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.
Well Charlie outranks you in mentioning reasons for his conclusions approx. a million to zero. Does that make you a parasite?
Regards,
Rink
Montecito to launch in July at an amazing clock speed of 1.6GHz.
http://www.theinquirer.net/?article=32158
Regards,
Rink
Spaarky, there's nothing in your post I disagree with.
R
re: AMD's 4x4 is supposed to be their "top secret" answer to Conroe
I must disagree with the idea that it was meant as an answer to Conroe. 4x4 is mostly just a marketing trick to keep up image. It can even be somewhat effective in this. Alienware presented yesterday and it's not that big a step to believe that there's a good chance that they'll build systems based on this. The systems will be reviewed and attracked attention more than money. From AMD's point of view it's a no brainer (efficient use of money for the purpose of marketing). From Intel's point of view though this particular system isn't of that much importance at all (very low rev). I agree that Core 2 Extreme will fit the bill way better for far most gamers. What I disagree with is that this was meant to be AMD's "top secret answer to Conroe", as it is mostly marketing and almost no answer. It might be successful in a rather small niche.
90nm QC will have lower frequency, while most games are still single threaded. I doubt it would have had made much impact in the gaming market. At least 4x4 can be nicely upgraded over time - it's not exactly all bad. QC is relevant for the server market mostly. I agree that 90nm QC has about a zero chance to see the light of day now 65nm is close.
Regards,
Rink
AMD Announces Dual-Socket, Dual-Core '4x4' Platform: http://www.extremetech.com/article2/0,1697,1970271,00.asp
Dual socket, DC Opteron for gaming market, dubbed "4x4".
This has got to be the recently rumored 'trick' AMD had left up it's sleeve. I think it'll work for at least part of the gaming market. This solution can't be leveraged in the consumer market outside of gaming. It's here to keep the image going in the pc gaming market. It looks like it's going to be needed for at least a year and maybe quite a bit longer.
BTW, I think that 'key customer' might indeed be Dell as indicated as much in the article. Alternatively HP might fit the description too.
Regards,
Rink
Wouter, yes, still it allows AMD to use larger dies, which come in especially handy when Intel is a year or so ahead in transitioning to smaller nodes. Fab 36 and fab 7 are a way to produce more dual core than otherwise possible. Transitioning fab 30 to fab 38 might provide the same function when Intel transitions to 45nm.
Regards,
Rink
AMD to Invest US$2.5 Billion in Germany to Expand 300mm Microprocessor Production
Monday May 29, 4:01 am ET
Large Investment in Europe Underscores AMD's Leadership Strategy to Drive Microprocessor Manufacturing Innovation and Reduce Costs for End Customers
SUNNYVALE, Calif. & DRESDEN, Germany--(BUSINESS WIRE)--May 29, 2006--AMD (NYSE: AMD - News) today announced that it plans to expand its microprocessor manufacturing capacity over the next three years by adding additional 300 millimeter wafer production capabilities in Dresden through the implementation of three new projects.
More here : http://biz.yahoo.com/bw/060529/20060529005047.html?.v=1
Pragmatic choice: Same employees (smaller learning curve). Lower cost than fully new fab. Smaller time to volume production.
Regards,
Rink
AMD to Invest US$2.5 Billion in Germany to Expand 300mm Microprocessor Production
Monday May 29, 4:01 am ET
Large Investment in Europe Underscores AMD's Leadership Strategy to Drive Microprocessor Manufacturing Innovation and Reduce Costs for End Customers
SUNNYVALE, Calif. & DRESDEN, Germany--(BUSINESS WIRE)--May 29, 2006--AMD (NYSE: AMD - News) today announced that it plans to expand its microprocessor manufacturing capacity over the next three years by adding additional 300 millimeter wafer production capabilities in Dresden through the implementation of three new projects.
More here : http://biz.yahoo.com/bw/060529/20060529005047.html?.v=1
Pragmatic choice: Same employees (smaller learning curve). Lower cost than fully new fab. Smaller time to volume production.
Regards,
Rink
AMD will invest to upgrade current fabs to the tune of 1B euro's, and it's currently highly uncertain if it'll build a third fab next the the other two. That's at least how I am reading current stories.
EDIT: This one mentions the 1B upgrades, and says the building of a third fab is not planned (at least not now): http://www.wallstreet-online.de/dyn/community/postingaction.php?post=21811304&action=drucken&...
