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.
Mas, you could well be right (eom)
Anyhow since Groo_NLI signs himself as Charlie, and was vigorously defending the Inquirer story, and other ACE board members claimed that NLI stood for "Not Logged In", I made the reasonable assumption that it was the same person.
I tried. 'NLI' is not added to my alias when I'm not logged in. http://www.aceshardware.com/forums/read_post.jsp?id=120083103&forumid=1
Regards,
Rink
The author Groo_NLI is unlikely to be Groo/Charlie (eom)
Mas, just for fun, is this your portrait? http://www.xtremesystems.org/forums/member.php?u=28557
Very funny if true.
Regards,
Rink
Wbmw, tapeout of K10 was halfway through Aug '06 (not Sept): http://www.siliconinvestor.com/readmsg.aspx?msgid=22720176
Also AMD has a NEW mobile cpu + platform in the pipeline. It was mentioned in last falls tech day slides that also included a die photo with explanation of K10. FAFAIK it's scheduled around the same time as the DC version of K10. Educated guesses have circulated that it might well lack the added 128b SSE functionality but have all other K10 benefits. Q2 '08 is two quarters later than I'm expecting it currently.
Regards,
Rink
Mmoy, just curious which direction AMD will end up tomorrow?
Regards,
Rink
Klaus, re: Singapore plays into the equation with low feature-yields, from what I hear.
I haven't seen this being talked about anywhere. No references in articles I read. I saw you write this before, so where does this info come from if I may ask?
Regards,
Rink
Drjohn, The few laptops I've opened have the sides and front taken by hard disk, card readers, dvd writer, battery, and even memory. The middle and the rear are places for processor and graphics unit. Cooling is usually at the back. A graphics card under the palm rest? That might be unusual if the few systems I saw so far are anything to go by.
My laptop is running pretty cool, but that is partly because I don't game much. In the game I play 3x a year it does heat up. It's not as warm though while rendering movies. Most other stuff doesn't take that much processing. Just a thought: If the OS requires 3d (like Vista) then the graphics card might be on all the time.
Regards,
Rink
Wbmw, FP_rate improved 40% over baseline QC Xeon (50-55% over top DC Opteron)
So FP_rate remains a strong selling point for Opteron based systems. Still the demo left me wanting too.
Regards,
Rink
Michael, AMD Analyst day is Dec 14. It was refered to in their last earnings CC where they said they'd talk over all financial implications of the merger with ATI (the earnigs CC excluded all of those explicitly). http://www.amd.com/us-en/Corporate/InvestorRelations/0,,51_306_7654,00.html
Regards,
Rink
Windsock, It's quite probably that the xbitlab author is interpreting the same quote as the others that mentions producing first silicon that will be send back to design after first production is completed (the quote is in his article in a similar fashion as in the other articles - I don't think he has anything else). It means tape out has happened. It's not possible from that quote to draw any conclusion that the first products have come out of the factory, nor that the first penryn is functional.
Regards,
Rink
Sarmad, HP server rev decreased, Dell's server rev increased QoQ as a percentage of total.
I was suggesting this might have something to do with:
1. higher Dell presence in the lower end of the market in combination with better Intel server products, and
2. Dell's increasing AMD server portfolio
Compare the "Worldwide Server Systems Factory Revenue" tables for Q2 and Q3 in the below articles:
Q3 IDC server figures: http://home.businesswire.com/portal/site/google/index.jsp?ndmViewId=news_view&newsId=20061121006....
Q2 IDC server figures: http://biz.yahoo.com/bw/060823/20060822005945.html?.v=1
Regards,
Rink
Sarmad, I think it's fair to say that better Intel products in addition to new AMD based products allowed the increase in results.
Separately, from the IDC numbers one could get the impression that Dell might have gained some AMD marketshare from HP.
Regards,
Rink
Elmer, Market share gains were also at a time when there were less SKU's. With Dell Lenovo and more in general the increasing commercial DTs and NBs AMD can continue to gain unit market share. Margin might decline until H2 '07 when Barcelona and derivatives are suppost to impact results.
Regards,
Rink
Sarmad, X2 is almost twice the size of A64 and it'll therefor yield lower as well. This alone is eating up most of the capacity increase.
Regards,
Rink
Keith, Uhh, right, Apple is an AMD GPU (and chipset) customer.
