Analysis on IT trends and competitive strategies, with emphasis on micro processors, computer systems and networks. Based on latest news, backed up with real data, this site intends to provide a true and realtime picture of the fast changing IT landscape. This journal strives to be accurate on facts and sharp on criticisms. You may email your opinion to sharikou@yahoo.com or post comments here, be cool and intelligent.
Freelance journalist on IT matters. Some of my writings have been published on online IT journals. Any original content on this journal is Copyrighted, but it's free for non-commercial use. Any Trademarks used on this site belong to their respective owners. Some of the pictures are links. If there is any issue with the content of this site, please email sharikou@yahoo.com .
The truth is out. 2P 3GHZ Opteron beats 2P 3GHZ Woodcrest in Specint_rate2006 by 2.5%.
posted by Sharikou, Ph. D. @ 4/04/2007 10:59:00 AM
102 Comments:
This is the fundamental issue - all we CAN do is imagine what Barcelona will do because there are no valid third party benchmarks...
And what about those validations on your ph.d? Any progress?
Intel's best cannot compete with AMD's current generation. Barcelona will leave Intel in a grotesque pool of shame and embarrassment. To make things worse, Intel will surely try to twist the situation with their fabricated benchmarks. Only fools buy Intel.
I thoroughly agree with penix. Also Intel's Penryn is a joke. It's just a simple die shrink. Inteler's, you should not get your hopes up because Intel will not have a new architecture in near future.
what is the point? AMD is dead. The stock has lost 66% of its value. The whole thing is moot.
what is the point? AMD is dead. The stock has lost 66% of its value. The whole thing is moot
If you want to discuss stock price go to some other blog you MF.
Lately I see a lot AMD flamers in this blog, I wont be surprised , if thay are all paid by Intel , to write BS on this blog. The way Intel marketing machine works is , they wont leave any store unturned , even if it means hiring some hacks to post on blogs etc. It is so obvious they do operate worse than mafia style. I even read some comments today that goes by "Intel to bury Barcelona blah blah " , I think those idiots that Intel can fool this time is probably those idiots like tim and roborat posting on this blog and who are definitely paid Intel hacks.
Interesting comments:
penix
Barcelona will leave Intel in a grotesque pool of shame and embarrassment.
Tommy
Intel's Penryn is a joke. It's just a simple die shrink.
In the best of scenarios... and I truly hope your right and I'm wrong. But being a 'realist' I have to be ready for alternate possibilities.
The "what if" scenarios...
1. Barcelona is NOT great, but simply good?
2. Penryn, although just a die shrink with minor tweaks is NOT simply good, but is GREAT?
What then? Does that make each processor on par? Or still gives Barcelona a very small lead?
What if Intel drops down prices by 50% to 75% by the end of the year?
Will AMD still be able to sell profitably in the high end?
Will Torenza products be ready? If so, will they help in sales or selling at a premium?
netrama
"If you want to discuss stock price go to some other blog you MF."
I'd go to http://trackingamd.blogspot.com/ but it closed a while ago, perhaps after it saw how AMD stock kept lowering and R600 constantly pushed back.
$12.71 Today.
On it's way to $0.00
Dear netrama,
don't use abusive language.
The way AMD has handled Intel is pathetic. AMD had a superior product for at least two years and that took them nowhere. Finally AMD managed to get into Dell, but the market perceives C2D as at par with AMD.
Does nay one know how well Dell/AMD machines are selling? What is the situation in the server space? Sun's courting of Intel was another blow.
Did anyone here praising AMD actually read the full article? Their choice of benchmarks, test setups, and marketing words are just as bad if not worse than Intel's last presentation. AMD just presented numbers in a similar fashion as Intel did a few weeks ago and nobody around here is going to call them on it? By the way, Barcelona is showing about a 8~10% advantage in general applications and game testing over the C2D at same clock speeds in internal testing now. That is why you are not hearing anything about it.
Intel releases quad core cpu's for embedded computing.
http://www.eetimes.com/news/semi/showArticle.jhtml?articleID=198702225
How long before AMD issues a press release yelling "Me too! Me too!"?
By the way, Barcelona is showing about a 8~10% advantage in general applications and game testing over the C2D at same clock speeds in internal testing now.
If that is true, then it's a total disaster for AMD.
By the way, Barcelona is showing about a 8~10% advantage in general applications and game testing over the C2D at same clock speeds in internal testing now.
Could you please clarify the term "internal"? Does that mean alpha or beta?
And if what you say is true, I would say we are headed for a uniplayer market!
Making the market equal again. An Barcelona isn't even out yet. To stock buyers the AMD-K6 stock if i remember correctly was at $3. So what does all the crap in here mean...nothing..we will still have 2 key players in the cpu business...live with it..
No ace up their sleeve huh?
:), yeahhh!!!
Go Opteron!!!
Go AMD-ATI!!!
P.S.
There's a new idiot on the forum!
Hello bubba!
Intels mercenary!
My team: SETI.USA has been using the performance of Intel Core 2 processors to benefit the distributed computing community.
Lately we tackled the project called Spinhenge. We first started AMD based computers were leading... however once our Woodcrest and Clovertown Xeons joined the fray the AMD's fell in ranking because they could not match the performance of Intel. SETI.USA has not been able to optimize this projects executables.
Do you have an answer as to why so many Xeons are now leading this previous AMD stronghold?
If you do, the AMD community for this project would appreciate your help.
Here is the proof:
http://spin.fh-bielefeld.de/top_hosts.php
Dude
Where did you get your PhD at a online school using AMD computers, LOL
Everybody knew in the netbust days that you could easily find a few benchies that Prescott/CedarMill could beat the AMD Operton, but it was common knowledge that AMD was a far suprior product at every price point. It was because of this it was embraced by performance user and they took 30% MS of the server by storm.
Its also obvious today that Core2 architecture is far suprior then the current Operton. But it is also no surprise that there are few benchies where Operton still wins. But the bottom line is AMD had to drop prices greater then 50%. Why because the weren't competive anymore with INTELs top end lineup. Sure there is a benchie here and there AMD wins but over all AMD product value was cut in half by Core2
Barcelona as been 4 years in coming. Frankly if it isn't as huge a step forward as Core2 was to Netbust then it will have been a failure and all the AMD folks should be fired and a Phd with credentials similar to the pretender here should be hired to design its replacement.
Barcelona will be 20-40% faster then Operon. Oveally it'll probably beat Core2 at same pwer envelop by maybe 10%. But that simply won't be enough. AMD is going to try and ramp this complex new quad-core on a new process, with new tools, in a new factory. Yields will be low and excursions often. It'll be 2008 before they iron all this out. In the meantime Penrym and Nehalme have already taken everyone's mindshare. It was brillaince on INTEL's part to annouce all the details about Nehalem and Penrym. That plus all the hoopla about HighK metalG and 4 45nm factories will keep everyone waiting for Penrym, even if it is only marginally better the damage will have been done to AMD, everyone will have waited and since they waited they will wait again for Nehalem since it will be there soon. IN the meantime Barecelona volumes will start to arrive but everyone will be thinking Nehalem.
AMD should have released benchmarks and demo'd the product to build anticipation. Now INTEL has stole all the thunder and got everyone thinking Tick Tock Tick Tock and they delivered Core2 everyone believes they will deliver Penrym and Nehalem. They got HighK metal gate glamour.
It is SO obvious what AMD should do. If Barcelona was so good they would release it and let the benchmarks leak. 40% speead up and 10-20% speed up over Core2 would sieze all the headlines back. People would believe AMD got its mojo. The only reason AMD is with holding is because those benchies aren't so great.. or they got some serious problems with B-step and they are praying a c-step fixes everything. THey got problems or they'd be talking.
Sorry PhD pretender your calls of "fire " don't work. I can't believe people actualy believe AMD doesn't have some seriouse problems. Even the die hard realist should acknowledge the very odd behavior and statement by AMD so close to a launch under such financial and credibility duress.
AMD got nothing right now if all they got is 10% on internal benchmarks that is a disaster and no wonder they aren't publshing any benchmarks. The stock would plummet to below 10 if that is true.
P.S.
There's a new idiot on the forum!
