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.
EP, why are you assuming it's due to SOI, though? There's a lot of word going around that it's due to speedpath issues, that the SOI part actually has done its job re: power and speed.
Paul
chipguy, I don't think he was being specific, just referring to the availability requirements for submitting/publishing the SPEC scores. Doesn't that require availability within 6 months?
Paul
Hey, Petz, you better make that "bated" breath, or your loved ones will be running from the smell of worms!! ;p
Paul
chipguy, just because AMD is behind Intel in large on-die cache doesn't mean that dealing with the bottleneck of main memory is any less important. Besides, I'm sure they're working on larger cache CPUs, but until they have more of a foothold in that market, it wouldn't make a whole lot of sense for them to waste their limited capacity on such a chip yet. By the time they can really sell them, I'd be surprised if they didn't have at least a 2Meg cache CPU in the works. Yes, they're behind in cache, but they're not standing still either.
Paul
Joe and sgolds, not all DSL is like that. I have Sprint/Earthlink, and while it was rough for the first six months (I was a pioneer in my area) in the 2 years or so since, I've had only 2 or 3 service "failures", and never had a problem that wasn't fixed by just re-booting Windows or the DSL "modem". Except, of course, the time a school construction crew across the street hit our phone line with a truck and ripped my phone interface off the house! I would recommend, if possible, to get it from your local phone service provider, though, so there's no buck-passing as to who's responsible to do something. This primarily comes into play during installation.
Paul
yb, make that Opteron w/ registered DDR333 CAS2.5 memory. Another whole cycle. So it's like 1.5 extra cycles on top of the 66MHz memory deficit.
Paul
yb, nonetheless, it raises the question as to whether the Sanmina/SCI Newisys acquisition may have been driven by IBM, doesn't it?
Paul
Intel still having TDP issues with Prescott? Seems the TDP is considerably higher than they'd expected. The Japanese article, as usual, isn't handled very elegantly by Babelfish, but it does better than some other translators I tried.
http://babelfish.altavista.com/
http://pc.watch.impress.co.jp/docs/2003/0722/kaigai005.htm" target="_new">http://pc.watch.impress.co.jp/docs/2003/0722/kaigai005.htmhttp://pc.watch.impress.co.jp/docs/2003/07...
Anyone know offhand which rail(s) the Opteron draws its power from? The xbitlabs 144/SK8N review has me wondering about their power supply. In the BIOS screenshots, it displays the 12V rail voltage as being 11.126V! I've never seen any system with a voltage that far off spec!
Paul
http://www.xbitlabs.com/articles/cpu/display/opteron-1_9.html
Edit: Why they call it a 144 review, other than the uniprocessor config of the board, I don't know - you can see from the image that it's a 244.
Hmmm.. interesting in light of Sanmina/SCI acquisition of Newisys.
Paul
http://crn.channelsupersearch.com/news/crn/39304.asp
yb, how do you figure that? amdzone.com has at least 4 game benchmarks on the 1.03 version SK8N and 1.8GHz Opteron with a Radeon 9700. It did quite well and will be quite impressive with 2 or 2.2 GHz. Especially when it loses a CAS cycle with unregistered DDR400. :)
Paul
http://www.amdzone.com/articleview.cfm?ArticleID=1313
rsteiner, it's not at all beyond the realm of possibility. I haven't personally done it, but with a high quality heat sink, I see no reason you couldn't, assuming you can drop the voltage sufficiently. I probably would not go for something like the Thermalright SLK series HSs unless you can use a fan of some sort, because they really need at least some air forced through them to work properly because of the fine pitch of the fins. Although with the appropriate case fan configuration, it's possible it may be able to pull enough air through to suffice. If you don't have something like that, what you might try is a Zalman ZM-F1(80mm) or ZM-F2(92mm, requires a little ingenuity to fit it to an SLK-800) on the HS. Zalman has an inline resistor rig to drop the fans to silent mode(even at full voltage, they're very quiet). I believe the FB123 is a 92mm with a bracket and the inline resistor, which is a better deal than separate items. As long as you use a fan of some sort, it's hard to beat a Thermalright.
Paul
Scroll down to the AMD section in the listing below to see the performance ratings
http://www.overclockers.com/articles373/
Hester - Financial problems not the reason for Sanmina/SCI Newisys purchase.
Paul
http://www.statesman.com/business/content/business/071703/0717newisys.html
So, is the consensus that the bad news was already priced in, or is it about to dive again?
Paul
Keith, all the hand-wringing and trumpeting we've been hearing about the "death" of Newisys seems to have been a touch premature. ;)
It was encouraging to hear that SOI gave them their fastest silicon yet, if I understood correctly. Also, that there is a 2-way transfer of technology between IBM and AMD. The short lifespan of 130nm sounds encouraging as well. No, I'm not Mr. "pie in the sky", I just figure there have been plenty of negatives focused on already.
