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.
Paul, based on your mention I tried to find it. I posted what I came up with. I don't think I can answer your question.
(BTW, I posted it at SI because I can search that site and find it back when I'd need it again)
Regards,
Rink
Paul, CJ, here's what I dug up about AMD's relation with KLA-Tencor: http://www.siliconinvestor.com/readmsg.aspx?msgid=22136506
Regards,
Rink
Yep, always good to see significant inside purchases.
My position in SPSN is only an initial position. It'll increase when each the first two conditions you mentioned are being met (90nm ramp + 65nm EOY, and ORNAND ramp). I sell once I get any flash overcapacity signal (probably too late but I'll try non the less).
Regards,
Rink
Tx (eom)
Chipguy, That puts it in a much better context. I couldn't believe that the 1.5GHz I2 outperformed a DC Opteron by that factor 2 on a whole range of HPC apps. This though is much easier to believe.
Rereading the thread I admit I missed your remark about outperforming DC Opteron "by anywhere from a little to over two to one". If I would have read it better I wouldn't have questioned you. Mes excuses.
I don't know any of these apps first hand, and this lower level I2 is doing better than I had guessed it would in this area.
Thanks for providing the context.
Maybe one more question: Can I take it these apps are typically multi threaded?
Regards,
Rink
Mmoy, what new stuff if I may ask?
Regards,
Rink
Chipguy, can you at least mention the applications then in which 1.5GHz I2 / 4MB outpaced the DC Opterons by that factor 2?
Regards,
Rink
Chipguy, re: I recently read a competitive analysis report which showed how the low-medium 1.5 GHz, 4 MB bin of the old 130 nm I2 outperforms dual core Opteron on a whole range of widely used HPC applications by anywhere from a little to over
two to one.
What's the die size? Twice that of Opteron as well?
It ofcourse depends on the exact parameters but I think there is an awefully good chance that it is in contradiction with what I read. It looks to be in contradiction with SPEC results as well (not the same as HPC apps I admit, but still your remark is so out of wack with the SPEC results I'm rather doubtful it is true). How recent was that report? What DC Opteron? How many processors / system? Actually could you please provide a link (maybe you are mistaken and such a report does not exist?).
Regards,
Rink
True. Anyways, the gossip sites forgot to mention too that in addition to a sequential decline in rev ($26M decline to $144M), and operating profit (2$M decline to $28M in operating losses), Accounts Receivable declined as well with $23M to $70M as well as Inventories that declined with $13M to $62M. Decline in inventories is ofcourse normal for a seasonally strong quarter, just like risen GM. Decline in Accounts receivables though particularly in this quarter is not a good sign. Suffice to say neither is the rev decline. Ofcourse there's always new products down the pipeline, some of which have very recently been released (the slip in release of these new products could explain some of the backlog growth), while Montecito based products will have to wait till summer. All in all SGI almost certainly will continue to bleed out of its main arteries for two more quarters; at least imo till Montecito based revenue starts making up for current IRIX/Itanium(Madison) System Rev declines. SGI does have some good tech for a good price. If anyone (NEC/Fujitsu/Bull/...) would want it now is the time.
Regards,
Rink
Silicon Graphics loses CEO, blames Intel
Itanium champion shows further losses
http://www.theinquirer.net/?article=29405
Was to be expected. In the off chance that H2 doesn't get them substantially better results this spiral of death will soon deplete their cash. In all likelyhood the Osbourne effect has only begun to roar its ugly head.
Could in fact become a nice mid term play. I can't consider it seriously though as I don't have access any ticker not listed on Nasdaq/NYSE.
Regards,
Rink
Buggi, re: Could be that Conroe is 5% faster
than X2 under 32Bit, but my feeling (today) goes, that X2
could lead under 64BIT OS/64BIT APP by as much as 50% under
special applications (or more) and lead with 10-25% on
avarage - depends on how we put the benches together.
Why do you think the 64b implementation in NGA will be that bad? Do you perhaps think they borrowed from 64b in Netburst?
Rink
Rink
Enlightning report. Tx!
R
Or perhaps rev F will be available for notebooks and servers first, but servers take more time to evaluate. Anyways, like CJ and Keith, I'll believe it when products are actually sold via Dell's website.
Regards,
Rink
re: It´s shame that the notebook isn´t sold worldwide.