Hello bubba!
Intels mercenary!
Funny, I'm called the idiot, yet I've never resorted to name calling.
So, since you're smarter than me, please tell me what AMD's stock price will be after a $1B bond placement.
By the way, Barcelona is showing about a 8~10% advantage in general applications and game testing over the C2D at same clock speeds in internal testing now.
It's all clear now. Barcelona is already pre-fragged by Intel's current Clovertown CPUs. With this new development Intel won't even need Nehalem, Penryn will be 45% faster than Clovertown ensuring AMD's BK. AMD is currently burning through cash at an incredible rate, they cannot afford to shrink to a 45nm process.
Maybe Hector Ruiz needs buy himself a MacPro; the fastest x86 based workstation in the world. He can work these predictions out much quicker on Intel rather than with AMD's Crapteron.
AMD BK Q2'08.
P.S. See that share price? AMD is worth less than $7bn now! A few more months of this and the combined DAAMIT will be worth less than what AMD paid for ATI last year!
yeh.. definitely... when both of them aren't running the same operating system (Linux for AMD, Microsoft Server for Intel). its a really good comparison of apple to orange,
also, on the side note, the link you posted actually question the validity of the benchmark, since AMD used mixing and matching method to present their slide.
http://blogs.zdnet.com/Berlind/?p=417
:)hm,
giant, lex
you have always the same stories.
You come here to do what???
To defend Intel??
If Intel is so strong, why are you protecting him?
You are so affraid, so paid just to speak bullshit and tell lies!!!
Reuben, be patient.
One month away of the 'AMDs Barcelona' first benchmark.
lex said...
'If Barcelona was so good they would release it and let the benchmarks leak.'
They are not like Intel, that published the first benchmark of Conroe in 2005, 9 months before you could buy one.
Do you want to know why?
Because they were so miserable, years and years living in lies, (today is the same thing) thay thought that nobody would believe them.
AMD is playing smart, like in poker. Why would they show their cards.
We all know that Intel is copying AMD.
LOL
dr blog
if that's the case, then please show me a benchmark that really represents a true comparison between 3Ghz Xeon and 3Ghz Opteron, with very similar system configuration, and very similar testing software.
also, please show me a site that prove Microsoft WIndows is written for Intel. let's say that I know nothing about CPUs, and would like a proof to reinforce your statement.
quoted from David Berlind, "But, the third chart that AMD is showing today is special. That's because neither Intel nor an Intel-based system manufacturer was the sponsor of the Xeon 5355 test. AMD was. Normally, I'd say this is great. But what confounds me is how, when presented with the opportunity to control the configurations of both an AMD system and an Intel system, AMD did not configure the systems as identically as it could (so as to, as best as possible, isolate any performance differences to the processors). Instead, whereas the AMD-based system is configured with 8GB memory, a 250GB SATA disk drive and Novell's SuSE Linux Enterprise Server 9 SP3, the Intel Xeon 5355-based machine is configured with 16GB memory, an 80GB SATA disk drive, and SuSE Linux Enterprise Server 10. While the non-processor aspects of the two configurations are probably closer to each other than in any of the other benchmarks AMD is showing today, it's still not close enough. In other words, it's not apples to apples."
in other words, AMD was the one configuring Xeon 5355s machine, which may compromise the neutrality of the test.
another thing to be noted is that, SPEC_fp rate 2006 is not a program specifically written to compare the raw performance of CPU, but rather what server configuration will suit a business best. in other words, SPEC_fp rate can be used for platform comparison, but not an ideal program for component comparison (i.e. CPU).
Dr. Blog,
You will almost surely lose all that money. You do know what a private bond placement is, right?
You know bonds are secured debt, and stocks aren't, right?
The day AMD announces a private equity infusion, their stock price will drop 20% at a minimum.
On a somewhat related note, it'll be interesting to see Intel and AMD's financial results on or around the 17th of April, in less than two weeks.
Hey Sharikou, what happened to those "massive operational losses" that were supposed to start in Q3'06? Call me crazy, but it seems that for Q3'06 and Q4'06 Intel made a profit of well over $1bn each time.
Intel is gaining marketshare in all segments, it's ASP is rising as more people buy high margain products and leaves AMD to fight a hopeless price war against an opponent with far more cash and far better products. AMD loses money while Intel maintains a nice profit margin in excess of 15%! Who was going BK in Q2'08 now Pretender?
AMD has taken on billions in additonal debt and is burning through it's cash at an alarming rate. Imagine what will happen to AMD's already awful share price when they have to borrow money and take on additional debt to take on a relentless opponent like Intel.
AMD will be severely punished for trying to take on Intel. The punishment is having their company driven into the ground through "fair and open competition" just as AMD wanted. Good riddance to AMD. AMD literally sat around doing nothing through 2006. They started off 2006 with dual core 2.6Ghz CPUs. They finished 2006 with dual core CPUs at 2.8Ghz. The only enhancement was a new socket that let them use DDR2. The result? A less than 10% increase in performance. Pathetic. They release slop like this throughout the whole year and wonder why their profits disappeared?
Intel's 45nm fab in Oregon is already cranking out 45nm products. Intel can stockpile these ready for the Penryn launch sometime in the second half of 2007. As soon as AMD does a paper launch of Barcelona in "late summer" Intel can just answer right back with a hard launch of Penryn: AMD is instantly fragged. The situation will not get any better for AMD over time.
AMD will post massive operational losses in less than two weeks. This will continue until AMD is BK.
AMD BK Q2'08. Sooner if they can't find anyone to lend them more money.
Are you stupid yomamafor2?
What are you talking about?
Linux, Windows??
Do you realize that the new architecture from Intel is the same or even slower then 4 years old Opteron!!
Please, shut up.
Intel's Penryn is a joke. It's just a simple die shrink
Not true at all. Penryn will be 45% faster than existing quad core CPUs. We know that Barcelona will max out at 2.3Ghz and be 10% faster per mhz than Clovertown. Therefore Barcelona is as fast as a 2.53Ghz Clovertown. But Clovertown goes to 3Ghz! Clovertown will be 18% faster than Barcelona. But Penryn will be 45% faster than Clovertown. The result? Penryn will be over 63% faster than Barcelona.
AMD is still using ancient 90nm processors. Intel is 100% 65nm, aside from the extremely advanced 45nm processors coming later on this year.
AMD is old and outdated, old and outdated junk gets thrown out.
AMD BK 2Q'08.
I also read reports that AMD is rigging the benchmarks. This comes as no great suprise to me. AMD has faked benchmarks on it's website comparing desktop processors. They include lots of graphical tests and then give the AMD proccessors a faster video card. What kind of crap is that?
The sooner AMD and it's deceitful and dispicable business tactics are gone the better off the industry will be. AMD is forced to sell X2 3600+ at $65. Even at that price no one wants that crap. They are willing to pay the extra for a Core 2 Duo to get real performance. Not this ancient 3-issue cores crap that AMD is continuing with, even with Barcelona. AMD stock is about to crash. Anyone with AMD stock, get out now if you know what's good for you!!
Calling Lee.. I can only wish I can play poker someday with a brillaint guy like you.
Do you know the difference between poker and business.
Do you realize part of the game of poker relies on getting your opponent to fold without every showing your hand.
Last I check that don't work too much in the land of business. Oh maybe it does.. You got no hand so you keep it secret so you can milk favorable financing for you company as you are in a huge cash crunch.
Imagine the impact if you were to show bone crushing benchmarks in demos and it got leaked you were going to crush your opponent. Damm what do you think the bankers would be doing..
Imagine now if your benchmarks were crap.. and you had billions in debt and billion in cash flow issues... what kind of financing would you get now.. would you find any equity buy out possiblities.
You are right AMD is playing poker it can't show its cards because it got nothing. If it had anything it would be in its best interest to show it.
You and everyone here already knows Penrym and Nehalme are locked and loaded nothing AMD will show can change their performance now. Just like Prescott, Cedar Mill, Smithfield etc. etc. Nothign AMD showed changed the crap INTEL produced. It takes too long for any company to react to a 6 month peek at a benchmark.