Paul
As long as Mike's been around, one would think folks would have picked up on that by now.
Paul
chipguy, wrong! It was being done with the AMD750 chipset, which while it was no incredible barnburner, was respectable and was quite robust and reliable. Besides, by your logic, what should have been done with that paragon of reliability i820/MTH ?
chipguy, P4's target audience when it was introduced didn't know what SPEC is, much less care about it or ever run it, so I think if you look at Anand's review, which used the normal gamut of benchmarks, it's clear it wasn't very competitive with 1.2GHz Athlon at all. That changed, of course, but my point was that P4 was not competitive when it was introduced, and Opteron most certainly is.
Paul
Tenchu, agreed that Dan3's statement re: P4 performance is absurd, but if I'm understanding you correctly, so is yours re: Opteron. Yes, they're having problems ramping up the clock, but even with that problem, Opteron is one of the top performing chips. So, its IPC is hardly irrelevant. Now, when P4 was first introduced, was not in the top ranks. At intro, Opteron is and when they do clock it a little higher, Opteron/A64 will be the top performing chip. Overall, that is; granted there will be limited instances where Itanium (or another) will pull off wins. Now, whether AMD can actually make them or not is a whole different story...
"Madison - a chip that the AMD partisans feel inclined to label as having "sucky" performance."
C'mon, wbmw, be serious. No one who matters thinks Itanium's performance sucks, except on legacy software - where it undeniably uhh... sucks. At least in its current form.
Paul
Haddock, yes, I was surprised at how little response there was when I posted on this yesterday.
Paul
http://www.investorshub.com/boards/read_msg.asp?message_id=1197444
BUGGI1000, this seems to back up what Magee said about the sockets. If you look on the sign, you'll see it says both Socket940 Opteron and Socket940 Athlon64. Hmmm......
Paul
Semi, I don't necessarily agree with that. But, it is possible that the margin on the Centrino laptops makes them less attractive to sell than other products. Just a thought, doesn't mean they aren't selling.
Paul
Keith, try hunting around here, if you haven't already.
Paul
http://www.aceshardware.com/SPECmine/
I suggest you look again. Circulars do have items that are on sale, but they also have many(often more) items included which, although they may be perceived as "sale" items by the customer because they are a good value, are actually normal price items with good margins for the retailer.
Prescott incompatibility may not be the whole story. We'll know soon enough, I imagine, but Ed's got a point.
Paul
http://www.overclockers.com/tips00424/
Intel isn't the invincible monolith, they still make stupid mistakes, too. A number of folks are going to be furious!
Paul
http://www.anandtech.com/#20014
sgolds, I'm just saying that if the rumors re: A64 being 940/939 pin are true, then this chip and these MBs may well be obsolete engineering samples(perhaps with their own performance issues), and therefore irrelevant.
Paul
yb and drjohn, don't put too much stock into this, we know next to nothing about the system. From the bandwidth, we might be able to surmise it's PC3200 RAM, but what are the timings - 3-3-3-7? It looks like it's probably an overclocked 1.8GHz chip. And since it's socket 754, it may be irrelevant, if the recent rumors are true.
Paul
sgolds, to take advantage of additional registers?
Paul
SiS 755 AMD64 chipset announced.
Paul
http://www.sis.com/news/press/755HSE.htm
Opteron/SK8N performance numbers out at AMD Zone. Encouraging, but not yet overwhelming.
Paul
http://www.amdzone.com/articleview.cfm?articleid=1304
CJ, it seems they approved PC3200 in December, at least according to TwinMOS. Still trying to track down direct info on JEDEC's site.
Paul
http://www.twinmos.com/about/press/press2003_0109.htm
kpf, the only way it would seem to make sense to me would be if it were one pin short of Socket940 if that were a pin that enabled MP configurations. That would enable the dual channel memory interface and full HT link. But then how would they justify going back to Socket754, unless that's introduced later as the new Duron/opponent to Celeron? Hmmm.....
In any case, it seems way too late to do anything but minor changes from either of the two layouts long since established. The MB makers would throw a hissy fit! And justifiably so.
Paul
Haddock, re: no changes. Not if Mike is right about this. Frankly, I think he'd had a few too many whenever he picked up this info - it makes no sense for things to go this way.
Paul
http://www.theinquirer.net/?article=10353
Jerry R, of course, the difference is that a number of developers have noted that converting their apps to 64bit for AMD64 required minimal effort, which one could certainly take to mean at least possibility of minimal cost.
Paul