AMD's Asset Smart Explained
Although Hector Ruiz mentioned the term Asset Smart at the Q1 2008 Earnings Report he avoided explaining what it meant. There has been a lot of speculation in this vacuum but all of it that I have seen has been wrong. Most theories seem to focus on either the idea of selling all or part of a FAB to raise cash or on the idea of making graphic products at Dresden. It is actually something quite different.
Asset Smart deals with manufacturing. Starting from 1998 you have:
1998 - SPC
2000 - APC
2003 - APM
2005 - LEAN
2008 - SMART
The term Asset Light referred to AMD's changes within the Lean project (although some of these had actually started back with the APM project). The next project is SMART. Asset Smart simply refers to similar changes within the Smart project which begins in Q3 2008. Mostly, these projects deal with the process of making chips and trying to reduce manufacturing costs. The groundwork for this next phase was actually begun by AMD back in June of 2007 during SEMICON West when it hosted meetings of the Next Generation Factory (NGF) group. This continued during ISMI in October 2007 and later at SEMICON Japan. For example Semiconductor International covers the Austin meeting here:
In an effort to place a more intense focus on 300 mm fab productivity improvements, Advanced Micro Devices Inc. (AMD, Sunnyvale, Calif.) is hosting ~75 people at its Austin campus today for the second in a series of Next Generation Factory (NGF) meetings.
The day-long Austin meeting will include particpants from six integrated device manufacturers — AMD, Freescale Semiconductor, IBM, Qimonda, Renesas Technologies and Spansion — and 16 semiconductor equipment vendors, said Gerald Goff, senior member of AMD’s technical staff. Six academic experts in fab productivity were also expected to attend, Goff said prior to the meeting.
“The suppliers and IDMs used to work more directly together on productivity issues,” Goff said, adding that the NGF group is intended to complement efforts within SEMI and the International Sematech Manufacturing Initiative (ISMI). “ISMI is not moving fast enough. We have to push this 300 mm fab efficiency issue harder as an industry,” Goff said, adding that ISMI project manager Denis Fandel will be among the attendees at today’s event. “In no way do we want the takeaway from this to be that we are against ISMI,” he said, adding, however, that the growing industry emphasis on 450 mm wafers is “concerning to us.”
Goff said that because AMD was “relatively late getting to 300 mm wafers,” it may have more interest in productivity gains at the 300 mm wafer size than its competitor, Intel Corp. (Santa Clara, Calif.), which seeks momentum behind the transition to 450 mm wafers in ~2012.
You can get more information about this directly from AMD by looking at Doug Grose's Keynote Presentation at the 4th ISMI Symposium.
Back in 1998 AMD was at 2.5 days per mask layer. After SPC, APC, and APM, FAB 30 was down to 1.5 days per mask layer. With Lean, by the time FAB 30 shut down in mid 2007 it was down to just 1 day per mask layer. What AMD wants to do is reduce cost by reducing cycle time just as it has been doing for the past 10 years. As a result of Lean, wafer starts per week have jumped 31%, while labor productivity (monthly activities per operator) has climbed 72%. Monthly wafer costs have dropped 26%, and the already mentioned cycle time per mask layer has been trimmed 23%. However, FAB 36 is still at 1.4 days per mask layer. AMD is hoping to reduce this down to 0.7 days per mask layer (a 50% reduction) by shifting to small lot manufacturing.
The basic strategy involves replacing batch tooling with single wafer tooling and reducing batch size. AMD wants to drop below the current batch size of 25 wafers. AMD figures that this will dramatically reduce Queue Time between process steps as well as reduce the actual raw process time. Overall AMD figures a 76% reduction in cycle time is possible so a 50% reduction should be reasonable. Today, running off a batch of 25 wafers is perhaps 6,000 dies. Reducing batch size would allow AMD to catch problems sooner and allow much easier manufacturing of smaller volume chips like server chips. Faster cycle time means more chips with the same tooling. It also means a smaller inventory because orders can be filled faster and smaller batches mean that AMD can make its supply chain leaner. All of these things reduce cost and this is exactly how AMD plans to get its financial house in order.
35 comments:
My apologies to anyone who was hoping that Asset Smart meant selling off pieces of the company. Sorry.
Oh my, I was seriously thinking that they are going to sell all fabs to make money. I guess.... I was wrong.
Scientia,
this type of competing based on costs only won't get AMD far.
They need to compete on performance and efficiency first of all. Even if AMD manages to cut the costs in half, they are still pushed in the lower end, where profits are lower and lower, by Intel's agressive pricing policy.
Cutting costs alone won't solve AMD's problem in the long run, it can stop a litle of the bleeding in the short run and that's all.
sparks
I decided not to bother with a reply to your comment (which you never intended to be in any actual spirit of discussion).
Roborat's blog is a great home for trolls which he fosters with his reliance on anonymous posts. The comments there about this article were somewhere between irrelevant and nonsense:
1.) Scientia must be wrong because no one knows what Asset Smart means.
