Thoughts of an ex-AMD cpu designer 1997-2006 (Cliff Maier) about AMD from about a year ago but maybe still relevant.
AMD has sucked for several years, and they still have nothing worthwhile in the pipeline. They went horribly awry 5 years ago, but it took a couple years for their failure to be noticed (when you cancel an important project or two, no one notices until 2 years later you're still spinning the same chips).
Because AMD has no roadmap. Their engineering is in complete disarray. They're selling chips that were designed a decade ago. They have no chief architect. Stick a fork in 'em.
The team that did all the great design work for Athlon 64/Opteron is gone. Forced out or quit in disgust. They now work at Apple, Oracle, misc. startups, and, in my case, changed careers.
Don't believe everything you read. Also don't believe AMD can execute.
The team that designed the K6-2 was the CMD team, which was formed by the acquisition of a company called Nexgen. That team also designed Athlon 64 and Opteron (Athlon was designed by the TMD team). By 2007, all the key CMD folks were gone. The team that was left sucks, and has accomplished little since then other than shrinks to smaller technology and bolting more of the same cores on.
Hey, I left on my own. AMD has not made a real profit (other than 2 one-time events) since Q4 06. They had the performance and performance/power lead with Opteron and Athlon 64 (K8). Now they lag in both performance and performance/power. They lost so much money they had to spin off their fabs into a new company just to get cash to survive.
Call my bluff all you want, but everyone here knows the facts, and knows that AMD actually had a lead over Intel for 2 or 3 years due to its high performance 64 bit cores and integrated memory controller and point-to-point bus. Then they sat on their heels and did nothing because the K7 (Athlon) team took over when Jerry Sanders retired and was replaced by Raza (the guy who decided to spill corporate secrets to everyone so they could make money in the market, apparently), and the folks in Texas resented the folks in California.
They had the performance and performance/power lead with Opteron and Athlon 64 (K8). K9 never happened. Ever wonder about that? Now they lag in both performance and performance/power. They lost so much money they had to spin off their fabs into a new company
Yes, that is true. I designed the 2-core, but left before the 4-core.
Yeah. Like replacing the 40-man team that designed A64 with a 250 man team for minor design changes. And like throwing away all the EDA tools because the vapor tools from the texas design team would be better if they ever were finished.
Intel's yield is much higher than AMD's.
Anyone who thinks Apple is going to buy AMD is high.
1) Apple would not be allowed to manufacture x86 processors. The license AMD has to Intel's patents is not transferable to a new entity (whereas Intel gets to keep it's AMD64 license in such a case).
2) AMD's processor design teams (can't comment on the GPU teams) are bloated, inefficient, and weak - the best folks had gone to Sun, PA Semi, Apple, Montalvo, Metaram, and various other startups years ago (which is why AMD has flubbed the lead it had with Athlon 64).
3) Apple would make an enemy of Intel, and AMD does not have a range of processors to fill all of Apple's product lines (even assuming somehow it was still able to make x86 processors).
4) Many of the best remaining employees would leave. Managers would be locked in, but the last thing Apple wants is AMD's managers.
5) AMD would make an enemy of nVidia, meaning it would have to transition all of its machines to AMD+ATI very quickly.
6) AMD has never delivered on its roadmap.
There's no way it's happening.
Apple can demand all it wants. AMD can't deliver.
The evidence is that when the DID compete they had people like Fred Weber and his team. That team left years ago, and AMD hasn't competed since. And the people they currently have don't even care.
Except the AMD and VIA licenses to not apply if the design is not an "AMD" or "VIA" design. However, you forgot the most obvious choice - IBM. They have a license to fab anyone's x86 design. All you have to do is pay them.
Nonsense. You act like I'm not still plugged into what's going on there. That I don't go drinking with some of the current employees. That I don't know the people in charge VERY well having had years of experience working with them before (and while) they were in charge. And 2002, 2003, 2004, 2005, 2006 is a lot of years of screwups while I was still there. 2007, 2008, 2009 AMD continued on the path from the previous 5 years - they did not tape out anything interesting, and continued making spins on designs from early 2002. They spun off their fabs, so they no longer can influence the process they way they used to. Their CEO had to resign in disgrace. The new CEO is a guy that once told the design team (when he was a manager in Texas) "if you don't like it, quit" - and 60 out of 100 people in Sunnyvale quit within a month. They used to hand place and hand instantiate each cell in the design for maximum efficiency and speed - now they rely on tools which perform 20% worse than humans (in order to save money).
I know EXACTLY what chips they are currently working on and WHO is working on those chips.
Think what you want. I find it hilarious that if someone says something bad about a former employer, everyone wants to assume it's some sort of evil grudge. It's just honesty.
That sounds like Dirk Meyer upsetting the natives. I always thought/stated that AMD's cpu problems laid with him rather than the more common scapegoat Ruiz.
My friends work there because of golden handcuffs.
I can't speak as to the ATI division, but, yes, the processor division is doomed. You don't have to believe me. Look at their track record for the last 8-10 years, and keep in mind that any new microarchitecture that is sold takes 2-3 years of design time. Then look at their 10K's, their desperation moves (selling the fab, Arab investors, etc.), the well-publicized defections (Fred Weber, etc.) and put it all together. Look at the benchmarks over time. Look at their stock over time. If you choose to write off my statements, there are plenty of objective facts out there for you.