REASON AMD DOESN"T SHOW THE HAND IS BECAUSE THE HAND IS CRAP AND WILL RESULT IN SERIOUS STOCK AND DEBT FINANCING ISSUES THAT AMD IS SEEKGIN TO FIX THEIR CASH FLOW PROBLEM
AMD GOT NOTHING
Sharikou,
You should remove some of the idiots flaming AMD here , they are clearly Intel paid mouse-clickers. They repeat the same non-sense time and time again. And sound like an extension of the Intel marketing (aka lying) machine. Your blog is often reviewed by some big time decision making idiots , who might take these lying flamers verbatim. Such false negative spin wont do any good for AMD.
Why? Does the truth hurt you? I agree that it's never been a worse time to be an AMD fanboy.
AMD BK Q2'08.
You know bonds are secured debt, and stocks aren't, right?
The day AMD announces a private equity infusion, their stock price will drop 20% at a minimum.
No debt is secured!! Some will be secured against some assets, but each round of investment will usually have priority on assets over others. It's easy to manipulate 'secured' assets.
Second, do you actually think that AMD can afford bonds if they can't afford a margin? Think!
The only types of bonds they can afford are convertible bonds. Convertible into stock at maturity. Sure stockholders would get diluted, but in my opinion, if these bonds could wipe out the debt, that's exactly the kind of company I would want to be in. Like I tell people... I'd rather have 50% of something than 100% of nothing.
Third, if AMD is smart, they'll dilute in waves.
Issue convertible bonds:
$1 billion due in 5 years
$1 billion due in 10 years
$1 billion due in 15 years
$1 billion due in 20 years
Each would be convertible in stock at a discount of the 6 month average stock value at maturity. Actual discount price would depend on length of maturity.
This way AMD wouldn't have to pay interest, wipe out their debt and yet would dilute stockholders very, very slowly.
Some of you that are crying wolf on the issue of benchmarks need to take a step back from the computer screen. Since when has ANYONE ever fully trusted In house benchmarks, people usually never do and their meant to be taken with a grain of salt. Even when C2D stuff was released early I looked at it but didn't take it as gospel like some people here seem to do, wait until somewhat unbiased 3rd party websites can get a hold of some ES chips or something, but my god, just becasue a certain benhcmark says this doesn't mean its 100% for real on either wide intel or amd
This comment has been removed by the author.
"Do you realize that the new architecture from Intel is the same or even slower then 4 years old Opteron!!"
if you're so certain that Core architecture is the same/slower than K8, then please show me a series of benchmarks by third party sites that support your statement.
As they already used different OS'es and vastly different hardware I wonder why didn't they use this one fr Intel. It would have beaten AMD by ~6%.
i just observed something on this forum:
"I hate Intel, because they suck. I like AMD because they rock"
response from other users: "I'm glad to see someone who has true knowledge to be posting."
"AMD is in deep trouble, because the current Core 2 line handed their arses to them. There is a possibility of AMD going bankrupt by Q2 '08."
response from other users: "omg you're such an Intel fanboy. Shut up and get out of here."
More bad news for AMD. They're going to need to raise another billion dollars to keep from going BK this year.
http://blogs.barrons.com/techtraderdaily/2007/04/04/amd-merrill-cuts-estimates-sees-need-to-raise-1-billion-to-meet-07-cash-needs/
That link said:
"ThinkEquity’s Eric Ross warned that 2007 is likely to be “extremely weak” for AMD; he repeated his Sell rating and reduced his price target on the stock to $10 from $12. His 2007 EPS estimate goes to a loss of 65 cents from a profit of 40 cents; revenue go to $6.55 billion from $7.12 billion"
I sure hope AMD doesn't cripple itself too much. If it does then say bye-bye to cheap CPUs.
No debt is secured!!
You are ignorant beyond comprehension.
"Dr Blog said...
Its also obvious today that Core2 architecture is far suprior then the current Operton. But it is also no surprise that there are few benchies where.......
Did you mean "far superior than"7
Why is it that the Intel "Fanboys" cant spell, or have any idea of the correct grammatical usage of the English language? It must be their "Superior Intellect" I guess.
The Doctor"
As opposed to what? Sharidouche's blatant falsification of doctorate credentials? And blog posts in broken english that sometimes border on incoherent. You're like the pot calling the kettle black, my friend. Think before you type.
"Dr Blog said...
So, since you're smarter than me, please tell me what AMD's stock price will be after a $1B bond placement.
Not sure Buba, but I see acquiring their stock as the best "Get Rich Plan Ever!" I plan to sink every dime into it with every forthcoming paycheck I receive. I'm looking forward to early retirement.
The Doctor"
Tell me how that cardboard box works for ya...
Dr Blog Wrot: "I love this stuff! Microsoft was practically written for Intel processors, and when AMD kicks their ass all over the OS the Intel FB's cry foul. Some things are just "Priceless!""
Lol. *YOU* do know that 90% of the code for XP 64, Server 2003 64, and Vista was written on AMD systems right, using AMD64 as the 64 bit instruction set? Not Intels 64 bit code set (which is just AMD64's code under existing cross licensing agreements).
Systems manufactured with the AMD TV Wonder Digital Cable Tuner have a bug that has halted shipments of the new PCs.
Quality so low their partners can't even ship their systems.
http://www.dailytech.com/AMD+OCUR+CableCard+Bug+Delays+Media+Center+Shipments/article6805.htm
Merrill Lynch comments on AMD (NYSE: AMD - News) after they a chance to sit down with the company's top management recently to see just how deep the hole is. ...
The market may still not appreciate just how much money AMD is likely to lose in Q1 and Q2 as it struggles to work off 90nm product inventory. They are revising their earnings estimate for 2007 down again, to a GAAP loss of $1.29, and they expect AMD to burn through about $900 million in cash by the end of June. The firm is Neutral - AMD's near-term problems are too great to support a more positive stance even at the stock's current level.
The bad news is that detailed analysis of AMD's cash situation indicates that the company likely can't get to the end of the September quarter without an equity financing in the $1 billion range. Investors need to remember that AMD's ability to offer additional debt may be constrained by the need to pay back the company's existing bridge loan.
Sure looks like AMD is going to bankrupt Intel.... LOL.....
OMG, this is getting too easy!
Intel launches 10w ultra mobile Core 2 Duo.
http://www.eweek.com/article2/0,1895,2111633,00.asp
Maybe AMD will re-release K5 and call it a low power cpu.
Food for thought...
"A sputtering start to 2007 has resulted in more than 100 layoffs at Markham's Advanced Micro Devices.
AMD, a leading chipmaker and Silicon Valley archrival of Intel, has been forced to operate much leaner as spending soared as quickly as the California company's debt.
The result: 375 employees from AMD's 16,000-strong global workforce have been handed their walking papers.
Nearly half of those were former ATI Technologies workers.
About 130 workers at the Hwys. 404 and 7 building have been laid off from various departments since last December, AMD spokesperson Dave Erskine said..."
"In the meantime, more job losses at AMD in Markham are likely, he said."
Here is the story.
Is it all so clear now why there are NO Barcelona benchmarks..
AMD is desperate to get financing, and release of crappy benchmarks will kill their chance to get the financing they need to feed cash crunch.
AMD BK in 2008
:)hm, how many idiots!
We'll see in 20 days!
Go AMD-ATI!
LOL
Although technical benchmarks are of great interest SALES performance is more important.
Intel while selling core 2 and quad core 2s took a 4 billion dollar decrease in sales for the 2006 year.
AMD while selling slower opterons had a 4 billion dollar increase in sales for 2006.
Intel had lost market share to AMD for 6, yes 6 years running.
I assure you that Intels loss of 4 billion dollars in cpu sales in 2006 is very, very, very, very, serious business.
Intels expenses are thousands of times higher than AMD and intel is in serious trouble financially due to the 4 billion dollar loss to AMD.
AMD 4 billion dollar increase in sales for 2006 have left AMD in fairly good financial shape with the future looking brite.
Intel may well be the next DEC or Enron.
Maybe Intels past has caught up with them?
Maybe they can make stuff, they just cant seem to sell it anymore.
Many writers say Intel is doing alright, but Intels sales figures indicate they are loosing ground faster than ever.