This is strange indeed because the same crowd had no trouble accepting a defintion right away when they thought it meant selling off AMD's FABs. Yet, that idea only came from a single source, IDG, in 2007 and was refuted by AMD just one month later. The latest retread of this idea again has only one source, Theo Valich, and no supporting evidence. The documentation that I linked to uses the SMART project name.
2.) Scientia must be wrong because he doesn't work in chip fabrication.
Again, these are AMD's documents and conclusions, not mine. Anyone who can actually read can look through the ISMI presentation and see what AMD's conclusions are. AMD concludes that shorter cycle time will save money. I would have no way of verifying this.
3.) These ideas won't work because they are not unique to AMD or because they don't apply to every company.
I'm sure each company adopts similar industry practice ideas if they benefit them or ignores them if they don't.
Intel claims that double patterning with a dry process is cheaper for them than immersion at 45nm. AMD claims that immersion is cheaper for them. I doubt either one is lying. Similarly, I doubt small batch manufacturing would help Intel. Presumably, Intel has seen a better benefit by moving to 450mm wafer size but that doesn't mean that AMD would benefit in the same way.
4.) The numbers are useless because they do not include industry averages.
These numbers were presented to several other companies who would obviously know if they were good or bad whether I did or not. It is reasonable to assume that if AMD had nothing worthwhile to say then the NGF meetings would have ended after the first one at SEMATECH West.
5.) Switching to smaller batch or single wafer tooling isn't free.
I would think this would be obvious. I have no idea how far along AMD might be in this plan or even if the tools are available. My assumption was that they needed the group in order to get enough volume demand to have some of these produced. Again, I have no idea how feasible this project is; just that this does seem to be what AMD has in mind and has been working on for nearly a year.
Other statements are good indications that posters there don't bother to read very well such as complaints that FAB 30's cycle time was lower that FAB 36's. Actually, I specifically mentioned this. FAB 30 had a cycle time of 1.0 at shutdown while FAB 36 is 1.4. Again, it is AMD's goal to reduce this to 0.7. Other comments are odd such as NGF is unnecessary because of Prime300. This does seem to ignore the fact that AMD was the keynote address at ISMI which presumably would not have been the case if all they had done was rephrase Prime300 goals. Semiconductor International also participated in the NGF meetings. A lot of the pseudo arguments seem to be that if Intel isn't doing it then it can't be a good idea. This could be true but time will tell.
Pop Catalin
"this type of competing based on costs only won't get AMD far.
They need to compete on performance and efficiency first of all."
I agree. AMD needs to get a good volume of the B3 stepping out and release the 2.6Ghz speed and try to stay on track with Shanghai (which hopefully is faster). And, hopefully AMD will be able to continue to increase speed during 2009.
" Even if AMD manages to cut the costs in half, they are still pushed in the lower end, where profits are lower and lower, by Intel's agressive pricing policy."
Again, I agree. AMD's primary actions which I mentioned before like tri-core and L3 cacheless quads are low end strategies. This is a bit like tossing out a second anchor to stay off the rocks. But, AMD still needs to get the engine fixed so it can move away.
"Cutting costs alone won't solve AMD's problem in the long run, it can stop a litle of the bleeding in the short run and that's all."
Well, hopefully it will stop all of the bleeding.
1st in the interest of discloser, I am and have been an INTC shareholder. Therefore, I do not by any stretch claim to be objective in these maters. I am in it for the performance of the stock, the company, and its products.
However, having a material stake in the company puts me in a fairly objective position. I will not stand by idly and watch any company I own tank for the sake of “fanboyism”. I will get out when deemed prudent.
Many things have transpired since the 2nd quarter of 2006 to the present. As you well know, fortunes have changed dramatically for both companies, in technologies, execution, product offerings, and ultimately, profits. The numbers don’t lie.
Here’s the point as far as “DOC’s”, if you will, blog. The essays contained on his timeline have been remarkably accurate, despite naysayer, analyst warnings, and paid media hypes with an agenda. In fact, with all sincerity, I defy anyone to go through Robo’s timeline and find anything the was inaccurate or wrong. Frankly, everything I have read has come to pass with clairvoyant accuracy, Barcelona’s top production speeds and thermals, for example.
With all due respect to you and the Robo"s ‘trolls”, can your timeline support the same claim with same precision? Many claims were made from AMD during the last 6 quarters, many were posted here, few of which came to fruition. On the contrary, can you sight a specific example when INTC misled its share holders, partners, and customers with delays and misleading statements in the interest of self preservation?
I can’t argue from a technical perspective. I simply don’t have the foundation or background. The “trolls” as you call them have been remarkably accurate on process technology and engineering, qualified by my own independent research. The timeline is the evidence, my dividend check is the reward, and my new forthcoming processor is the proof.
Sincerely,
SPARKS
sparks
"In fact, with all sincerity, I defy anyone to go through Robo’s timeline and find anything the was inaccurate or wrong."