Regards,
Rink
Keith, Nah, Suggesting that Apple doesn't like Intel graphics might be correct but we'd probably have heard this rumor from other sources too and we didn't.
Sigh, even though I don't quite believe it at this stage it would make sense for integrated graphics specifically.
Regards,
Rink
CJ, the article has been corrected: http://www.channelweb.com/sections/allnews/article.jhtml?articleId=193500828&cid=ChannelWebNews
Regards,
Rink
Sarmad, congratulations with your epsconstest win. My estimates sucked quite a bit, especially AMD's stock price estimate.
What Keith said (50% growth in mobile cpu's both unit and rev wise) I heard being mentioned twice on the CC. I also think I heard a comment about increasing ASP but I think that it can't possibly have been significant -- as long as units and rev grew exactly an equal percentage ASP was flat. So what probably happened is that rev grew slightly more than units but both close to the 50% mentioned.
Regards,
Rink
Elmer, that's bull.
Duke, I much appreciated your insight.
Regards,
Rink
Wbmw, agree. Your rev H thoughts are a bit more detailed than my own where it concerns frequencies - tx.
Just posted this on SI: Intel to marry processor to memory (Through Silicon Vias / TSV): http://news.cnet.co.uk/desktops/0,39029662,49283968,00.htm
It mentions marrying the 80 core prototype chip to 256MB memory (prototype would be marketed in some five years from now). Interesting idea to reduce latency / increase bandwidth.
The article mentions they'll use the same technique for other things. Maybe they mean cpu to cpu.
Meanwhile I still speculate that AMD works with SUN on their proximity connectivity to speed up communication between chips (Intel's technology is for in between several dies in a single chip). AMD can also use older techniques (wires) to link dies in a chip like Intel is about to use for Core 2 Quad (two Core2 dies in a single chip). Intel's new technique should have vast latency and bandwidth benefits, and might only be limited by how much dies can be fitted in a single chip. I like Intel's approach.
Regards,
Rink
re: 3GHz is a launch frequency for Core 2. AMD's 90nm K8 launched at 2.2GHz and single core. It's naive to think that Core 2 won't improve in speed faster than the EOL 90nm K8.
You're comparing launch frequency of a new design (Core 2) on a very mature Intel 65nm process, with that of a mature design (initial 90nm K8) on a very new (i.e. very new when initial 90nm K8 was launched) AMD 90nm process!?!?
I agree that K8 is nearing EOL. Still rev G will improve faster in speed than 90nm rev F because I think it will see more process improvements. Fafaik the roadmap for CMW doesn't feature major speed jumps either. For now it might not be unreasonable to assume that until AMD comes with either the new mobile core, or rev H, that the current performance gap between Intel's CMW line and AMD's K8 line will remain approximately as big as it is now (comparing DC CMW's with DC K8's), with the gap being larger in desktop and mobile than in the server segment.
re: The issue here is the *launch* of 65nm relative to what it can do for AMD's *current* competitiveness. The answer is: nothing. 65nm G-step is less competitive than 90nm F-step, so it's a net step backwards for AMD.
65nm rev G allows AMD to shift to dual core X2's at a fast pace. X2's even at somewhat lower speed perform better and are hence more competitive than single cores. So in the sense that rev G allows a much richer DC/SC ratio rev G is more competitive. Only if you compare an 'average' bin 90nm DC with an average bin 65nm DC you are probably ('probably' because we don't know what 'average' is) right that it is a step backwards.
Obviously, this will improve over time, and H-step has a much greater chance of outclocking Rev-F. But H-step is still many quarters away from mass production.
It seems you're saying that rev G will outclock rev F some time (e.g. two to three quarters or so) after launch. If so I agree with that. I think the design goal of rev H was obviously to improve IPC, not clock speed perse. Therefor evolutionary jumps in speed are more likely imo (comparing DC rev H with DC rev G) than any revolutionary jumps.
Regards,
Rink
You mean next generation 45nm process tech?
Wbmw, yep. Still Conroe uses a very mature Intel 65nm process, while Brisbane uses the very initial AMD 65nm process. AMD's process will improve pretty fast during 2007 (both power and speed).
Regards,
Rink
Most make it public (eom)
I've suspended more AMD investors on SI than Intel investors.
