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.
Some interesting thoughts here, although it seems to me there may be some math errors in at least one of the tables(of course, it's very early and I don't have my contacts in yet, either ;) ). Comments?
Paul
http://www.chip-architect.com/news/2003_08_29_Cache_efficiency_for_SPEC2000.html
Don't know if this has been verified here yet, but thanks to Asus, it looks like Ahlon 64 FX will be socket940, unless they plan on releasing both socket754 and socket940 versions. Seems that there is a recent stepping of Opteron as well, since the BIOS adds support for it, and it's from 8/20/2003.
http://www.theinquirer.net/?article=11169
Keith, looks like ALi is hoping for a resurgence in the AMD K8 market, after having ceded the K7 market. I hope they do well, the more solid competition, the merrier!
wbmw, it is interesting that AMD gets better output transmit power with lower transmit power consumption. AMD appears to be more efficient. If you note the difference in units of measurement, you'll realize that what you had listed as an advantage for Centrino is actually a substantial advantage for AMD. Higher transmit output power with lower transmit power consumption is a very good thing. So, Centrino gets the nod in idle power consumption, and just ekes out a win in the sleep power consumption, AMD gets the rest.
Paul
wbmw, I was particularly referring to the HP part. That is a good sign; not as good as an Opteron win, but it's good in and of itself.
Paul
WOW! (if true) eom
May be a couple of interesting Opteron articles coming tomorrow and Saturday at http://www.tecchannel.de/index.html (Scroll down and check the "Themen Vorschau" section). On Thursday, a longer term server test, and on Saturday a Via-based Opteron workstation test.
Keith, the author from ocprices attended the presentation. I'm not sure if anyone from L'Inq attended or if they heard about it. If they were there, I'd probably lean toward believing them. Anyhow, it's all info to consider.
Paul
Ixse, probably mostly 2 factors:
1) laptops have a lot less user interface with processors, and therefore less chance of damage.
2) They probably don't want to add the additional thickness the heat spreader adds in a low profile environment like a laptop.
Paul
Ixse, Corsair directly gave him the info, although I imagine he may have misinterpreted it. It appears their intensive testing procedure is the limiting factor in output, from what I understood, at least. Makes you wonder why they don't invest in more testers, though.
Paul
One answer and some contrarian info from a Corsair presentation. The second item may be no big deal, perhaps Corsair is just poorly represented in that market. The third item is troubling, if true.
Paul
...-PC3200 Registered will be the next big AMD product released, naturally for the Athlon 64 platform, expected in Q4 2003
-There was a general lack of confidence in AMD's decision to go with Registered DIMMs.
-If Corsairs latest board samples are anything to go by, AMD's Athlon 64 is not ready for release. ...
http://www.ocprices.com/index.php?rev_id=173&cat_id=69&action=reviews
DDB, that fab switch sounds risky. I hope they know what they're doing.
Paul
8-/ , c'mon, now - mobo and drivers for keyboard/touchpad have nothing to do with Centrino! The chipset is highly unlikely to be the culprit, more likely the BIOS. I doubt the mobo is made by Intel - or HP for that matter - it's probably farmed out to Taiwan.
Now I'm certainly one to hold Intel's feet(as well as their supporters' :p ) to the fire when they screw up, but this is bogus. How about we save our excoriations for the real thing?!
Paul
Dan3, stop discrediting yourself. You know very well that the odds this has anything to do with Centrino itself are slim to none.
Paul
You're not kidding - it's got it all. AGP Pro 8x 110W, fast 64 bit PCI-X, SATA RAID, GbE, 8 DIMM slots / 16GB. Now, if only it supported reg DDR400 - well, can't have everything!
Paul
yb - for you! ;) re: Doom3 hardware needs
Running on the fastest hardware money can buy right now (as well as a few components you can’t quite find yet), the gameplay was noticeably choppy when it got very busy. Granted, we were not seeing a slideshow by any means though that is a distinct possibility for those who would try to run the title as it stands on lesser hardware. One thing is for sure, the title will definitely need a killer gaming rig to be played the way developers intended. If nothing else, this demo was a reminder to start saving pennies for the upgrades when the title ships.
http://www.hardocp.com/article.html?art=NTA0LDI=
More complete A64 images. 0320 production date. So, in May or so they had a 3300. How far have they come since then, I wonder?
Paul
http://www.hardocp.com/image.html?image=MTA2MTAyMjA5NjIxcGhDVkY2RzJfMl84X2wuanBn
http://www.hardocp.com/image.html?image=MTA2MTAyMjA5NjIxcGhDVkY2RzJfMl85X2wuanBn
BUGGI1000 - Hmmm.... first shot of A64 mobile - interesting!
Paul
DGagnon819, my point wasn't that info was being withheld, but that the way the power grid is designed to automatically route and transfer loads, it's not surprising to me at all that it would happen very quickly. Of course, computer problems could cause it as well.
On a side note, from something I seem to recall you saying on RB, I'll bet you're either at Seymour Johnson, Pope or Bragg, right? (WUNC's mine, too - Fuquay Varina - home, sweet home).
Paul
sgolds, actually in an NPR story this morning about the '65 and '77 blackouts, one of them(I forgot which) had started up in Canada, I believe, and spread to Rochester, down to Boston, then to NYC in the space of 10-15 minutes.