A 4 billion dollar decrease is huge by any accounting estimates.
Dude..
4 billion is huge, recall INTEL was doing 8 billion profits and this past year they dropped that to what 4 and change..
What happened.. got mad that chimpzilla was stealing their bannas. Got going, released Core2, dropped prices stopped market share errosion, acclerated silicon and factory builds, pulled in 45nm HighK metal gate and will have 4 factories humming on 45nm pumping penryms and Nehalmes. I expect Penrym 6 Meg to be going for 200 bucks in 2008.
Sure AMD sales were way up.. and what happened? Saw a little green.. got billions in revenue and what hundred million or so increase in profits.
DUDE it takes 4 billion to do the 45nmn factory, 2 - 3 billion to develop the process and another 500M or so to take a x86 product line to market. What did AMD do.. waste 5 billion on ATI another loss leader.
THey have a bankrupt strategy.. no win no way...
That is the sad facts of the business. Bencmarks along don't win. YOu need a business plan, Hector blew his load when he was hard and now he is limp as a noodle and no matter what he tries he won't find postive cash flow for all of 2007 and 2008 and that is the problem
30% bencmark wins on Barcelona can't change the money situation, the fab situation, nor the technology situation.
I do admire the blind faith or is it laugh at the stupidy in not being able to see the facts!
AMD's stock was at over $35 last year. Now it's at $12 and crashing. It will certainly fall well below $10 after AMD reports a massive loss in a few weeks time. Analysts are predicting a loss as large as $700m.
I think that the "smarter choice" would be to buy Intel stock. When Intel reports excellent profits and that it has regained marketshare in all segments it's stock will be well on it's way up.
Soon AMD will be worth less than SCO. Then they go BK.
$700m is just the beginning. AMD will bleed much more than that throughout the year. They are running out of money. No one wants to loan them more money.
AMD cannot afford to transition it's fabs to 45nm or 32nm. They will be stuck at 65nm until they BK.
P.S. Clovertown is up to 4x faster than AMD's Opteron. http://www.intel.com/products/processor/xeon/competitive_guide.pdf
:o...ahhhhh,
always the same stories,
always the same stories,
always the same stories,
always the same guys,
always the same guys,
always the same guys!!!
Please.
There's no more question!
They are Intels mercenarys which must write about 70-100 words every time they want to post some text.
Of course to spit on AMD.
People who read this blog, don't believe to these people, these bunch of Intels liars!
They always write the same thing, repeatedly!
Dr Blog said...
Real genius, you truly are. Sharikou it's time to speak up, or close this blog, please make your decision soon.
Who do you think you are?
Who are you?
Who are you to tell someone to close his blog?
Migrate, open your own blog!
Intels 'FANBOYS':
'ajhj.ertj,rthar ezzq hg Ašćždhgf kjščk,t eroiht erer;gtern --rerl eri4čš0m ert eržđp-egwer.'
YES, we understand you.
Get lost!
We are priceless, like AMD.
LOL
This thread has had a fair amount of chatter. Let me see if I can add anything.
First of all, trying to find comparable spec 2006 benchmarks is certainly not easy. There are lots of Linux specs for AMD and lots of Windows specs for Intel. I looked but I couldn't find anything that matched in spec Int rate. The closest thing I could find was in spec Fp rate sponsored by Fujitsu Siemens.
RX220 Opteron 280, 2.4Ghz, 36.9
RX300 Xeon 5160, 3.0Ghz, 45.1
36.9 / 2.4 * 3.0 = 46.125
This is a difference of only 2%. There's no fragging; it's just a tie. I suppose that could mean that Barcelona will do better but then I'm sure Nehalem will too. I'm not really mentioning Penryn because I'm not so sure that more cache and faster FSB is going to help that much in this case. These scores are only dual core so there should be enough memory bandwidth. One would guess that Penryn would help most on quad cores. I'm not sure about dual and quad socket because that depends on the chipset.
It is true that AMD is still making 90nm. The truth is that FAB 30 can't make anything else so AMD is pretty much stuck making at least some 90nm on FAB 30 until early 2008 when they get the 300mm wafers rolling and start calling it FAB 38.
Someone claimed that TrackingAMD had "closed". I was just there and it is still there. The authors just aren't posting any new articles until June.
Someone else claimed that AMD's CPU values had been cut "in half" and that Barcelona wouldn't have high yields until "2008" because it's a "new quad-core on a new process, with new tools, in a new factory".
I think for the first point, the poster obviously forgot about the normal price drops with new releases. The prices would have dropped anyway with 6000+. Overall, I'd say the drop was closer to 20% than 50% but there is no doubt that C2D had an effect. And, poor yields for Barcelona into 2008? Well, it is a new quad core; that is true. However, the tooling and factory have been up and running since late 2005; that isn't what I would call new. The process is a bit newer but 65nm seems to be working fine for Brisbane. Yields won't be a problem but the volume isn't going to be up for desktop until Q4.
I've seen a lot of talk about AMD's finances. In just 2 quarters back in 2002, AMD lost 66% of its revenue. Now that was a bite. This is nothing. AMD is going to take the biggest hit in Q2 because those price cuts will kick in but AMD won't have any K10 desktop chips to make up the difference. Q3 should be a lot better because the K10 server chips will be reasonably available. Q4 should be better because there should be more K10's for the desktop. However, the last two quarters won't make up for being down the first two. 2008 should be fine for AMD; I don't believe you will see Nehalem desktop chips until 2009.
BTW, is anyone else puzzled about Intel's plan to keep the northbridge chip and FSB on the single socket Nehalem's? AMD used an IMC even on the socket 754 chips. Maybe Intel can't change over its chipset production that fast. The only other theory I've heard is that if Intel puts the GPU on the die then the Northbridge would be completely eliminated with an IMC and somehow this would cause Intel to lose money.
Oh, and is anyone else also wondering what Intel is going to do to screw up the design and keep competitors away? In other words, it seems like nVidia could just take an AMD chipset and put a simple CSI translator on the HyperTransport port and use essentially the same chipset for both. Intel seems very territorial about its chipset business so you have to wonder.
I see someone is still saying that buying ATI was a mistake. I've ran through AMD's potential strategies over and over and I still haven't come up with anything that would have worked without buying ATI. As far as I can tell, the ATI purchase had nothing to do with arrogance or bad judgement or being overly optimistic. As far as I can tell it was the only option. According to AMD, their biggest sales growth area is commercial and also according to AMD they needed an all in one solution for both continued growth in mobile and in commercial desktop. Giving up the ATI purchase would mean giving up new volume share until at least 2009. That would cost more than the current situation.
And, as far as buying AMD stock goes; I wouldn't. AMD's stock is going to drop a bit when revenue falls in the second quarter. Stock prices should rise after that but they may not if AMD raises money by issuing new stock. This could keep AMD's stock price down until sometime at least mid 2008. This would mean holding the stock for at least a year which doesn't sound like getting rich quick.
Finally, why are so many here almost pathological about giving credit? When AMD bumped the IPC up from K7 to K8 this was pretty good. When Intel got the IPC of Dothan and Yonah up close to K8 this was a lot better than the falling IPC of P4. C2D bounced the IPC by probably a similar 20% and this should prove that they aren't going back to hyperpipelining. I'll have to admit though that doubling the cache bus and SSE performance compared to Yonah was a master stroke.
If anyone is familiar with my position on AMD then you'll know that I've been sceptical about AMD's prospects of catching C2D with K10. I figured AMD would erase about 75% of the IPC deficit. Then I starting hearing that it could actually match. Matching C2D in IPC and SSE would be reasonable performance on AMD's part. Now, the rumors are claiming that K10 will surpass C2D in IPC. If this is true then this would certainly be an A+ effort on AMD's part. Bumping IPC by more than 20% is very difficult to do.
However, I'm also not so sure that Nehalem will match this. In other words, I'm sure Nehalem will catch K10 but I'm sceptical that Nehalem will increase IPC by 20% over C2D. I'd like to give Intel more credit but the current talk is about symetric multi-threading. If Intel only gets the extra speed when running two threads then this isn't quite as good. I mean, are we really going to be able to find ways to keep 8 threads busy on a quad core? And, what about dual socket? So, I'm not so sure that Intel is going to repeat the success it had with C2D with Nehalem. Nevertheless, I'm sure it will catch up with K10.