Regards,
Rink
Plyngso (and Petz), This article puts initial server rev H (H1/07) before desktop and mobile rev H which in turn seems to be put before rev H with direct connect 2.0 enabled (start) 2008.: http://www.siliconinvestor.com/readmsg.aspx?msgid=22827569
Regards,
Rink
PS, Petz I saw your post on the subject and thought you might want to see the interpreted indication provided by this article as well.
EDIT: Just saw FPG also mentioned this article.
Nope, No direct words from the horse's mouth, but then I would not expect that directly as that would amount to loosing face.
Regards,
Rink
Chipguy,
That was a direct quote from Charlie. Link was provided. I know you don't agree.
You mentioned in the past having worked on a certain new memory type, was it this one?
Regards,
Rink
Chris, did you read this?
AMD drops FB-DIMMs: http://uk.theinquirer.net/?article=34412
Intel pulls back from FB-DIMMs. Dropped from 90% of servers in 2008. ... Intel is telling everyone who will listen that they are going back to rDIMMs [i.e. registered DIMMs] in 2008 over about 90% of their server line. This is due to them not being able to hit power numbers, seemingly a problem with the tech itself vs implementations, and of course cost. ... http://www.theinquirer.net/default.aspx?article=34220
BTW, AMD never committed to use FB-DIMMs exclusively for servers. They said at times that they'd support the FB-DIMMs if and when appropriate, and at other times that support for it would become available but OEM's would not be forced to use them and would be able to opt for rDIMMs instead.
Regards,
Rink
Plyngso, I think all is said on this topic.
Regards,
Rink
re: How long is a timeframe?
And here I was thinking that you knew that Q2 2007 will be about, UHMMM, about three months long.
re: From Charlie's story, it sounds like Barcelona is Opteron-only. ... But that would mean no desktop new-core until Q4. Not so good.
Psst. I let you in on a secret. Barcelona is Opteron-only.
Regards,
Rink
PS, maybe Barcelona can be used for limited desktop as well. A DC version for desktop would require another code name. I'm not sure if Charlie knows all code names, and the two sources I provided are reasonably sound.
re: The other interpretation of that slide is that at least ONE of server OR desktop will be released by mid-2007. Does a yellow band on a slide imply AND and not OR ? Look at the "enhanced RAS" bar below it... does THAT apply to desktop too, or just server?
Well you have to admit it's second order logic at best, and that it becomes third order logic after looking at this quote from Henri Richard: "Our plan is to have both dual-core and quad-core K8L-based processors within the same timeframe." http://digitimes.com/print/a20060628VL201.html
Just curious: How many sources do you have for claiming that DC versions of K8L/revH will only start to appear in Q4 2007?
Regards,
Rink
PS, Not saying it can't happen, I'm just saying current info doesn't allow you to have a currently plausible basis for your claim.
And there I was thinking you tried to stretch 'mid-2007' to 'q4' for desktops...
Rink
re: From Charlie's story, it sounds like Barcelona is Opteron-only. ... But that would mean no desktop new-core until Q4. Not so good.
Analyst day slides clearly suggest somthing else: http://www.epscontest.com/presentations/06q2_analyst-day.htm?slide=49&a
Regards,
Rink
AMD Roadmap:
Dec 2006: Rev G, first 65nm
Q2 2007: Barcelona, Rev H, first QC with the 128b FPU
Q4 2007: Budapest, Rev H but on a HTT v3.0 platform, both single socket Opteron + A64
Q1 2008: Shanghai, Rev H dual and quad socket Opterons on a HTT 3.0 platform
The above's taken from this article: "AMD quad cores: the whole story unfolded": http://uk.theinquirer.net/?article=34433
Regards,
Rink
Cool! Thanks for asking Ruth. Just yesterday I again went over your viewpoint at SI and I'm greatful that it's sufficiently cleared up for me now.
Regards,
Rink
Duke, it's clear you have no idea what you can say about that article because you don't know more than a couple of words German. Considering your remark it might actually surprise you considerably that German and US laws do have the same origin. So besides that you don't seem to know any German, you might as well know equally little about German law. So who are you to imply that German law in this particular case can't have influence in the US (e.g. on the AMD's US lawsuit against Intel)?
Regards,
Rink
Chipguy, tx! (eom)
Chipguy, This slide puts production of 65nm in Dec. Do you know if this means production start or production out?
http://images.dailytech.com/nimage/2035_large_amd_65nm_updates.jpg
Thanks,
Rink