Things roborat himself was wrong about or things wrong posted on his blog? Roborat himself has been pretty accurate about Intel and mostly accurate about AMD. Things said on his blog by Intel employees about Intel have of course also been accurate. Things said on his blog about AMD much less so.
"Frankly, everything I have read has come to pass with clairvoyant accuracy, Barcelona’s top production speeds and thermals, for example."
Frankly it is difficult to have any respect for someone who encourages flame posting and trolling. I don't do that here.
"With all due respect to you"
You have none; why pretend?
"and the Robo"s ‘trolls”
Who often make up things that I never said, pretend to disagree with me when they actually repeat what I already said, and apparently skim through my articles with very little comprehension.
At any given time usually half the things there attributed to me are simply incorrect. That doesn't even count all the insults and personal attacks from children hiding behind an anonymous post. Again, hard to have respect for such people.
"can your timeline support the same claim with same precision?"
My grasp of things and predictions were very accurate from 2003 through 2005. I have been far less accurate since C2D was released. I have been doing better in the last four months. We'll see if that trend continues.
"Many claims were made from AMD during the last 6 quarters, many were posted here, few of which came to fruition."
That is certainly correct. AMD has failed its projections again and again. This has been pretty much the case since Q4 2006 through Q1 2008. There is little positive worth noting from that time. Very recently this has begun to look better. Again, we will see.
"On the contrary, can you sight a specific example when INTC misled its share holders, partners, and customers with delays and misleading statements in the interest of self preservation?"
I could name quite a few of them but the number has dropped to only a fraction since Intel released C2D. Again, hopefully Intel's accuracy will stay fairly high and AMD's will improve.
SPARKS
"1st in the interest of discloser, I am and have been an INTC shareholder."
I have never owned either Intel or AMD stock. I don't work for either company or any subsidiary or any company that does direct business with either AMD or Intel. I have no stake in the performance of the stock from either company. Nor do I have any vested interest or stand to benefit in anyway by the peformance of either company.
I have purchased both AMD and Intel processors individually and systems using both AMD and Intel processors. My computer system purchases have been: Intel, Intel, AMD, Intel, Intel, AMD, AMD, Intel.
To be honest I would have had trouble recommending K10 prior to the recent B3 stepping. I see quite a bit of overlap between Intel and AMD systems. But, then I don't overclock nor would I recommend it.
"Therefore, I do not by any stretch claim to be objective in these maters. I am in it for the performance of the stock, the company, and its products."
Good to know.
“I have never owned either Intel or AMD stock. I don't work for either company or any subsidiary or any company that does direct business with either AMD or Intel. I have no stake in the performance of the stock from either company. Nor do I have any vested interest or stand to benefit in anyway by the peformance of either company”
Then the question begs, why such passion?
From your previous post “personal attacks and insults”, I agree with you whole heartedly. There is no need for that at all. Further, I haven’t, to my knowledge, attacked or insulted you on a personal basis. Yet, you claim I have no respect, “You have none; why pretend?” couldn’t this be interpreted as an unwarranted personal attack?
“Good to know”
Academic, my disclosure was out of respect to you, your site, and ultimately, myself. Why therefore, is it good to know? Will it change the value of my opinion and/or analysis of either company, perceived or in fact? Actually, if the previously mentioned fortunes changed as dramatically as they did in 2006, I would look at this as a great opportunity, not a detriment. A change of position is a few mouse clicks away.
If this makes me an opportunistic capitalist slime, well, no ones perfect, I’m in good company. More importantly, however, it gives me an edge, as I have much more to gain (or lose) than, shall we say, mere pride? I cannot afford to be that altruistic; I don’t have THAT kind of money. Consequentially, and ironically in my position, objectivity is mandatory.
That’s what’s good to know.
What I object to is misinformation disclosed, wittingly or unwittingly, whereby investors and shareholders are led to believe their positions are sound, when in fact; their positions have been compromised, directly in the face of evidence to the contrary.
SPARKS
sparks
I'm sorry but it is just very difficult for me to understand why someone would prefer the locker room/clique mentality of roborat's blog. Apparently you like that environment and that is your choice. I'm also baffled at the way I am constantly compared with Sharikou over there. I would assume that most people who can read could see that my articles bear no ressemblance whatsoever to Sharikou's.
There is far less about roborat's blog that is concerned with accurate, objective information as you suggest and far more that has to do with being one of the boys and being part of the crowd.
The really ironic part is that if I wanted to I could certainly make my blog very pro-AMD but I have not done that. It amazes me that the people who post on roborat's blog and Roborat himself are not smart enough to see that. You yourself suggest that I have passion for AMD. That actually is not the case.
Right now I could see the Q6600 and the Allendales as pretty good bargains. I don't see the Conroe's as bargains because by the time you get up to that price you could get a Q6600 for just a little more. I don't see the AMD tri-cores as bargains because they are still B2. The B3 steppings are finally a reasonable quad offering from AMD. I'm a bit divided over Black Edition versus stock Intel quad. The Intel chips have more OC'ing potential but you have to overclock the FSB and this is more involved than just increasing the clock on the Black Edition chips so both have their place.