So, maybe all of the pro wrestling hype is unnecessary. No massacres, no slaughtering, no fragging, no killing, no destroying, no crushing, and no bankruptcy. Just stiff competition.
kudos to you scientia... that was a he*l of a read back there =D.
I guess one more thought. Let's compare AMD with Chrysler.
Chrysler's current value $4.5 Billion
2006 Earnings - Lost $1.5 Billion
Announced Layoffs - 13,000
Unfunded Pension Liability - $22 Billion
Loss of Value since 1995 - 80%
It appears that being bought by Daimler Benz hurt more than helped. Yet, at least two different groups want to buy Chrysler. Does anyone here who is talking about AMD being in such bad shape seriously see any similarity between AMD and Chrysler?
Nice post scientia.
I do have two points I would like you qualify with additional info. a link would be fine.
K10 Server availability - Ruiz has said in an interview that K10 would not ship in significant numbers in 2007. Is so, why do you think it would provide a significant boost to income?
K10 on desktop. Hasn't AMD said that desktop for K10 is 2008?
leecooper...
don't take this personally... but is it just you, or all of AMD fanboys are so desperate to save AMD that they even shoot one of their own?
Dr. Blog is actually more AMD than Intel.. he was just telling sharikou to post something meaningful, and actually facilitate the comments. if he doesn't then there is no point in having a blog where everyone just flame each other.
here a thought for you. how come everytime Intelers make a statement, they back it up with real data, and they are called mindless fanboys.
however, when AMDers say something without backing it up, and those words are taken as gospel of truth?
bubba
I do have two points I would like you qualify with additional info. a link would be fine.
No problem.
K10 server availability - Ruiz has said in an interview that K10 would not ship in significant numbers in 2007.
Okay, I know the interview you are talking about. AMD’s New Processor Micro-Architecture in 2008 – Chief Executive.
AMD’s Hector Ruiz Predicts New AMD Micro-Architecture in 2008
and this is the part you are referring to:
“We’re going to introduce a really new architecture that will work well with our partners for the best performance. We’re going to start sampling it at the end of 2007 and roll it out in 2008,” said Hector Ruiz
And, I can see how you could think that he was talking about Barcelona and that would indeed push K10 way back. However, he wasn't talking about Barcelona; he was referring to Direct Connect Architecture 2.0 which is due in 2008.
K10 on desktop. Hasn't AMD said that desktop for K10 is 2008?
No, actually you can see this in the same article. Remember, this interview was in 2006 so "next year" would be this year, 2007:
The products to debut next year will include a quad-core design for servers, workstations and high-end desktops, and a dual-core design intended for mainstream desktop markets. These next generation processors will be built using AMD’s 65nm silicon-on-insulator process, and include a broad range of functionality and micro-architectural improvements, including a new ability to dynamically alter the frequency of each core on the chip to match application workloads and thereby reduce overall power consumption.
The oldest roadmap for 2007 is this vr-zone roadmap.
You can see that the clock frequencies have been overstated for both AMD and Intel. In this chart Antares is Kuma and Altair is Agena.
We have the correct names in this Dailytech roadmap.
Agena is AMD's K10 equivalent of Kentsfield. Kuma is the K10 version of X2. Notice the roadmap shows both in 2007.
This HKEPC roadmap only shows Barcelona but you can see that it says mass production in June. You fall back another quarter for mass production for Agena and Kuma and this would probably be September.
Okay, this January Dailytech article says:
Agena launch at 2.4-2.6Ghz in Q3
Presumably Kuma would then be Q4.
In the Forte Interview Randy Allen, AMD VP of Servers and Workstations said:
AMD will begin shopping the Barcelona chip around to customers in the April-June time frame
This fits with the HKEPC roadmap indicating launch in May and mass production in June. Again, the Agena and Kuma chips should trail this by about 3 months.
No debt is secured!!
You are ignorant beyond comprehension.
So I guess you've never heard of junk bonds? So secured... they decided to call them 'junk'!
Bubba, did you get your name before or after you got out of jail?
The only other theory I've heard is that if Intel puts the GPU on the die then the Northbridge would be completely eliminated with an IMC and somehow this would cause Intel to lose money.
Great point! And this is the kind of stuff that Intelers don't even seem to understand themselves. Just like their reason not to integrate GPU on CPU in fear of loosing some more chip sales.
That is why I was telling 'ho ho' a little while back that Intel has a lot to loose and AMD only to gain.
---
Actually I don't understand Intelers? Were all just a bunch of geeks following technologies. But there is a main difference between AMD and Intel fans!
AMD fans are usually people who don't hate Intel products, but just Intel's way of doing business. We are confident and usually prefer to root for the underdog. Because we know that it is the little guy who adds balance in the ecosystem.
On the other hand, you have Intelers, who actually 'hate' AMD. Constantly you hear them wish AMD would go bankrupt and disappear. Outright refusal to accept the greatness in competition. Please, Intel fans, respond with YOUR reason why you 'TRULY' want AMD to fail and/or disappear cause I really don't understand you?
how come everytime Intelers make a statement, they back it up with real data, and they are called mindless fanboys.
Right, so show me data/proof for comments as such:
"Intel massacres AMD"
"AMD architecture sucks"
"AMD only copies Intel"
"AMD is going bankrupt"
And no, your phony, Intel compiled, single threaded, Intel sponsored benchmarks don't count!!
however, when AMDers say something without backing it up, and those words are taken as gospel of truth?
We're smarter :)
jeach!
"Intel massacres AMD"
link: http://anandtech.com/cpuchipsets/intel/showdoc.aspx
?i=2903&p=5
a processor running 1.8Ghz is able to generally outperform X2 4600 @ 2.4Ghz.
"AMD architecture sucks"
that i have no comment. first of all i've never made such comment. if you're directing your fire towards AMD fanbois, please do not quote me.
"AMD only copies Intel"
well the only thing AMD copied Intel was its quality of benchmarking. AMD is becoming less genuine and informative about its benchmarks. Just as the link Sharikou posted, the OP of the article actually questioned AMD's move to make benchmark comparisons more ambiguous.
"AMD is going bankrupt"
i did not say that, though there are a lot of wall street-ers speculating.
http://www.247wallst.com/2007/04/an_adva
nced_mic.html
personally i would hate to see AMD going bankrupt. we might see another Netburst era if AMD went bankrupt.
however, i would say AMD's current financial situation was because of AMD's poor decision making.
1. buy ATi using cash instead of all-stock
2. too much ego in releasing all level of R600 at the same time. if the top model R600 is ready, why don't they just release it to the public?
3. not showing Barcelona's benchmarks. to be honest, if AMD already benchmarked Barcelona to get the 40% performance, why don't they just release ES to third party site to save AMD from sinking deeper?
AMD fans are usually people who don't hate Intel products, but just Intel's way of doing business. We are confident and usually prefer to root for the underdog. Because we know that it is the little guy who adds balance in the ecosystem.
oh really? the person who owns this blog, has been forecasting Intel's bankruptcy all the way back in 2005. "Intel BK in 2Q 06". "Intel BK in 2Q 07". "Intel BK in 2Q 08".
to be honest, back in K8 vs. Netburst era, how many AMDers actually believed in Intel's bankruptcy? i would say tons.
not a lot of people believing in Intel's BK at the moment is because it would simply be too stupid to forecast such thing when C2D generally trump AMD's offering.
saying that you hate Intel's way of business doesn't make you less than a fanboy. to be honest, i find your utter ignorance, yet trying to disguise yourself as a neutral person quite amusing.
We're smarter :)
no. we're all the same, Intelers and AMDers. but we are all smarter than those who are blinded by their own bias and ignorance, yet refuse to admit it.
yomamafor2
a processor running 1.8Ghz is able to generally outperform X2 4600 @ 2.4Ghz.
