be in Rev F core. A little more likely for K8L.
Joe
chipguy,
Sounds interesting. But if the FP resources effectively doubled, wouldn't you expect somewhat bigger gain of SpecFP?
Joe
Tenchu,
If it makes you feel any better, I sold my Qualcomm shares before their explosion to make the down payment on the first home. I could have paid cash for the whole thing if I waited a year or so...
Joe
Can you enlighten me? I have not seen specs of what instructions Intel added. I was going by second hand one liner. My assumption was that they were adding 128 bit precision floating point.
Joe
Tenchu,
Only AMD processors? How about my Northwood? How about single-core Prescott?
I am presuming they are innocent, because it is within a realm of possibilities that a multitreaded application may need more than one processing cores to achieve certain level of performance. I would agree that it is dumb to hardwire this in the EXE. So innocent and dumb.
It isn't AGAINST AMD's processors per se
How can you possibly say that, after the code has been published?
But even if it wasn't, it's not going to stop AMD from wanting in on every deal that Intel makes with its business partners.
Sure, every exclusive deal that Intel strike, that specifically excludes AMD processors is illegal. If the test is on the merit of the processor, it can be legal, if it is a test for "GenuineIntel", it is illegal.
Joe
wbmw,
Gelsinger just let slip that Core micro-architecture has 128-bit SSE capabilities.
That's good news. Hopefully, AMD will follow up quickly. I didn't like the idea of losing 80 bit capabilities.
Joe
Durl,
Judge has to spend a couple of days reading and studying and cite checking only to find out they are meaningless, paranoid drivel?
Hmm, didn't you just chide me for trash talking. Shouldn't you, as a high priest of the law keep mum until all the facts are in?
How do you think he will rule on the next AMD motion?
Which one?
Joe
Joey,
Planning on holding your AMD at these levels or buying more??
I am holding most of my shares. A tiny portion got assigned at the expiration, and I decided not to chase it, so I have about 9/10th of my max AMD share holding.
Joe
wbmw,
I think you are right. I think I misinterpreted some of the rumors.
Joe
Durl,
I can read the code and I can read Clayton Act. My conclusion is that the code alone is the smoking gun.
Joe
Temp,
How do you think the current Skype story may affect the outcome of AMD complaint? Do you think it may make a material difference?
On one hand, it is just a complaint #69 or so, but it also shows ongoing conduct.
As far as immediate impact, that may be limited, since it may take a while before the subpenaed documents arrive, and it also depends if AMD just adds them to the complaint, or if they try to bring it to immediate attention.
Joe
sparky,
Is it so difficult to image that Skype, looking at the target users of this app (businesses), and the market share of Intel in the business desktop/mobile space (90% + last numbers I saw), couldn't be bothered with ponying up the significant cost of validating and supporting a complex, multi-threaded application on a second platform? Validation costs are a significant portion of software development these days.
Thanks for your brave attempt to excuse an inexcusable. As far as I know, Skype is the only company (outside of Intel) that cripples their product based on the test for "GenuineIntel". Trying to portray this as a general trend is disingenuous (or ignorant) on your part.
Joe
Misen,
From what I have seen, the Clayton Act affects pricing and rebates. What am I missing?
It addresses exclusive dealing, especially if their effect result "substantially lessen competition or tend to create a monopoly in any line of commerce."
Joe
mas,
Conroe has 3 ALU/integer units
Weird. Wasn't the consensus that there were 4? Could it just be a mistake.
Joe