Paul
Don't know if this has been mentioned yet, but I was just browsing through some Opteron BIOS tech docs and came across mention of PowerNow! and a new trademark I hadn't seen before - AMD "Cool'n'Quiet" technology for the desktop environment. A lot of people will like that, especially in light of Prescott's problems. If they run into snags fixing it...
Paul
drjohn - and your point is...?
Welcome Ixse! I've seen you posting over on SI(I lurk there); you'll find we have some very good discussions here as well, although it can be a little bit more confrontational and noisy at times. But it is definitely a worthwhile board.
Paul
Tenchu, because the average Joe, who'd be buying the peripheral, may not know about that - he just wants to be sure it'll work with his stuff. Kind of like all the meaningless Y2K ready labels - for someone informed, it's silly, but it's comforting for the buying public.
Paul
Well, the new logos are okay, but I liked the old ones better; they look classier.
Signs of some actual marketing...
http://www.infoworld.com/article/03/08/12/HNamd64logo_1.html
AMD64 logo program launched
Campaign aimed at improving awareness of its 64-bit processors
By Gillian Law, IDG News Service August 12, 2003
PC manufacturers expect a new marketing campaign announced by Advanced Micro Devices (AMD) on Tuesday to improve awareness of the company's 64-bit processors among consumers and corporate buyers.
The campaign will put an AMD logo on PCs, software and peripherals, identifying them as being "ready for AMD64." It is being launched to raise awareness of the platform, which includes the AMD Opteron processor for servers and workstations, and the AMD Athlon 64 processor for desktops and mobile PCs. The Athlon 64 processor is due to be launched on Sept. 23, the Sunnyvale, Calif., company said in a statement.
In addition to running new 64-bit applications and operating systems, the AMD64 platform can also run older software written for the 32-bit x86 processor architecture. This means users can upgrade their hardware while continuing to use legacy applications, AMD said.
The campaign will help desktop, notebook and server manufacturers like Evesham Technology of Evesham, England, push AMD products in a market where Intel's name is far better known, said Carolyn Worth, a spokeswoman for Evesham...
Taunting? Give me a break! eom
EP, you'll be missed, but I haven't found any dearth of free discussion on this board.
Paul
Keith, although things are obviously a little different with nForce3, with integrated memory controller and aHT, nForce2 has the important feature of having all the peripheral busses and AGP separately clocked from the FSB and memory. I can't imagine they'd want to backtrack on that, it's a strong selling point.
But then, it is a different beast. As far as I've heard, though, the OCing has been purely multiplier based(which due to the K8 design, results in various memory speeds, depending on the multiplier), but the HT frequency isn't changed at all, so there's no reason to assume anything but the memory and internal core-to-memory bus is running faster.
Paul
chipguy, it's interesting that you and others chastise folks on this board for comparing Opteron to Northwood, not the upcoming Prescott, but now you're comparing Northwood to K7, not K8. Consistency, please.
Paul
Haddock, wbmw, be careful with that; the board design is a big part of RAM performance and stability. Just having quality chips onboard is not enough to guarantee quality modules. In this case I would probably tend not to worry so much, I suspect MDT is a German company and they tend to be sticklers for good design. Klaus and Keith, can you help us out here? By the way, the link in the post was invalid, some extra characters wound up tacked on the end. This one works:
http://www4.alternate.de/html/nodes_info/iaidu1.html
CJ, apparently he doesn't visit Bunny's SuperPi listings very often. Seems to me I recall P4 took the lead back, although it seems to require at least 1400MHz gap, not to mention obscene FSBs to do so. Regardless, P4 sure seems to do well with "AMD-optimized" SuperPi!!
Paul
CJ, it's the secret part that has me thinking your scenarion 1 is right, although I suppose they might have wanted to slip in under the radar if they had the contractual issues and thought that if it were made public, they might have some purchasing competition.
wbmw, good analysis. Interesting times!
Paul
Auf wiedersehen, EP! Don't wait too long, you'll be missed.
Paul
CJ & wbmw, Apparently Opticom/Thin Film are working with Intel, but they've run into some problems with volume production that Coatue hasn't, but Coatue isn't yet fabricating 3D structures yet. It sounds like Opticom's emphasis is pure plastic and they're diverging from Intel a little. It's also interesting in light of Opteron's compatibility claims, that Coatue says their process is fully compatible with existing semiconductor fabrication technologies. Ease of conversion again.
http://www.electronicstimes.com/tech/news/OEG20030513S0044
http://www.siliconstrategies.com/story/OEG20030512S0013
http://www.coatue.biz/
No, wbmw, I think you're mistaken here. That's Coatue's whole point. Their polymer memory is fast and cheap, so it can take the place of both system memory, and to some extent storage, depending on how much is needed, of course. I haven't heard yet how far it's likely to scale in terms of capacity.
Glad to hear you liked it, some guys on SI thought it was cheesy.
Anyone catch a glimpse of the Athlon 64 logo(s) that The Inquirer had up for a few minutes before withdrawing it(them)?
Interesting thought re: IBM/AMD cooperation results.
http://www.siliconinvestor.com/stocktalk/msg.gsp?msgid=19190210