I'm sorry but this statement is incorrect. I've looked over the Anandtech tests and there is a problem. Although you can use the tests to relate C2D to C2D performance and K8 to K8 performance you cannot relate them to each other. I would love to be able to look over the scores and give you a good estimate of how these two architectures match up, but unfortunately, there isn't enough data. I can certainly understand why you would mistakenly think that there is enough information and you would not be alone. Most people would just assume that the scoring numbers are completely comparable. However, that is not the case.
I don't know if you are familiar with black box investigation. Not in the sense of a flight recorder but in the sense of an unknown process. Since I don't have the code for the benchmark it is a black box; I can't see what is inside it. However, if I have enough information I can infer the process and this then would allow me to connect the scores for K8 and C2D. But, there isn't enough data. To rule out cache sensitivity I need an overlapping cache range but there isn't one. The K8's only have 512K while the C2D's have 2MB's. This leaves too big of a gap. When K10 is released it has a 2MB L3 so this should overlap enough to determine cache sensitivity.
However, there is also the question of the compiling method for the code. I can assure you that small differences in compilation can change code speed by 20% quite easily. To rule this out we need something that allows us to eliminate compilation as a variable. Since SysMark is strictly a Windows product and we can't change how it is compiled the only thing we can do is profile a number of benchmarks compiled with something else like Linux benchmarks compiled with GCC. Then you see if some benchmarks stand out as statistically different. It isn't simple. The bottom line is that there is not enough information in the Anandtech tests to tell how good the testing actually is.
1. buy ATi using cash instead of all-stock
I'm curious. What makes you think that AMD didn't try a stock deal first? My guess is that the ATI board of directors insisted on cash.
2. too much ego in releasing all level of R600 at the same time. if the top model R600 is ready, why don't they just release it to the public?
This actually has nothing to do with ego. In some respects the introduction of a new product works the same as advertizing. AMD said that its vendors requested the all level release for the biggest effect rather than a series of small releases with hardly any notice. Keep in mind that AMD is not doing this out of ego; it is counting on the R600 release to partially offset the decrease in revenue in Q2.
3. not showing Barcelona's benchmarks. to be honest, if AMD already benchmarked Barcelona to get the 40% performance, why don't they just release ES to third party site to save AMD from sinking deeper?
Well, this is a double edged sword. If you preview a product that isn't ready to be released then you run the risk of killing your current sales. This is what happened to Osborne and they went bankrupt because of it. Secondly, AMD's vendors would get very irritated if previewing an unavailable product caused the value of their current K8 merchandise to drop. AMD's vendors want K10 information suppressed until they have K10 in stock.
a processor running 1.8Ghz is able to generally outperform X2 4600 @ 2.4Ghz.
Generally C2D outperforms K8 by ~20% clock for clock. Thus a Core 2 E6300 is roughly equal with an Athlon X2 4200+.
No, I think I mean 20%. Anyone can point to one page within a review. For instance: http://www.hexus.net/content/item.php?item=7913&page=5
Nice try though.
As I explained before I haven't seen enough information to actually come up with a firm number.
I've seen averages as low as 11% and as high as 22%. I'm pretty confident that C2D has at least 14% more Integer IPC. You cannot actually pinpoint one benchmark more accurately than 3%. To say that the actual number is 20% and not 18% or 22% is impossible with the current data.
And, are we talking 20% with 32 bit or 64 bit code; are we talking Windows or Linux? I'm hoping that we can get some good data soon with Barcelona running under Vista and start getting enough information that we can actually say.
Ideally though the benchmarks would be compiled on both Vista and Linux with Portland Group which does AMD and Intel. If we could compare PG Linux with both Suse and Intel Linux then we'd have some good data.
Scientia,
No I was refering to the interview here:
http://www.crn.com/sections/breakingnews/dailyarchives.jhtml?articleId=197007770
In February of this year he stated he's not going to produce many Barcelona cpus.
Given this, I don't see AMD having positive cash flow for a year.
scienta nice post...
What is your take on why no benchmarks or private demos and leaks on Barcelona benchmarks. AMD's history has been to tell early, why the change and for what reason.
Two extreme possibilities depending on your views of things
1) Benchmarks are a toss up +/- a few %
2) Barcelona really delivers double digit over Penrym...
Either way AMD simply won't have enough volume nor pricing flexibility in the server space to cover the ramp to 65nm and fund 45nm pull-in. Something has to go to fix margins and staunch cash burn.
I think you underestimate INTEL's ability to push once they have set a direction. In the netbust days once they decided Meghz was king they ended up doing a decent job of pushing it, only forgeting the minor detail about power and power efficiency.
Now that they have swung hte focus the other way the resources they bring to bear I'm sure Penrym, Silverthorne and Nehalem will come out on schedule hell or high water and hit their performance goals. Netbust failed due to miss directed management and marketing direction.
I'm skeptical Brisbane ramp and health is indicates process readiness for Barcelona. I only look at the Pentium M and Pentium IV ( Netbust ) circus on 90nm at INTEL. You discount the complexities of process and design interaction for new vs shrink designs. The tick tock with 9 to 12 months lag is the right strategy as at these dimensions the design and process interactions are just too subtle to cover on a single product.
My belief is if INTEL really did get some cost improvement from their layoffs this past year then they have more wiggle room to drop prices and I expect them to severly punish AMD cash flow to insure their ability to continue to design and process cadence is permanently ended. 45nm and 32nm node investment is just astronomical unlike the competitive landscape at 130 and 90nm where equipment vendors could offer some signficant help, by the tool and get the process and plug and play. At 45nm and 32nm that simply isn't the case anymore. Without investment early and deep AMD simply doesn't have cash to burn.
IBM is no longer a player a big player so they help but it won't be enough.
My last comment to you fanbois.. how many of you are studend and fans of the CPU wars vs. actually have a clue about what it takes to bring one to market either from design/test/validation to manufacturing/ramp on the technology side. I love watching those fanbois ( those with no clue ) lap the interviews of Patty or hector, or take as gospel from dailytech, or other trademags from silly journalists with no clue. Ever wonder why a journalists only reports vs have a better more lucrative job watching the industry?
bubba
Oh, that interview. Okay.
RUIZ: This is an incredibly important product transition. We don't expect the ramp [this year] to be dramatic because it's a new core, new micro architecture and platform. The biggest impact it will have is that we'll see a large number of customers and partners align themselves behind the technology. We expect that ramp to follow along the same lines as when Opteron began to get adopted. So I expect it to follow the same patterns. Over 2007, it will have a significant impact on what I call design wins. People are committed to the architecture and product, and [it will be] a very significant part of revenue and earning in 2008.
Yes, he is saying that the ramp won't be dramatic. In other words it will be less than half of production by end of 2007, 30% at most.
In terms of revenue this should be fairly clear. Let's say you started at mid year and ramped production up to 50% by year's end. So, your average would be 25%. However, you've only produced that average for half the year so your actual average for the whole year is only 12.5%. This is why even with aggressive ramping the volume is still too low to have much effect on yearly revenue. Therefore, the real revenue effect will be in 2008.
Right, but you said that Barcelona would be significant this year. Q3 should be a lot better because the K10 server chips will be reasonably available. Q4 should be better because there should be more K10's for the desktop.
So which is it? You're talking out of both sides of your mouth.
they used 32-bit Windows XP which is obsoleted OS as it stands today.
Which is used by 99% of the worlds pcs and will be for a long time.
scientia said: "Yes, he is saying that the ramp won't be dramatic. In other words it will be less than half of production by end of 2007, 30% at most."
That's funny, i tried looking at a dictionary and "less than half but 30% at most" isn't one of the meaning of "dramatic".
Truth is nobody knows exactly what Ruiz meant and your wishful guess is as good as anyone's. Bottomline is, AMD will be sacrificing both marketshare gains and revenue by chosing to load Barcelona at 65nm. It isn't cost effective. AMD itself even confirmed themselves that all they plan for Barcelona for 2007 was to get design wins. I really don't understand how you can associate "design win" volume to 30% production volume. You're contradicting AMD's CEO. Unbelievable.
AMD is so desperate to sell it's products that even the top 6000+ is selling for just over $300. No one wants to buy their ancient high power consuming 90nm processors when they can get cool energy efficient Core 2 Duos (6000+ uses 125W - All C2D uses only 65W).
AMD's ASP is going way down. One thing is for sure: AMD is going to post a larger loss than was earlier predicted.