Agree. I think it would be even more of a hit in Europe.
Regards,
Rink
I_banker, something similar here. BTW here's where I recently started to post most relevant Spansion info: http://www.siliconinvestor.com/subject.aspx?subjectid=56255
Regards,
Rink
Keith, here's some talk on that 7M block / move above $40 events (too late to edit previous post; I should've read SI first like usually):
http://www.siliconinvestor.com/readmsg.aspx?msgid=22103749
http://www.siliconinvestor.com/readmsg.aspx?msgid=22103790
So theories for the move above $40 that I see are:
1. Institutional block took away $40 resistance.
2. Santiago Dell story reiterated on 'news' that Dell still is 'open' to using AMD.
3. MSFT earnings speculation / leakage
I think we can't puzzle the answer together without more info. There were simply too much things going on yesterday.
Regards,
Rink
Keith, the 7M block trade yesterday was discussed at SI as to probably have taken away resistance at $40. The trade didn't influence the price of AMD common immediately (so one institution sold to a willing buyer). Hence I thought the resistance breaking thinking made sense.
Regards,
Rink
Thanks Keith! Re: Galaxy volume started ~12/14/2005: http://www.investorshub.com/boards/read_msg.asp?message_id=8877749
I remembered that post but couldn't find it back (not a paying member here). Hence I knew volume availability was late in Q4, but this date is even a lot later in Q4 than I remembered. Thanks. In this perspective the 8000 Galaxies in Q4 weren't that bad at all.
On top of that the Galaxy line isn't complete as the high end (lower volume) systems are still missing. They're starting to get overdue now.
Might have been mentioned here before: It's possible that some of growth in SUN's backlog is because of Galaxy's volume availability ariving a lot later than most expected it. Does anyone know / can make an educated guess as to what the biggest components of that backlog growth are?
Regards,
Rink
For those interested: Over $40 and shooting up.
Regards,
Rink
Bobs, you're invested in anything besides this?
Regards,
Rink
It's not certain that it'll have a fourth execution unit that you mention. The part is known to have a four issue decoder, meaning that it can decode 4 x86 instructions at the same time. This doesn't mean much for as far as the amount of execution units goes.
Regards,
Rink
Didn't mainstream Galaxy only become available in volume halfway through Q4?
Also the Galaxy line isn't exactly complete yet.
Regards,
Rink
re: For 2007 he estimates earnings of 51 cents on revenue of $6.7 billion.
What?! Drop of more than $1.2 in profit compared to 2006, on 8% more revenue. He obviously expects Intel overtake AMD rather significantly end 06.
Regards,
Rink
He doesn't follow SGI under the microscope either. When I mentioned SGI wasn't exactly growing it's server lines a while back (comprised of both MIPS and IA64) he didn't know what I was talking about.
Regards,
Rink
Doug, actually Intel said that ("AMD is eating its future") before wbmw tried to make it a personal trademark. At the time that Intel said it there was a grain of truth in it. I think it was at the end of the 180nm cycle. Since then this hasn't happened anymore, not to the same extent at least fafaik. And Intel never said something along these lines again either. Wbmw however didn't stop.
Regards,
Rink
Woodcrest will make a very interesting combination with IBM's X3 chipset. Because of this IBM has very little incentive as of H2 to use Opteron except for in the niche where it currently already is.
Regards,
Rink
EDIT2: Thanks Alan. I didn't know that at least some of Intel's 200mm facilities do 90nm. I'm positive I saw that sheet you referenced before though. I think it has been updated. It mentions fab 18, 200mm, doing 90nm flash (fab D2 might as well but it's much less clear as they're doing 180nm as well, and produce logic too and are a development site on top of that). Anyway fab 18 is listed as starting with 90nm flash in Q4 just as had previously concluded (I thought it would come from a 300mm line though which turns out to be incorrect).
Regards,
Rink
90nm 300mm "communications" is flash.
Intel is producing 90nm flash as of Q4 last year I think. It can only come from a 300mm facility. Hence I think there's a good chance that "communications" means flash (for cell phones). This should at least be a partial answer. BTW Intel was also producing flash at 130nm (not only 180nm as was previously implied). They never had flash on 110nm like Spansion.