A dual core is the lowest I would go in bargains right and both the Allendales and K8's look good. The Allendales have lower performance than Conroes so they are a lot closer to the K8's and you could use 35 or possibly 38 series for Intel or a 690G board for AMD. The C2D chips are stronger in raw SSE but if you are looking for a bargain then you will probably stay with onboard graphics and the AMD chipsets are certainly stronger there. The 9850 Black Edition with 7xx series chipset looks pretty good for a mid range system but you could put together an Intel system of similar value with again stronger SSE but worse graphics and video playback. However, as soon as you decide to go to discrete graphics the AMD chipsets lose their advantage.
I'm sure your buddies at roborat's will be whooping and cackling over the idea that I am not a die hard AMD fan. But being sophomoric does not make you right. I would have no trouble suggesting Intel based machines depending on what someone was looking for. The reason I tend to talk about AMD is that you can always find articles, reviews, and editorials favorable to Intel. Also, the average computer buyer will never add discrete graphics and for these people AMD's better chipsets would be good choice. However, this choice will tend to be drowned out by the cacophony of Intel followers.
I apologize for not being clear.
“Then the question begs, why such passion?”
The question wasn’t implied, or other wise directed, towards your passion regarding AMD. Not at all, the passion I refer to is your site and how you handle this blog. Allow me to preface this by saying it yours do what with you will and it’s none of my business.
Look, from my perspective, a blog, on a post by post basis, represents a chain of thoughts and idea’s, given freely. There is a flow, an exchange of ideas that each poster contributes to on an argument by argument basis, which will take a natural course if left unimpeded. It really doesn’t matter, if the individuals are factually or empirically correct or incorrect, that will sort it self out naturally, by each contributor.
That’s if it's allowed to run it's course, a free exchange of ideas, if you will.
However, categorically dismissing someone ideas or opinion because they are not in agreement, or fulfill specific criteria, does not lead itself to a free and open exchange. Frankly, this is precisely where you and The Doc differ.
For example, I believe I have the distinct honor of have my first post deleted. I freely admit, it was sophomoric and condescending, but it was my first post that led us up to this discussion, the chain was broken. No one will ever know how or why:
“That guy SPARKS can be an asshole”
Some may have enjoyed the wit, some may have hated it, and the choice was theirs, rightfully so.
By deleting that post you’ve denied me the opportunity to be whom and what I am, right or wrong, to others who read and contribute to the blog. You have denied others to make that determination themselves and the luxury of responding to it. This is not free exchange, it’s moderate and filtered by what you personally deem as an acceptable or unacceptable contribution.
If read things correctly, this is what bothers the guy’s at Robo’s blog most. There is no free exchange, here. You yourself admitted to being “less accurate”. Perhaps they saw the inaccuracies and tried to argue the point. Then again we’ll never know. Some things, obviously, may have been deleted. Some may have given up posting here at all.
The Doc has no such agenda. He rarely posts, almost never deletes, no one to my knowledge has ever been banned. His ‘hands off’ strategy has fostered an extremely rewarding, and in my case educational, experience. I have made some friends (there are not many people I can talk computer with), and indirectly developed a repartee with you.
He takes nothing at face value and nothing personally. The number of posts is the evidence and my respect for the man goes beyond words.
SPARKS
Your right about the gains that AMD will get by continuing to lean out their process. Goal is single part flow rather than batches. Batches by their nature increase cost, cycle time, and quality.
I use the "simple penny experiment" with clients (Discretionary Thoughts) that goes like this:
-Use 5 pennies and model the two different production methods. The "plant" has four machine stations (corners of a desk) and one second to process a penny and one second to move a penny from one station to the next.
Batch operation:
-all five pennies are processed at each corner station as a group (five seconds to complete each station).
-You'll note that four pennies are sitting idle at each corner station while one is being processed.
-You'll also see that repeating this over time that each corner has four idle pennies (for 16 pieces of excess work-in-process).
Single part flow:
-one penny is processed at each corner station.
-no pennies are sitting idle (other than the pile at raw goods at the beginning and finished goods at the end)
-you'll see that total cycle time to process those five pennies (one order) is much lower than the prior batch process.
I've described the above quickly, hopefully you can follow it, but what it means is a manufacturing plant practicing Lean principles will begin to approach the reduction potential of 70% less WIP/in-process-inventory, 70% less floor space, much lower quality problems (scrap out one bad penny instead of five).
Then that's also not challenging the equipment changeover issues (often even more dramatic of an improvement than going to single piece flow).
Done right, AMD can reduce costs, increase quality (and increase yields which is lower costs) get to market faster and achieve stellar stock performance. If the product engineers are freed up from having to fight manufacturing issues, since the quality has improved with Lean practices, then they can really focus on building the next chip generation.