Regards,
Rink
Berkeley paper assessing the IBM/Sony/Tosh Cell cpu for scientific use:
"Overall results demonstrate the tremendous potential of the Cell architecture for scientific computations in terms of both raw performance and power efficiency," the authors wrote. While their current analysis uses hand-optimized code on a set of small scientific kernels, the results are striking. On average, Cell is eight times faster and at least eight times more power efficient than current Opteron and Itanium processors, despite the fact that Cell's peak double precision performance is fourteen times slower than its peak single precision performance. If Cell were to include at least one fully utilizable pipelined double precision floating point unit, as proposed in their Cell+ implementation, these speedups would easily double.
http://www.hpcwire.com/hpc/671376.html
The great potential of Cell keeps being confirmed (even though it might be harder to program for). This is why I speculated about a year ago that cell-alike features will make it to x86 processors. So far I have been proven wrong.
Regards,
Rink
I read a couple of times before that AMD is sharing Geode (embedded x86 processor) technology with China. Because it is a couple of generations behind especially performant wise compared to current mainstream x86 technology almost nobody at that time reacted in a strange way. I think this 'article' is about this same topic.
EDIT: I see CJ and Smooth went over this already.
Regards,
Rink
Yep, last earnings call AMD said they were supply constraint in Q1 a couple of times (components/materials/whatever...).
Regards,
Rink
Kate, IBM didn't drop AMD. Consequently your post is largely incorrect. And yes they didn't exactly embrace AMD either, but that is not what you said.
Regards,
Rink
Are you anywhere near sure those orders will happen? I mean current AH prices are not that close to yours. (Just try to learn something here).
Regards,
Rink
I bought back into AMD a week ago (as posted on SI). This was just a short term trade (I'm not planning to post all short term trades but will post my long term ones). It's just a quarter of my portfolio (jan '08 $30), but still very nice.
Regards,
Rink
OK, I stand corrected in that it isn't a sure thing. Tx.
Regards,
Rink
Wbmw, Hans wrote about starting at 100nm going to 90nm.
The 15% perf/watt increase I quoted indirectly from AMD (42% increase over no strain, while DSL strain is 24% of that 42% and new SiGe based strain takes care of the rest i.e. 18% additional or ~15% improvement over current DSL).
Hans wrote extensively about Prescott and Yamhill. See: http://www.chip-architect.com So after K8 and Prescott he didn't write as much on his site anymore, but you hardly can't accuse him of being a focussed on AMD exclusively.
re: Another example is the die size estimate of his, which claims a 65nm quad core at 150mm^2. Intuitive estimates would put it at >200mm^2, given that it's twice as many cores and cache as K8, along with a sizable number of improvements, and only one process shrink to make up the difference. So I think Hans' methodology (in which he scales two die photos) uses a lot of assumptions which may not end up proving true.
I trust his estimates over anyone elses here. He's proven himself extensively in this area. You may want to consider the recently released die photo of a 65nm K8 with much denser L2 cache (a newly designed denser type of L2 cache) than expected with a normal shrink. The denser cache is an important reason why the 65nm die is quite a bit smaller than normally expected with a straight shrink. Intel has that kind of denser cache for years already so it's about time too.
Anyways, before saying Hans isn't reliable check his rather flawless record.
Regards,
Rink
Chipguy, much appreciate that. I'll have to be very cautious then with my current thinking that K8L might outperform NGA.
Regards,
Rink
Chipguy, I would never have made that statement to you. It was meant for someone else.
May I ask if you think K8L will slightly outperform or slightly underperform compared to NGA (both native dual core versions of them)?
Regards,
Rink
Wbmw, I stated that I believe K8L will outperform NGA. I meant to say 'somewhat' after that sentence. This is close to your statement that you think they'll be close. You didn't factor in IMC in your post which I think is relevant considering AMD still has that as an advantage if most features are more or less on par with NGA. Everything else being equal the IMC and to a lesser extent the xbar will let K8L outperform first gen NGA. Just my 2c ofcourse.
Couple of minor points:
- Shared L3 is I think quite significant for QC, because the L2's aren't shared.
- re: 16. New instructions: big whoop. AMD has had little success with new instructions outside of x64. I think MS will continue to strive for a unified instruction set, meaning that when AMD introduces new instructions they might well have asked and gained the support of MS.
re: The other "sounds" you are hearing are from people who are still in denial that Core 2 will unseat the current K8 from its performance crown. So to them, it's obvious that K8L will be a lot faster.
I take issue with that. Those people some of them who post at RWT, some at Aces, all long believe NGA will outperform K8. It's no use making accusations without facts.
BTW most of these guys now think that NGA's lead will only last till K8L. This is both in line with what you are saying as well as what I am saying.
re: In fact, Intel is also vulnerable here if they don't get their platform approach to work. So IMO, they have to find ways where AMD can't follow, due to the fundamental lack of a means to influence their 3rd party component vendors to support features as fast as Intel can do internally.