All this means that AMD is finished.
AMD BK Q2'08.
Bubba said...
Which (Windows XP) is used by 99% of the worlds pcs and will be for a long time.
No. If you can back your statement by some source that would be better. Oh wait... all Windows OS's don't even account for 94% of worlds PC's.
Giant said...
AMD is so desperate to sell it's products that even the top 6000+ is selling for just over $300.
Wrong. Everyone wants to buy that technology. Thats why AMD's market share grows at 2.4% per Q. You can believe what you want you know, but numbers say a different story. Wanna bet AMD will take some more of intel's business in 2007Q1?
Giant said...
...consuming 90nm processors when they can get cool energy efficient Core 2 Duos (6000+ uses 125W - All C2D uses only 65W).
Firstable, it's a 89W chip. Secondary AMD's 90nm technology is more energy efficient than intel's 65nm: DDR2 memory controller takes ~20W of those 89W, thats 69W already, then lets take in to the account the fact that intel uses "average" TDP while AMD uses maximum. Thats another 10W. Can you argue that intel doesn't use "average" TDP or that they don't include MC wattage in their chip's TDP? If you do, bring something to back you up - i got plenty of sources that can confirm my claims.
AMD lost marketshare in the most profitable segment last quarter, servers. AMD is losing marketshare in all segments this quarter. As The Inquirer reported, AMD's CPU revenue will miss expectations by 300 -> 400 million. We could argue endlessly about this, but we shall just wait until AMD and Intel report financial results for Q1'07. That should reveal all.
Saying AMD's 90nm is more efficient than Intel 65nm is plain wrong. The 6000+ is 125W. http://www.amdcompare.com/us-en/desktop/Default.aspx
Go to that site, choose AMD X2 processor, and 6000+ and it will give you the TDP.
To further ilustrate this point, let's take a look at that review you linked to earlier. Here's the graph of power efficiency: http://techreport.com/reviews//2007q1/athlon64-x2-6000/cine-power-peak.gif
That's nearly 80W more. That is the power drawn from the entire system, only the motherboard and CPU are different for the systems.
Dr Blog,
You actually sold people K5 and K6 cpus?
Hvae a nice time in hell.
Superior Architecture (AMD) is King
http://www.anandtech.com/tradeshows/showdoc.aspx?i=2956&p=5
To summarize, Intel quad = hack job, bottleneck, doesn't scale resulting in push for higher 1600MHz memory bus and copying of AMD integrating memory controller.
Intel has gone from spinning the advantages of "higher clock" of yesterday to now "die shrink" since they're not regarded as the architecture king. King of architecture (AMD) teamed with king of process technology (IBM) spells trouble for Intel.
Firstable, it's a 89W chip. Secondary AMD's 90nm technology is more energy efficient than intel's 65nm: DDR2 memory controller takes ~20W of those 89W, thats 69W already, then lets take in to the account the fact that intel uses "average" TDP while AMD uses maximum. Thats another 10W. Can you argue that intel doesn't use "average" TDP or that they don't include MC wattage in their chip's TDP? If you do, bring something to back you up - i got plenty of sources that can confirm my claims
It's really funny that when talk about the power efficiency, the AMD fanbois would again and again saying that Intel TDP is average while AMD's TDP is Max ... blablabla .. do you have any idea what does the TDP means? It is called thermal design power, not the power consumed by the chip. For simplifying the motherboard design, the TDP is quoted for a range of CPU frequency, do you seriously thing E6300 consume the same level of power as E6400 even they are rated the same TDP?
second trick that AMD fanbois would use would claim the IMC and the northbridge differences. Yes, the Northbridge would consume some energy, but in total the current Intel platform is still more power efficient. You do not need to claim the extra energy used in Northbridge or a particular type of ram in the server case; you just need to measure the power at the plug (power source)! IIRC, Tech report (not sure if the name is correct) has shown benchmarks with reference to the total power used, and Intel's platform is more efficient, at least for the current version.
Barcelona will beat intel's top dog.We will see incredible performance of barcelona over intels quad_core.
bubba
"Right, but you said that Barcelona would be significant this year. Q3 should be a lot better because the K10 server chips will be reasonably available. Q4 should be better because there should be more K10's for the desktop.
So which is it? You're talking out of both sides of your mouth."
No. I haven't changed what I said; you are confusing two different things: quarterly profits and yearly profits.
I expect AMD's quarterly profits to be down Q1 and Q2. Then with the K10 release I expect Q3 and Q4 to be better relative to the first two quarters.
However, because K10 volume for the entire year will still be low, Q3 and Q4 won't be enough to give AMD a good yearly profit.
giant
"AMD lost marketshare in the most profitable segment last quarter, servers."
True.
"AMD is losing marketshare in all segments this quarter."
I'm not so sure about this. Let's say that AMD is likely to lose revenue share but may be able to hold onto volume share.
"As The Inquirer reported, AMD's CPU revenue will miss expectations by 300 -> 400 million. We could argue endlessly about this, but we shall just wait until AMD and Intel report financial results for Q1'07. That should reveal all. "
Oh, I have no doubt that AMD's first two quarters will be down. I expect Q2 to be worse than Q1. In fact, I'm guessing Q2 is going to be down sharply over Q1. But, Q3 should be better. Do, you disagree?
roborat
"Bottomline is, AMD will be sacrificing both marketshare gains and revenue by chosing to load Barcelona at 65nm. It isn't cost effective."
You've completely lost me here. What are you talking about? Are you trying to say that Barcelona can't be manufactured on 65nm? Producing Barcelona as soon as possible seems like common sense. Producing it on 65nm rather than 90nm also seems like common sense. What exactly is your objection??
" AMD itself even confirmed themselves that all they plan for Barcelona for 2007 was to get design wins."
Yes, several HPC contracts for example. 30% is what I estimate would be the maximum reasonable ramp. It could be lower. Again, the volume for 2007 will not be a large fraction of production. If AMD can reach 30% volume by end of 2007 this would only be about 7.5% of the year's volume. I think anyone with common sense would agree that if 92.3% of your production is K8 then you can't expect too much from K10.
Okay, I see a lot of back and forth about whether AMD's or Intel chips are more efficient. The truth is that it doesn't matter. Nor does it matter whether the chips are 90nm or 65nm. The offerings in terms of TDP from Intel and AMD are very similar in ranges of 30W, 50W, 90W, and 120W. The only question is how much performance you can get at that TDP. Intel does currently have an advantage in SSE performance at similar power draw on all platforms and increased Integer performance on desktop systems.
scientia from amdzone:
I apologize for not present a fully researched argument. Let me do a little more research to give you a response =)
copying of AMD integrating memory controller.
Another dumb kid that hasn't studied history. Intel has an IMC long before AMD. Look it up.
pointer said...
It's really funny that when talk about the power efficiency, the AMD fanbois would again and again saying that Intel TDP is average while AMD's TDP is Max ... blablabla .. do you have any idea what does the TDP means?
http://www.silentpcreview.com/article169-page3.html
...And the most telling quote of all, contained in both documents:
Analysis indicates that real applications are unlikely to cause the processor to consume maximum power dissipation for sustained periods of time. Intel recommends that complete thermal solution designs target the Thermal Design Power (TDP) indicated in Table 26 instead of the maximum processor power consumption. The Thermal Monitor feature is intended to help protect the processor in the unlikely event that an application exceeds the TDP recommendation for a sustained period of time.
What this means is that Intels TDP is actually lower than the maximum power dissipation of the processor (and as youll see later, it can be significantly lower). This is in stark contrast to AMDs TDP numbers, which are higher than the respective processors maximum power dissipation.
As I've said AMD uses maximum possible TDP value while intel play games and shrugs 10-15%. Unlike your post i am able to back my point by professional opinion which includes real life testing, thus i prove my point and disprove yours. You are corrected Mr. Pointer.
Pointer said...
do you seriously thing E6300 consume the same level of power as E6400 even they are rated the same TDP?
Same applies to AMD CPU's as well thus your point is irrelevant.
Pointer said...
Tech report (not sure if the name is correct) has shown benchmarks with reference to the total power used, and Intel's platform is more efficient, at least for the current version.