John Smith
sparks
We are talking about two different things. I will not allow the use of profane language here regardless of someone's feeling of self importance, free spiritedness, that other places allow it, or diatribes about free speech. This is a publicly viewable blog, not a pool hall or a locker room. You may see that as being a forum for free discussion; I see it as laziness on the part of the blog owner.
I also don't like dismissive personal attacks (which is routine for the crowd at roborat's). I'm not biased in this; I don't allow personal attacks against Intel enthusiasts either. There really isn't much of any other limitation. I've never complained when people posted news favorable to Intel in the threads or news that was bad for AMD.
The bottom line is that if you can't have a discussion without calling someone a fanboi or using foul language then you don't need to post here.
sparks
"However, categorically dismissing someone ideas or opinion because they are not in agreement, or fulfill specific criteria, does not lead itself to a free and open exchange. Frankly, this is precisely where you and The Doc differ."
That is where you are incorrect.
I have no problem with someone posting here who believes that Intel has good products; they do. Or someone who notes that Intel is dominant or doing well financially; they are. Or someone who realizes that Intel has been on track and pretty close to what they said they would deliver for the past 2 years. That is quite true. I have no problem with anyone who says that they feel that Intel is better at OC'ing right now, quite true.
Frankly, I'm baffled what idea that you think I'm concerned about.
However, there is no reason to have an endless succession of "Intel rulz, AMD sux" posts. There is no reason to have constant "AMD BK" posts or "you're just a stupid fanboi" posts. That accomplishes nothing. If that is what you feel open discussion is all about then you are better off on roborat's blog.
In all honesty, the difference between roborat's blog and mine is that he encourages flames and I don't. He let's posters hide behind an anonymous post and act like children; I don't. I'm sorry but that is no improvement for roborat's blog.
john
"Your right about the gains that AMD will get by continuing to lean out their process. Goal is single part flow rather than batches. Batches by their nature increase cost, cycle time, and quality."
Well, yes and no. Batches work great with memory products and are essential to Foundries. I would imagine that this is the least cost method for Intel. AMD believes that smaller batches will help them. We'll see if that is true.
"I use the "simple penny experiment" with clients (Discretionary Thoughts) that goes like this:"
Thank you for the description but wafer manufacturing is quite a bit more complex than that. To be honest I have no idea how AMD's cost structures and processes are cost limited. It would be difficult to have a real discusion about that.
"I've described the above quickly, hopefully you can follow it, but what it means is a manufacturing plant practicing Lean principles will begin to approach the reduction potential of 70% less WIP/in-process-inventory, 70% less floor space, much lower quality problems (scrap out one bad penny instead of five)."
Yes, I understand the principles with no problem. I just don't have enough specific information on cost, workflow, and time studies to put any reality on the discussion.
Being in manufacturing (Mil Aero) and having been at a commercial contractor during my career, I can fully appreciate what AMD is trying to do. If your supply chain is in order, eliminating batch orders is a great thing to do.
Typically when you are running in batches, you'll have one work order into the process followed by another. If you get to the testing phase on one batch, while another batch is finishing build and another is in a state of build, and you find out you had a bad part (say a PWB in circuit board manufacturing, or in AMD's case a CPU wafer) then you have just lost three batches. If you are running in piece parts, then you'll generally have a lot less scrap.
It also will give them flexibility to do two things: 1) meet varied core orders (say notebook chips versus server) and 2) apply continuous improvements to their CPU's. One of the biggest problems AMD had with Barcelona was the amount of time it took to perform a re-spin (3 months roughly). As they work on the time to reduce the lith mask production and the time it takes to apply each mask to create a CPU die, as well as the time to get a working sample to test, they're ability to respond to bugs in the product will greatly increase.
Had AMD already implemented these prior to the Barcelona sample runs, I suspect a B3 Barcelona would have been produced back in November or possibly earlier depending on how long it's currently taking them to get new masks made (which also depends on how many layers are required to be redone). As it is, we saw them painfully, and slowly, crawl out of the gate with B3. Instead of having worked on speed improvements to get K10 up into the 3GHz range (where it becomes very competitive on everything except SSE4), AMD instead has basically dropped K10 development for K10.5. Now they have to hope that 45nm will be as good as they believe their samples are showing.
I suspect that the court issues for Intel in the US and EU may end up awarding AMD billions of dollars (more suspect the EU to do this than the US). If that happens, AMD can immediately liquidate debt and become a profit maker again, even in it's current position. This is going to be an interesting year for AMD. Let's hope they get themselves up off the floor and get back in the race.
richard
I'm not an expert on wafer manufacturing so if someone has more specific information feel free to provide corrections.
My understanding is that batch tools have been the only tools capable of delivering the necessary volume. Single wafer tools were created when particularly difficult processes needed to work on one wafer at a time but this was not ideal. The compromise has been mini-batches or tool clustering.
It could be the case that smaller batches work better for AMD because of external factors. I'm certain that the current batch system works well for Intel. But, if Intel has an issue with a die they work it out at D1D and then send it to other FABs. AMD does run test wafers inline with production so it is possible that this could benefit AMD more with testing. However, the normal carrier size for wafers is 25/pod so this would have to be broken down further somehow. This is what AMD says but I don't know what the method would be.