Regards,
Rink
EDIT: One example of 90nm flash (didn't go as far as looking up the details): http://www.intel.com/design/flcomp/prodbref/m18.htm
Spansion CC relistened / CC Notes: http://www.siliconinvestor.com/readmsg.aspx?msgid=22082967
Regards,
Rink
Keith, Ruiz used to be optimistic about Dell as well. Didn't happen. At least not in the timeframe he had in mind. I do see this as an exception though as by far the majority of things he has been positive about have happened.
Regards,
Rink
Keith, re: Inventories.
Rough idea by Petz:
- http://www.siliconinvestor.com/readmsg.aspx?msgid=22074851
- http://www.siliconinvestor.com/readmsg.aspx?msgid=22074882
Regards,
Rink
EDIT: Keith, re: Spansion CC.
Ssee post #18 and #21 here: http://www.siliconinvestor.com/subject.aspx?subjectid=56255
Maybe a quick glance at the roadmap in #26 might help a bit as well.
As for the CC I did here it but was multitasking reading prelim AMD results that make up a much larger portion of my portfolio. I need to relisten, but the above gives you my first impressions.
EDIT: Yep too optimistic. I think it's has quite a good chance of being a very good long term investment. I still don't know why the $37M loss was higher than I anticipated which is why I need to relisten to the call.
Regards,
Rink
EDITED: SP, It's revenue share, not unit share. AMD probably has between 21-22% unit share. TWY and others tried to come up with a reasonable idea for unit share and ended up with ~21.5%. See here: http://www.siliconinvestor.com/readmsg.aspx?msgid=22075422
He also did a rough calc for ASPs for both Intel ($148, down $4) and AMD ($98, up $6): http://www.siliconinvestor.com/readmsg.aspx?msgid=22075675
Regards,
Rink
Chipguy, have you heard about Z-RAM, a 1T mem cell design that AMD just licensed? If so what is your opinion about it?
AMD licenses Innovative Silicon's SOI memory: http://www.eetimes.com/showArticle.jhtml?articleID=177101749 (via NicoV/SI).
Link to Innovative Silicon Inc.'s website: http://www.innovativesilicon.com/
Quote: In present System on Chip (SoC) applications, memory already dominates silicon area is steadily increasing with each generation. The most common types of embedded memory in current use are 1T/1C DRAM and 6T SRAM. As CMOS technology achieves sub 100 nm geometries, new memory devices are being considered for DRAM/SRAM replacement. However most of these new memories rely on the integration of exotic materials into a baseline CMOS process and require relatively large cells. Innovative Silicon has developed a true capacitor-less, single transistor DRAM - named Z-RAM for Zero Capacitor DRAM – by harnessing the floating body effect of Silicon on Insulator (SOI) devices. This technology is capable of achieving twice the memory density of existing embedded DRAM technology and five times that of SRAM yet requires no special materials or extra mask/process steps.
Roadmap extents to FD-SOI and FinFets as well: http://www.innovativesilicon.com/en/technology_roadmap.php
Regards,
Rink
CC notes from AMDZone's Chris Tom: http://www.amdzone.com/modules.php?op=modload&name=Sections&file=index&req=viewarticle&a...
Regards,
Rink
Ouch!
Chipguy, well, then we think differently.
Regards,
Rink
Chipguy, re: The phrase "high level microarchitecture" means it
is clear that AMD are still evaluating pretty basic
design trade-offs. Which means they haven't started
seriously coding this up in RTL let alone begun any
physical design.
Sorry, but that is BS. AMD, just like Intel, always has teams working on high level architecture. It doesn't necessarily say anything about when which design has or will be moved to coding in RTL, or physical design.
This morning I sold a bit more of AMD. Together with previous sales this month I'm half out of AMD now. I feel comfortable with my remaining position. Do you hold any AMD/Intel currently?
Regards,
Rink
But I don't find it serious.
Well the bug effects both sales and image. So what what on earth makes you think that Intel or anyone else will talk about it as "yawn", "bios fixes are common", "May hang, get it?", and "I don't think it's serious"?
I believe AMD has had their share of incidents like this (I know they have).
Sure. Pretty long ago, and back then it was serious too as it effected sales and image.
Just a bit more context: AMD has been executing rather nicely the last couple of years. Intel just showed that it is perfectly able to continue it's years long string of P4- and Itanium related flukes.
Or maybe you meant "yawn" in another context, like "Yawn, another blunder"?
Regards,
Rink