Sometimes a duck is a duck, and the fact that I am an ex-AMD employee calling it a duck doesn't make it any less of a duck.
Athlon 64/Opteron were excellent chips. So were the x2's and the initial x4's. But they waited too long and Intel caught up by putting its own memory controller on-board and by using a point-to-point interconnect, and they have no visionary. They don't even have a real architect.
Guys - I have nothing to be disgruntled about. I left on my own accord. I had great times at AMD making a64 and creating amd64. AMD paid me enough to pay for a house in silicon valley and a porsche. I quit on my own and was not pressured to do so. They were sad to see me go, and i was sad to leave
I am not disgruntled. I am sad. Sad that AMD squandered their lead because new management decided that the way to compete was design CPUs the way that chips in toasters are designed. And worse, to do it with giant teams of unskilled designers instead of a small team of highly experienced engineers who know how to design transistor by transistor.
Now the whole company is organized on this basis. It's not going to change. It sucks, because AMD had a huge opportunity. A superior product. A superior team that was eager to win. And a sleeping and disoriented competitor. Those three factors are not coming together again before AMD runs out of cash.
On paper bulldozer is a lovely chip. Bulldozer was on the drawing board (people were even working on it) even back when I was there. All I can say is that by the time you see silicon for sale, it will be a lot less impressive, both in its own terms and when compared to what Intel will be offering. (Because I have no faith AMD knows how to actually design chips anymore). I don't really want to reveal what I know about Bulldozer from my time at AMD. I know less about bobcat. From what I can tell it's just the latest name for a project that had been kicking around since 2005 when we acquired new design teams in two new locations. None of these designers had any experience with design in the <10W range. I don't know any of the people currently working on bobcat, but given the price it is likely to sell for, it's not going to make AMD a lot of money (and it will be competing not only against Intel, but against numerous ARM variations).
Sure, the playing field hasn't always been fair. But at the same time, in the lawsuit, papers from AMD that said even AMD's head of sales thought intel chips were better also came to light.
In any event, it doesn't matter why. The result is the same.
As for the accusation I am only making unverifiable statements, that is nonsense. I've referred repeatedly to things like stock price (once over $42), the lack of an on-going (as opposed to one-time) profit since 4Q06, the performance lead of AMD chips in the Athlon 64/Opteron days that has since evaporated, etc. All of these facts are completely verifiable, as are other facts that have been mentioned (AMD's financial performance prior to Athlon 64/Opteron, AMD's failure with K5 and the fact that it bought Nexgen which took over K6 and K8, and directed the development of K7), the exodus of key architects like Fred Weber, the current lack of an actual architect, the lack of cash on-hand, the sale of the fab business to raise money, the sale of the flash business to raise money, the desperation investments by arab sheiks, etc. Also, all verifiable facts. The ascendency of Atiq Raza who moved the headquarters from silicon valley to texas, gave away corporate secrets, and had to resign. In its 40 year history AMD has had about 3 good years. Easily proven facts. The future is not hard to predict based on all this PUBLIC evidence. I add my non-public knowledge, and it merely verifies what should be apparent to everyone already.
Of course. AMD's yields have always been terrible compared to Intel. Also, AMD has never been able to compete in the entire range (from workstation to laptop) meaning that big customers were unwilling to try us and risk pissing off Intel since they couldn't be sure AMD could pick up the slack.
What did happen is that management decided there SHOULD BE such cross-engineering ,which meant we had to stop hand-crafting our CPU designs and switch to an SoC design style. This results in giving up a lot of performance, chip area, and efficiency. The reason DEC Alphas were always much faster than anything else is they designed each transistor by hand. Intel and AMD had always done so at least for the critical parts of the chip. That changed before I left - they started to rely on synthesis tools, automatic place and route tools, etc. I had been in charge of our design flow in the years before I left, and I had tested these tools by asking the companies who sold them to design blocks (adders, multipliers, etc.) using their tools. I let them take as long as they wanted. They always came back to me with designs that were 20% bigger, and 20% slower than our hand-crafted designs, and which suffered from electromigration and other problems.
That is now how AMD designs chips. I'm sure it will turn out well for them [/sarcasm]
I'm not sure why people don't understand the difference between hopes and predictions. I've never said I don't root for those guys. All I've said is I predict failure. Every year I root for the NY Mets, but I know that chances are they are going to blow it. In this case I look at past performance, take into account things that I personally saw happen, take into account what my friends who are still there tell me, take into account how much money AMD has in the bank, and I make a prediction. It has nothing to do with me wishing them ill will or hoping that things go well for them.
BTW, you ask how AMD could have competed? Well, for one thing, the could have leveraged K8 and the K8 team's success and design techniques instead of wasting years of time on a project that eventually got cancelled using people that had never achieved any success. It took Intel years to come out with Nehalem, and AMD could have been so far ahead by that point that they'd have enough money in the bank that they wouldn't have to accept a low-ball settlement offer in the antitrust suit and they wouldn't have to sell off their fabs.
The source is me. I know the average ASP required for a profit (as does every employee of CMD and TMD - it was a regularly repeated figure that guided everything we tried to do).
As to your other comment about me, AMD does have three decent employees. They are waiting for their bonuses to kick in, and then unless they get new lock-ins that will be all she wrote.