Semiconductor International
Van Eck said the ultimate goal is to have multiple lots within a single 25-wafer front-opening universal pod (FOUP). A company doing prototyping, for example, might run six wafer lots to reduce cycle time. Theoretically, a lot could be a single wafer within a FOUP, or any number between one and 25 for that matter.
A 50% cycle time delay is another goal. If a two-month cycle time is common in today’s fabs, with rush lots requiring three to four weeks, Van Eck said, “We want a couple of weeks to be normal.”
If AMD can actually make this work then I could see benefits.
scientia
"In all honesty, the difference between roborat's blog and mine is that he encourages flames and I don't."
Funny thing is I've seen more flames on your blog than his.
"He let's posters hide behind an anonymous post and act like children; I don't."
This is evident from the wannabe-Hector posts sitting here for days.
"I'm not an expert on wafer manufacturing so if someone has more specific information feel free to provide corrections."
Check out Roborat's blog, you might actually learn something about the subject
Ho Ho
"Funny thing is I've seen more flames on your blog than his."
This is obviously wrong. I can find more flames and personal attacks in one of robo's threads than I can in ten of mine.
"This is evident from the wannabe-Hector posts sitting here for days."
The Hector posts were made by one of the children from roborat's blog. Again, this type of behavior is encouraged by roborat himself.
In a great blow to Intel fans everywhere ZDNet, George Ou
As of March 26, 2008, I have to say goodbye to ZDNet because of some corporate restructuring. I have no ill will towards any of my managers and I hold no grudges towards the company. This will be my last and final post on ZDNet so farewell my friends.
Well, I guess they still have Kubicki at DailyTech.
Ho Ho said
"I guess if you label people who say you are completely wrong in what you as flames then you might be right."
What you just said makes no sense at all. People can say that I'm unknowledgable, inexperienced, lacking in technical expertise, beyond my experience, mistaken, oversimplifying, distorting, or just plain wrong as many times as they like. None of these have anything to do with flaming.
"I agree that some times thay aren't that gentle with this but I can understand what they feel if they have to keep on reading the stuff you make up here."
You are so full of it. No one has to read my blog. This is just a very poor attempt on your part to excuse inappropriate behavior by posters on roborat's blog which is again encouraged by roborat himself.
"How exactly does he do it? I don't remember when I saw such spam on his blog."
Spamming my blog is no different from writing dirty words on bathroom walls. Posters routinely do the equivalent on roborat's blog. If you haven't noticed this then you are either blind or have no comprehension of English. Either that or you are once again excusing this behavior because it isn't directed at anything you care about. You are afterall part of the clique.
sparks
I know you think roborat's blog is brilliant and accurate (you've repeated this at least three times) and I know that ho ho will make excuses for everything that is posted there as will roborat himself.
However, it is quite easy to label a site as brilliant when you only count the fraction of posts that are authoritative and accurate and ignore the rest of the garbage.
For example, in one of the many diatribes there where I was yet again being told what an incredible idiot I am:
"I did like the interpolation of 9850 being a 2.5GHz chip....BRILLIANT!"
Obviously referring to me in the polite way that I've come to associate with roborat's followers. He goes on to say:
"I think we will probably never see a 9850 as AMD already would have 6 quad core desktop SKU's if/when the 2.6 GHz chip comes out."
I'm guessing that this is a typo and he actually meant that 9950 would be the 2.6Ghz model and that is what would never be released.
However, this is exactly the kind of mistake that if I had made it there would be at least three separate posts on roborat's blog cackling about how stupid I must be to make such an obvious error. Secondly, even ignoring the typo, this comment will have been completely forgotten when a 2.6Ghz model is released and posters there will again claim that only the highest wisdom exists there.
They too will gloss over the slow morphing of opinion over time as speculation changes to fact and they will conveniently forget their own incorrect speculation. Instead they'll spend time digging up anything that I got wrong and have a great many laughs over that.
There are legitimate corrections there but there is also a very childish and mean spirited delight in making fun of people in the most mocking, derogatory, and even crude fashion. And, whenever someone there makes such a comment the others close ranks and slap each other on the back and laugh right along with them. These are men who never outgrew their high school mentality; there is nothing to respect about such behavior.
I guess the most disappointing aspect is to realize that some of the people there actually work for Intel. Professional? Polite? Instructive? Not even close. I have to wonder if it is just the dregs of Intel who post there or if they are taking out some frustrations they have as Intel employees.
Hi Scientia:
What are your thoughts about this piece ?
erlindo
I don't really have any thoughts about it. It says they are speculating based on rumors.
‘I know you think roborat's blog is brilliant and accurate “
Yes, on the whole, I do. Some posters, however, haven’t been quite so accurate. And they’ve been told so, (including yours truly). When a poster continues to argue from a materially bad position, he will indeed get a full frontal assault, from all. Sometimes, true enough, it’s not pretty. Speaking for myself, I am wise enough to know my limitations, acquiesce, do a little research, get an education and move on. I have never been ridiculed, ever.
