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.
wbmw, as it's written, it looks like this refers only to the 1MB L2 cache Athlon64.
Paul
http://www.aceshardware.com/read.jsp?id=55000244
spokeshave, one apparent correction. According to Johan, Opteron and A64 are the same thing, just with HT links disabled. So, they're not really 2 separate products, just a different final(?) step.
Paul
Elmer, could you share what facts you are basing the 50% utilization figure on?
Thanks,
Paul
re: price move
Could it have to do with this? I'm not sure why it would be necessarily considered positive for AMD to spin out(same as spin off?) its flash business, though.
Paul
http://www.siliconinvestor.com/stocktalk/msg.gsp?msgid=18721449
cj, was it already known that Opteron and Athlon64 L2 cache had 128 bit bus? I can't remember if I'd heard this before. Planet3dnow's CeBIT report had a segment with John Crank in which he covered some Opteron/A64 stuff.
Paul
http://www.planet3dnow.de/artikel/diverses/cebit03/index.shtml
Jozef, one of your pet peeves may have been dealt with. Planet3dnow, in their CeBIT report, is reporting that some of the mainboard manufacturers specs were listing Athlon64 boards as having 6.4GB/s of bandwidth, which would correspond with dual channel PC3200/DDR400.
It also seems to say(3rd page) that system problems were why Athlon64 was delayed("You and I buy processors, because we love playing around with them. But normal users don't buy processors, they buy PCs and that's why we must secure, that everything runs fine, when Athlon 64 comes") - perhaps that's the AGP issue?
Paul
Article's in German, pick your translator
http://www.planet3dnow.de/artikel/diverses/cebit03/index.shtml
EP,yb - the link worked for me.
Paul
cj, re: Thorton
From the translation of the PC watch article, it seems that Thorton will be based on the Barton core improvements, which seems to have better power consumption than TbredB, but still have only the 256K L2. Translation link below(translation not too hot, I especially question the 8GHz part, maybe it was referring to 8way?).
Paul
http://www.worldlingo.com/wl/translate?wl_lp=JA-en&wl_glossary=gl6&wl_fl=2&wl_rurl=http%...
cj, I missed that. Who's Frank Weber? Or do you mean Fred?
avatar, one of the most interesting tidbits I found in this "report" was that the Hammer crossbar apparently already has the interface for the second on-die CPU. Now, I don't know how much stock I put in this, since some of the other technical info he got seemed either stretched, inaccurate or outright bogus, but we shall see.
Paul
Not until the "For testing purposes only" part disappears.
econo - Well said. EOM
econo, do you think AMD will miss anything by missing the back-to-school crowd with Athlon64? I'm not convinced they would even be interested. A64 will be high-end consumer stuff when it's intro'd, and most back to school buyers are not going to pay top dollar for the kids' PCs. I think they will do just fine where the intro is placed now, as long as it's a volume intro, and well advertised, whether commercially or by word of mouth.
Paul
wbmw, I would agree with you, that approach would make sense. The biggest values of Newisys are the ease of introducing a system, since they've done the hard work, and the prestige of the big guns they brought in from IBM. Newisys will allow them to ease into Opteron in blades, etc., while they can take the time to do their "big iron" designs their own way.
Paul
OT:edgarcayce & combjelly, you particularly may be interested in my last post. Using the script, if the last line isn't grayed out, you'll find the info there interesting! ;)
Paul
Ed at overclockers.com noticed something I didn't when I checked out this script site. Downbinning revealed?
Paul
http://www.overclockers.com/tips00316/
re: CeBIT
Has anyone seen anything else about the big Fujitsu Opteron system other than what's shown below at The Inquirer? (Which appears to have at least 7 1U units - 2P each?)
Paul
http://www.theinquirer.net/?article=8274
I usually like to keep the PCI below 35MHz, so I don't run long term over 138FSB, since my MB doesn't have a PCI divisor of 5. Is your MB an nForce2 board that can set PCI/AGP independently? If so, I like to run Memtest86 long term to see if any errors come up, then when I get to the point where they first start to, I'll back off a 2 or 3 MHz for stability's sake. Then there are a number of things you can run in Windows to test there, but the good thing about Memtest86 is that there aren't any driver/Windows instability issues to fog the results.
You might check out some of the OC databases for the same lot as your chip to see how they've done, so you have some idea of what to expect. Of course, take into account the cooling used!
Paul
http://www.vr-zone.com/guides/AMD/TBred/
If you need to be sure on a 1700+, go to excaliberpc.com, they have a couple of different lots of TbredBs in stock, just search on AXDA1700DUT3C and you'll find them. $65 & $66 shipped.