Intel's platform approach isn't that difficult to copy in general terms. I don't see how you come to the conclusion that AMD can't follow in general terms considering their partnerships with nVidia/ATI. From another angle: I don't see the first new platform product VIIV being a killer product. Yes Intel might sell more chipsets but who cares as long as those chipsets aren't better than what nVidia/ATI make for AMD platforms?
Regards,
Rink
re: That's not bad, but it might be because AMD tends to backrev their processes with future enhancements, so their current 90nm may be very close to 65nm already. If that's the case, they won't get a significant perf/watt boost until late next year or early 2008, when they start incorporating 45nm features in their 65nm process. It's also unclear how much extra power those additional FPUs will cost them in the K8L core. It might end up that with the added transistor budget, the 65nm shrink doesn't buy them all that much.
Hans de Vries claims that AMD started the 90nm node at 100nm and only gradually shrunk to 90nm. Also at 65nm AMD will add next gen strain (SiGe added to current DSL). This alone will help the perf/watt increase about 15%.
Regards,
Rink
Improvement K8L over rev F:
0. Native quad core
1. Hypertransport up to 5.2GT/s [HTT version 3]
2. Better coherency
3. Private L2, shared L3 cache that scales up.
4. Separate power planes and pstates for north bridge and CPU [allows higher frequency / lower power]
5. 128b FPUs - see 14,15 [probably better performing than NGA]
6. 48b virtual/physical addressing and 1GB pages
7. Support for DDR2, eventually DDR3
8. Support for FBD1 and 2 eventually
9. I/O virtualization and nested page tables
10. Memory mirroring, data poisoning, HT retry protocol support
11. 32B instead of 16B ifetch [Better prefetch, see also 12]
12. Indirect branch predictors
13. OOO load execution - similar to memory disambiguation
14. 2x 128b SSE units
15. 2x 128b SSE LDs/cycle [bit of an advantage over NGA as well]
16. Several new instructions
Coprocessors:
media processing
JVM/CLR acceleration
TOE, XML or SSL processing
The above and more can be found here: http://www.realworldtech.com/forums/index.cfm?action=detail&id=67239&threadid=67239&room....
Oversimplified K8L is to K8 what Merom is to Yonah.
And yes I think K8L improved enough that the dual core version of it will outdo NGA on most loads. This is provided Intel doesn't walk away frequency wise. Equally the quad core version will I believe outperform a native quad core version of NGA. I believe this because some people that know way more than me and that I've learned to rely upon up to a point are making similar sounds.
See also the Inq for a more detailed explanation: http://www.theinquirer.net/?article=31761
Regards,
Rink
Kate, NEC/Packard Bell are new to Turion based notebooks fafaik.
Regards,
Rink
(Turion X2 launched:) AMD Delivers Multi-Tasking Performance On-The-Go With First 64-Bit Dual-Core Mobile Processor
...Systems are initially expected from Acer, ASUS, BenQ, Flocity, FSC, Fujitsu, Gateway, HP, MSI, NEC, Packard Bell, Sotec and TongFang. ...
http://www.amd.com/us-en/Corporate/VirtualPressRoom/0,,51_104_543~108456,00.html
Strong list even without Lenovo.
Regards,
Rink
I just got an image of Icarus flying too close to the Sun with wings made of wax.
LOL. You made me visualize it during reading...
I'd say your image could be true, yet I'm not counting on it as a sure thing either. My take is NGA is somewhat significantly better than K8 until K8L. I think even more important than NGA being better in the mean time is the deadline for K8L, when it will finally become available in 2007.
(The fact that Intel is way faster to 65nm doesn't matter nearly as much because AMD has the increasing fab capacity to weather a dual core war brought onto it by Intel. That's doable imo.)
If K8L would ship around summer 2007 OEM's will know that by now and will keep their AMD server lines until that time on strength of current virtualization, basic performance that is still not bad at all, and upcoming K8L. Customers already owning Opteron gear will not all of a sudden start buying NGA again provided the price of the Opteron gear is right. Same goes for consumer lines but to a much lesser extent. Those lines will probably shrink but be maintained too. Generally speaking that is. AMD might well grow in mobile on strength and slightly nice timing of Turion X2, and it might even penetrate business accounts with their stable image program. If K8L is basically for start 2008 however the picture is much more dire for AMD imo.