But here’s the thing, and there’s no getting around this. Apparently you do read the blog. You must admit some of the technical discussions are brilliant. With the content, depth, and flow of some of the essays, there is absolutely no way this material hasn’t come from actual first hand knowledge.
I have gotten an insight into this very secret, complex, and fascinating world. I couldn’t have gotten that from anywhere else, all on a personal one to one level. Most importantly, I will never try to be anymore than whom I am, they would eat me alive.
However, if someone crossed the line into MY area of expertise, say powering up an entire FAB facility, from the ‘street’ to the ‘tool’, and got out of line, THEY would get eaten alive. Trust me. Perhaps, I can identify with their arrogance in their field; actually I get a kick out of it. I am just as arrogant in mine. Maybe more so, you ain’t gonna get those fancy machines running without good clean reliable juice from me.
That’s a fact.
At the risk of being redundant, I like those guys because I can get into a FAB vicariously through their experiences. I am getting education on a fascinating industry I have grown to love. I am a computer hardware freak/enthusiast. I really can’t really talk computer with too many people. And finally, I am an educated investor having fun.
Hey, I can take the good with the bad. If I don’t like what you’re posting, you’ll know. The DOC gives me Carte Blanche to say so. He has never deleted me, and I respect him for it. It’s not right or wrong. It’s my prerogative and his.
SPARKS
sparks
I'm sorry but I can't agree with you.
I remember the textbook we had on compiler design. The author had 30 years experience in programming but the textbook was awful.
I recall two men both with decades of experience in experimental aircraft who had completely opposite points of view on the idea of using automotive engines in light aircraft. The interesting thing was that even though they completely disagreed they were both right.
I've studied the subject of deep water sailing. You only need to look at a good reference like Alard Coles' Heavy Weather Sailing to see how many different opinions there are about how to deal with this situation. Many advocate running before the wind using warps or drogues. The Pardey's strongly advocate heaving to. Yet, I've seen heaving to dismissed by many who have a background in racing.
I can also think of plenty of engineering disasters (by supposed experts) from the Tacoma Narrows bridge to the opera house in Sidney to the Hancock building in Texas. I guess I could also mention the Liberty ships which broke in two before they were reinforced with a band of steel on the sides.
I've seen variation in professional opinion, differences in starting assumptions, and inability to apply knowledge and experience from one area to another. I can't accept the idea of arrogance or using bullying tactics to make a point. If that makes you feel at home then that is your choice. To me your arguements are like arguing in favor of full contact golf.
sparks
"But here’s the thing, and there’s no getting around this. Apparently you do read the blog."
roborat's articles tend to be short so I probably read a third to half of them. I seldom read the comments.
"You must admit some of the technical discussions are brilliant."
I don't recall ever reading a brilliant technical discussion on roborat's blog. However, this could be something that you've gotten out of the comments which I don't normally read.
"With the content, depth, and flow of some of the essays, there is absolutely no way this material hasn’t come from actual first hand knowledge."
That may be the case but I don't really have time to screen through the sludge of comments looking for nuggets of wisdom. This seems to be a lot of the same crowd who posted on Toms Hardware Guide. And, in those forums the ignorance about AMD processors was staggering. You could find mistake after mistake after mistake.
I have no doubt that the people who work at Intel who comment on roborat's blog are knowledgable about Intel. However, I'm not so so sure about the level of knowledge about AMD. I'm also highly skeptical about any level of objectivity.
You asserted that anyone who posted something false would be vigorously attacked. I think the truth is that this would only apply to positive things about AMD or negative things about Intel.
scientia
"I think the truth is that this would only apply to positive things about AMD or negative things about Intel."
Just read the comments and you'll see for yourself. The articles roborat posts aren't usually worth even a fraction what is written in comments. As I too don't have much time I have to limit what to read on the net. Between robos and e.g amdzone I've picked the first. You might want to do the same as you'll get much more out from your time that way.
Approximately how much money, give or take a few billion dollars, could this allegedly described scheme save AMD?
ho ho
"The articles roborat posts aren't usually worth even a fraction what is written in comments."
I see. Still looking for good information would require a lot of sifting; most of the comments are garbage.
nanikore
"Approximately how much money, give or take a few billion dollars, could this allegedly described scheme save AMD?"
I'm puzzled why you would describe it as alleged scheme. AMD has had four meetings on this since early 2007. If Intel had had one meeting you would consider it a proven fact.
Nanikore-
I don’t mean to criticize, but with a name like ‘nanikore’, I suspect you’re in the industry.
If so, you already know the answer, probably better than most. Sounds like your baiting ole Sci here.
However, if your question wasn’t completely rhetorical, the answer, be it “Asset Smart “or “Asset Lite “, it would be billions. Basically, given AMD’s serious financial condition, in conjunction with an extremely tight capital market, FAB affordability would be almost impossible at this juncture, NYS FAB construction incentives inclusive.