Paul
Johan de Gelas(Senior editor at aceshardware) has confirmed that Barton 3200+ will use 166(333)MHz FSB. It's a short comment, but he never says things like this unless he is certain. Usually he has the actual product in hand and is testing it.
Paul
http://216.87.214.213/forum?read=95024288
dlr_972, usually they use an actual photo, but with the 1700+, they have at times used the TbredA picture, while actually shipping the TbredB.
Paul
yb, It's been a toss-up from what I've heard regarding 1700+ at newegg. It seems to vary from shipment to shipment as to whether they have As or Bs. Even when their part # has said DLT3C(TbredA), at times people have been receiving DUT3C(TbredB). But Palominos seem to be disappearing fast.
Paul
yb, I think the main reason that 2200+ hasn't done so well is that enthusiasts avoid it like the plague in case it's a TbredA, since TbredA 2200+ has zero headroom. There's no way to tell the difference from the part #, since they're identical for 2200+ TbredA and B. 2100+ can be identified as A or B from the part #, and 2400+ and up are all B.
Paul
1)Certainly it will affect the max OC you can reach, but they've been overclocking to a decent extent at even less than default voltage.
2)Watch for any instability as well.
3)Sounds like they may be, a lot of people have been able to OC 2100+s to very high levels, as well as anything else. So, they're probably downbinning them. Looks like their production sweet spot may be above where the sales sweet spot is.
Paul
Anecdotal market share info. I went by Best Buy today to check out wha's on display. Perhaps better than what you'd expect for desktops, not so good for notebooks. At least the Durons are history!
Desktops: Celeron-4 Duron-0 P4-7 XP-5
Notebooks: Celeron-4 Duron-0 P4-8 XP-2
Once they ship, I don't think it'll take as long as usual to get systems on the shelves. The other components have been ready for a while. Qualifying with the latest speed CPU won't be that big a deal.
In any case, I intended my main point to be that it was Tom talking, so I don't lend it a lot of credence.
Paul
Very interesting article. Good PR for AMD, for sure.
BTW, in case you didn't notice, you can click on the image for a closer(or at least larger) look.
Sorry, Keith, I don't buy it. That price list has so much bad info on it, it's laughable. Just look at the frequencies related to the XPs and you'll see this had to be old speculation. Would be nice, but it's wishful thinking. AMD doesn't release stuff before they said they would anymore.
Paul
Don't be absurd, Elmer; you know very well the difference between availability of a CPU and availability of an OEM system based on said CPU. Besides, this is Tom's talking - not exactly the most accurate source when it comes to this sort of prediction. I doubt they have a clue about who's going to produce what when, just that if the OEM were to start from square one with the intro, they'd have to take until Christmas to get a product out. I find it a lot more likely that there will be systems available immediately, there's been plenty of both prep time and interest.
Paul
Last stepping? Remains to be seen, maybe they'll find a little more. In any case, very nice chip, I'm considering the same one.
Let us know how it goes.
Paul.
Now THAT'S a motherboard!(To the right) Wish they had a closeup shot! The label seems to have the Newisys logo, so I'm assuming that's their 2P baby.
http://www.forumdeluxx.de/gallery/showphoto.php?photo=816&password=&sort=1&thecat=515
Sure sounds like him(her?), doesn't it?
Tom's day2 CeBIT with image of Athlon64 mobile chip.
http://www.tomshardware.com/business/20030314/cebit2003_2-03.html
Ooops, my mistake, "run higher" does not equal 2 GHz.
Not that it's too comforting, but the article did say that they would intro with a 2GHz model.
SMP will be where it really shines, but the full-speed memory on-die controller is not something to be overlooked for 1P applications. If the Athlon64 price is right(which remains to be seen), enthusiasts will love the performance. The motherboards could be even cheaper/simpler to make than socket A.
Paul
I wouldn't put too much stock into this score without knowing a lot more about the development/stability of the drivers and the OS. Not to mention that it's a VIA chipset, and their first effort at HT-AGP interface. I'd be a lot more concerned if it were an nVidia chipset with that sort of score. Time will tell.
Paul
edgarcayce, if it's a "B", in WCPUID, the Family/Model/SteppingID will show 6/8/1 Standard and 7/8/1 Extended.
Paul
edgarcayce, if the part number is AXDA2100DUT3C, it's a TbredB. TbredA would be AXDA2100DLT3C. What version WCPUID are you using? If it's an older version, it may not have been updated for later processors, but it may be BIOS as well. Does it name it properly in the bootup screen? If it doesn't, then I'd suspect BIOS; if it does, I'd suspect the WCPUID version. I'll get back to you in a minute re: model number info in WCPUID.
Paul