You mean This report?
Well, it shows that system based on Athlon X2 3800+ EE consumes less power under load and in idle than any Core2Duo based system (as well as AMD's)
You mean This report?
Well, it shows that system based on Athlon X2 3800+ EE consumes less power under load and in idle than any Core2Duo based system (as well as AMD's)
Under load the system power is identical to the Core 2 E6300. At an idle state AMD has an advantage of THREE watts.
Lets have a little poll, who would rather have the E6300 and it's increased performance over a 3800+? Or how many want the 3800+ and it's three watt advantage at an idle state?
I'm not so sure about this. Let's say that AMD is likely to lose revenue share but may be able to hold onto volume share.
We shall see in just a short while, won't we?
Oh, I have no doubt that AMD's first two quarters will be down. I expect Q2 to be worse than Q1. In fact, I'm guessing Q2 is going to be down sharply over Q1. But, Q3 should be better. Do, you disagree?
I agree with this. If AMD can ship Barcelona in Q3 with sufficient volume to ensure it's customers can secure a steady supply of the processors AMD will have a better Q3 than Q2. Q4 should be even better for AMD when it releases K10 desktop CPUs.
Randy Allen said...
Under load the system power is identical to the Core 2 E6300. At an idle state AMD has an advantage of THREE watts.
Lets have a little poll, who would rather have the E6300 and it's increased performance over a 3800+?
In general, in 2006 more than 25% of people preferred AMD technology over intel's , thats a 35% revenue increase over 2005 acording to AMD's official letter to their shareholders, and we know you can't get any more serious than that.
AMD-2006ACO-Letter
Being able to back me up with some data I'm confidently saying that there are plenty of people who would/will choose AMD over intel because AMD has answer to what TDP or/and performance intel CPU's has to offer. Which was the point of my original post.
In general, in 2006 more than 25% of people preferred AMD technology over intel's
So in other words, three out of four people preferred Intel's technology? That number is rising now. By the end of the year it will be at least 80%.
Randy Allen said...
So in other words, three out of four people preferred Intel's technology? That number is rising now. By the end of the year it will be at least 80%.
Wrong. Intel market share decreases while AMD's is on the rise.
http://www.tgdaily.com/content/view/30834/118/
In plain numbers, AMD finished the fourth quarter of 2006 with a market share of 25.3%, up more than 8% over Q4 2006, in which the company held 21.4%. Intel, meanwhile, dropped about 3.4% points from 77.0% to 74.4%.
As you see in the last quarter AMD gained significant ground on intel. We do not have yet the numbers for 2007Q1 but they will be out in a few weeks, in which, on my professional account AMD should gain somewhere between 0.8% to 1.4% more in MS.
I believe that I've just proven your claim of: 'That number is rising now. By the end of the year it will be at least 80%.'- to be false.
AMD lost marketshare in servers last quarter. This quarter it will be in all segments.
AMD's shares were down today on news that it would miss its previous revenue guidance of $1.6 to $1.7 billion USD for Q1 2007. Despite admitting the shortfall, AMD did not disclose exactly how much it expects to come up short for the quarter.
http://www.dailytech.com/article.aspx?newsid=6351
The cuts, which will slash between 26% and 42% of the price of some chips in the Athlon and Sempron product lines, are an effort to "bolster flagging demand" and "pre-empt Intel's price cuts scheduled for April 22," according to Dhanda.
Lower prices for its microprocessors are hurting AMD. The company is calculated to chalk up a net loss for 2007, and estimates have been increasing in recent weeks.
Analysts forecasted by Thomson Financial forecast the semiconductor maker will lose 45 cents a share, up from a net loss of 12 cents previously estimated in late January.
How can AMD be gaining market share when it is cutting prices on processors to "bolster flagging demand" for it's processors? In addition AMD continues to build up inventory of 90nm products. AMD is forced to sell these processors at very low prices, leading to an even lower ASP and a larger loss.
AMD has no money to go 45nm or below, no one wants to loan AMD more money because they'll never see it again!
AMD BK Q2'08.
link to second article:
http://news.morningstar.com/news/ViewNews.asp?article=/DJ/200703211732DOWJONESDJONLINE001105_univ.xml
From the article you have linked Andy:
AMD warned March 5 its "unlikely to meet" its first-quarter sales target between $1.6 billion and $1.7 billion.
From what you've said Andy it is pretty clear that you confuse Market Share 'MS' with sales. I don't remember i ever said that AMD will rotate good money in 2007Q1 and in fact my prediction of MS 0.8% to 1.4% gain in Q1 is still stands. We can talk more about this when results are out.
You have also mentioned that AMD's server MS went down in 2006Q4, well how come you forgot that their desktop MS went up big time in the same Q?
When AMD needs to cut processor prices so much that an X2 3600+ costs $65 to "bolster flagging demand" for it's products it's pretty clear they're losing market share. (See link posted above for the full story)
AMD also has a lot of inventory, which led to further price cuts.
AMD will cut the prices of entry-level 90nm processors on 9 April by between 30 and 35 per cent, a US investment bank has claimed.
San Francisco-based Thomas Weisel Partners (TWP) recently told investors AMD's inventory levels of these chips remains high, suggesting to TWP analysts a further round of price cuts is in the offing, DigiTimes reports.
http://www.reghardware.co.uk/2007/03/23/amd_price_cuts_coming/
Now tell me, why would AMD need to cut prices to clear inventory and "bolster flagging demand" for it's processors if they were gaining market share as you suggest?
This comment has been removed by the author.
deleted the previous post because word missing
Azmount Aryl said...
You mean This report?
Well, it shows that system based on Athlon X2 3800+ EE consumes less power under load and in idle than any Core2Duo based system (as well as AMD's)
wow, what i can say is that u do learn a lot from the blog owner and this is another show case of choose and pick at its best.
1) you compared a lower performance CPU to a higher performance CPU and claims its power advantage
2) you used an EE version (35W TDP) compare to Intel earlier stepping of the E6300 (65W TDP) (the review link dated august 2006) : (a) even so, it only shows 3W advantage (not yet consider the performance) at idle only and isn;t this contradict your view of TDP? (b) new Intel stepping Allendales has much improved idle power. http://forumz.tomshardware.com/hardware/E6300-B2-L2-Stepping-ftopict227656.html
http://www.pricewatch.com/cpu/
What's going to happen to all the Intel NetBurst CPUs (look at url above)? The prices have been lowered to almost nothing but they still don't move. If Intel gave them away for free would anybody take them or reject them still since they're underperforming, consume too much electricity and generate out too much heat? At least Intel's latest generation Core architecture has caught up to AMD's two year old technology.
Bubba said...
Dr Blog, You actually sold people K5 and K6 cpus?
Yes, and that's where it really started to make sense. My customers came back and told me the K6-2 500 computer I sold them was the first computer they ever purchased that worked, never blue screening, or locking up on them. I retired my K6-2 500 workstation a year and a half ago, and at that time you could still buy the K6-2 500 for 35 dollars at Newegg. What a deal, $35 for something that actually worked! I don't remember a single BSOD in over five years of service.
The Doctortrabal
Bubba said...
Dr Blog, You actually sold people K5 and K6 cpus?
Yes, and that's where it really started to make sense. My customers came back and told me the K6-2 500 computer I sold them was the first computer they ever purchased that worked, never blue screening, or locking up on them. I retired my K6-2 500 workstation a year and a half ago, and at that time you could still buy the K6-2 500 for 35 dollars at Newegg. What a deal, $35 for something that actually worked! I don't remember a single BSOD in over five years of service.
The Doctor
Bubba said...
Dr Blog, You actually sold people K5 and K6 cpus?
Yes, and that's where it really started to make sense. My customers came back and told me the K6-2 500 computer I sold them was the first computer they ever purchased that worked, never blue screening, or locking up on them. I retired my K6-2 500 workstation a year and a half ago, and at that time you could still buy the K6-2 500 for 35 dollars at Newegg. What a deal, $35 for something that actually worked! I don't remember a single BSOD in over five years of service.
The Doctor
Would you delete the repetition Sharky, had problems with Google.
Post a Comment
<< Home