As you know, tools are EXTREMELY expensive, and from what I’ve learned recently, getting a production line qualified, up and running, and finally, producing a HV competitive product is about 4-5 Billion. That’s a LOW estimate. When you factor in depreciation, volume output production curves, (low initially, high hopefully), R+D, and a myriad of other unseen costs, a lot more. Then, you have to execute well enough to compete with Chipzilla, who is considerably ahead on the performance and process and HV curve.
I think AMD’s strategy is to manufacture a performance/cost solution by using the BIG foundries and their tooling, thus alleviating AMD from incredibly high initial capital expenditures.
From what I have been hearing, there are two major obstacles to this scenario, however. One, being less control over the manufacturing process itself, and secondly, x86 Licensing issues with INTC. I think the agreement is 20% of total x86 production.
There is another, and it could be argued by the way, TSMC and Chartered Semi are a bit behind INTC on the process technology scale. I think more than a bit, personally.
All said, if AMD can outsource a good portion of it CPU production without violating any licensing agreements, push out competitive solutions with good margins, they may interest someone with fresh capital to insure their continued sustainability. (By the way, the graphics division ATI has always been fabless.)
Perhaps, if AMD, optimistically speaking, can hit a homerun CPU product, simultaneously beat back NVDA, and/or develop killer platform solutions, outsource everything BUT 80% of CPU's, they may pull out of this financial death spiral. (They have till 2012 before the two 2007 share dilutions mature.) Personally, I have my doubts.
It’s too bad they didn’t take this strategy 18 months ago.
That said, time will tell, as all of us who have been following this corporate daytime drama are waiting for a clear and substantive outline of what AMD will in fact, be doing next.
The rest is conjecture and speculation at this point.
But, you already knew that. ;)
SPARKS
I really have no idea how much they would save but I think Billions is too high. Supposedly AMD will reduce its expenses by $500 Million a quarter but that is not related to SMART.
If they can actually pull that off then whatever they can get from SMART is extra. Assuming they can get back into black an extra $100 Million a quarter would be quite welcome I'm sure. This is also related to AMD's desire to defer moving to 450mm tooling. However that is not direct savings.
sparks
"I think AMD’s strategy is to manufacture a performance/cost solution by using the BIG foundries and their tooling, thus alleviating AMD from incredibly high initial capital expenditures."
I don't think this scenario is likely at all. In fact, everytime I've thought about it I've never been able to figure out a way to get it to work. And, the reasons it won't work should be obvious. First of all, TSMC's process is not the same as AMD's. This means that AMD's cpu's would have to be re-engineered to work on TSMC's process. I can't imagine that AMD can spare the effort to do this.
Secondly, AMD is short of revenue right now, not in six months or a year but right now. Yet, sending chips to a foundry would reduce AMD's revenues today, in fact, it would cut AMD's revenue's drastically. And, supposedly this is so that AMD could defer spending money on tooling? This makes no sense. AMD has an empty FAB and apparently is not hitting its limit on production capacity. The amount of tooling that AMD might have to buy in the next 12 months should be less than the loss would be due to using a Foundry.
"From what I have been hearing, there are two major obstacles to this scenario, however. One, being less control over the manufacturing process itself, and secondly, x86 Licensing issues with INTC. I think the agreement is 20% of total x86 production."
In addition to the fact that it makes no economic sense.
"There is another, and it could be argued by the way, TSMC and Chartered Semi are a bit behind INTC on the process technology scale. I think more than a bit, personally."
I don't understand how you could lump together Chartered with TSMC. Chartered is fully certified to second source AMD; that would be no problem. However, TSMC does not have the same process.
"Perhaps, if AMD, optimistically speaking, can hit a homerun CPU product"
I don't think AMD needs a homerun to become profitable again.
" simultaneously beat back NVDA,"
AMD's chipsets are better than nVidia's and Intel has had no trouble holding onto its chipset share even while being inferior to nVidia. Intel's integrated graphics are improving though and AMD is looking to expand its chipset range to mobile and servers. It seems to me that better integrated graphics will begin to cut into discrete graphics which is where nVidia makes most of its profit. Not only that but Intel will also release its own discrete graphics products in 2010 when AMD releases a server chipset. I see increased pressure for nVidia, not a rosy future.
"outsource everything BUT 80% of CPU's,"
I'm not sure here what you are trying to say. AMD can already do this by using Chartered and has been able to for the past year. How would this be a new advantage?
"It’s too bad they didn’t take this strategy 18 months ago."
What strategy are you talking about?
That said, time will tell, as all of us who have been following this corporate daytime drama are waiting for a clear and substantive outline of what AMD will in fact, be doing next."
Supposedly AMD will break even by Q3 2008; that isn't long to wait.
Is "Asset Smart" here?:
www.amd.com/us-en/assets/content_type/ DownloadableAssets/AMD_Fact_Sheet.pdf
Post a Comment