Even provided K8L would be for summer 2007 AMD's share's would still be weak and decline until prove that AMD keeps growing a couple of quarters after NGA launch. Similarly Intel's shares in that case will strengthen until any prove of sustained further AMD growth. If K8L is for summer 2007 I don't think that is impossible. Don't read me wrong: For now I'd give that scenario only a 25% chance (because I think K8L could be very late 2007 / start 2008, and because Intel might well regain significant share in the mean time). So I'd very much like to be sure about K8L's time line.
Regards,
Rink
Smooth, yep, it's one or the other.
BTW, I'm not sure that the info on that chart came from AMD / is true. I'm not sure about K8L supposedly being due summer 2007 either. Info is too scetchy for me so far.
Regards,
Rink
Chipguy,
re (Doug wrote): Now that we have you here, why don't you answer the question: Roughly what % of the performance improvement (clock/clock) from the K7 to the K8, on average (or break it down by categories of application) do YOU think is due to the IMC?
Didn't you write on rwt just before K8 arrived that you thought that K8 would have a 10-15% performance improvement due to the integrated MC? I.e. about half of the total improvement you expected? I remember you said something like this (might be off slightly) because you turned out to be quite right.
Regards,
Rink
SGI files for chapter 11: http://www.theregister.co.uk/2006/05/08/silicon_graphics_chapter_11/
Shares fell accordingly.
Regards,
Rink
Kate, re: No I declared a winner from what I read :) I'm right.
That's actually not impossible.
Regards,
Rink
Wbmw, we'll see. This article was definitely premature as you're adding information with regards to Intel's VIIV that's relevant.
Regards,
Rink
Dual core Turion delayed, say notebook makers. Not May, June. : http://www.theinquirer.net/?article=31427
Regards,
Rink
Wouter, I trust him a lot too. Tx! (eom)
Wouter, are you certain that rev G isn't K8L? That article isn't conclusive for as far as it concerns this (at least I didn't see it that way). If you're certain may I ask what the source of your information is to come to that conclusion?
Regards,
Rink
re: Nobody knows that now, but there is great deal of humor in Ronster and rest of this thread attacking Fleck's credibility, when in fact he was right, people on this thread were wrong and are holding a stock that has performed very badly.
Hey, not every current poster was wrong last couple of years!
Fleck has been pretty much ok for a while but it's a big question whether he'll be right this time too.
Regards,
Rink
Kate, Re: VIIV!! vs Live!
Somewhat in contradiction to what you are saying the article does not exactly declare a winner. In fact the only part where Intel seems to 'win' is that it is earlier available in the market place. The systems specs generally seem to favor AMD Live!. Besides that the article is comparing limited system specs, no benchmarks or other impressions. It is premature.
Regards,
Rink
Smooth, IIRC they'll be roughly $3K instead of $15K system containing such a chip. So still pretty high, but much more affordable than it is currently.
Regards,
Rink
AMD recalls "inconsistent" Opterons. Heat + x52/4 != good. http://www.theinquirer.net/?article=31354
EDIT: Just adding to wbmw's post earlier.
Regards,
R
Sold my remaining 40 AMD Jan 20's for about $14.10
- Too much uncertainty about K8L. Performance and introduction date. Charlie somewhat confirms that K8L might be a Q2 thing like I suspected already; basically K8L might well not be the first 65nm product.
- Added performance caused by 3rd gen strain (SiGe) might well not make it to rev F this year
- Intel pulling in NGA release dates (at least comparing to what I thought it would be)
- Possible very aggressive ramp of Woodcrest
- Intel is cutting prices significantly enough to limit AMD's upside to guidance this Q
- Most importantly sentiment might favor Intel's upcoming NGA for a while. The street might think all Intel's advantages in H2/06, H1/07 (until K8L) might outweigh AMD's massive capacity increases.
I plan to trade AMD when I think good news is coming. I might take a stake in Intel this Q before NGA, and possibly trade it afterwards. When we know more about K8L I might get back in. Also if I think sentiment might start changing again in AMD's favor I might get back in too (I think sentiment is often overreacting, but it influences the stock prices often rather significantly).
In addition to some Intel + some trades + owning Spansion shares I'm now looking for another stock with good leverage based on technical fundament. I own Spansion shares based on the same thing (tech fundament - shares, not options, because I think it'll go up nicely over a bit longer time). I'll probably trade Cray before DARPA funds are determined. I'll be looking at eastern europe as well (like Czech Republic, Poland, not Russia).
Regards,
Rink
I'm sorry you can't comprehend it.
R
(Yeah, I know, really last post this time)
Yeah, you take great care in disrespecting some and don't care about being disrespected. And on and on go the pointless debates. Great attitude!
Regards,
Rink
PS, last post from me on this subject.