Slashdot Mirror
Is AMD Dead Yet?
TheProcess writes "Back in February 2003, IBM predicted that AMD would be dead in 5 years (original article here), with IBM and Intel the only remaining players in the chip market. Well, 5 years have passed and AMD is still alive. However, its finances and stock price have taken a serious beating over the last year. AMD was once a darling in this community — the plucky, up-and-coming challenger to the Intel behemoth. Will AMD still be here in 5 years? Can they pose a credible competitive threat to Intel's dominance? Do they still have superior but unappreciated technology? Or are they finally old hat? Can they really recover?"
When AMD came out with low priced CPUs that were highly overclockable and great performance at stock they became *the* CPU for any serious geek. When they changed their mind and decided to price-match Intel causing massive price increases they alienated their primary sales force. Geeks selling to family & friends was a great system and without that AMD has been hurting. It's possible they would have died anyway sticking to the cheap, but they've never made a sufficient argument to their customers of why they can't keep the prices low like in the past without letting it on that they like all big business care more about short term cash than long term relationships.
And they went with trusted computing. That was the last straw for me. I would have continued to buy from them despite the flaws listed in parent post, if they just continued to build computers that I could trust.
I hope that AMD soon becomes the darling of the community once again, it's because of them that I recently got back into PC gaming. I had totally given up on gaming on the PC, I had bought a gen1 X2 4200 and AM2 motherboard right before the Core 2 Duos came out and I was cursing my bad luck ever since - until I realized that the real holdup in my system wasn't the processor, but my aging 6600GT. In fact, even though I had bought my AM2-based system almost 2 years ago (or longer! I can't remember when the platform launched) I still had a fairly recent system that could actually support even the newest AMD chips. The real kicker came when I bought my Ati Radeon HD3850. This thing, in my oppinion, should be getting just as much press as the 8800GT. For someone like me, spending $180 on a graphics card is a whole lot more reasonable than spending $250+ on an 8800GT just for performance gains in games like Crysis. My housemate dropped over $1000 on a new Intel Quad-core based rig with an 8800GT in it and my system keeps pace with his very well under almost all scenarios. There is a difference, sure, but considering my entire rig probably cost less than $500 (exluding monitor), I'd say I'm doing pretty well. AMD is doing a great job at catering to people like me who were about to be console-only gamers because keeping up to date on the PC side was getting expensive. AMD offers an affordable upgrade path at a lower performance point - but it's good enough to make my Xbox 360 jealous! I'm proud to say that I'm still an AMD fan. Will an X2 5000+ Black Edition beat a comparably clocked Core 2 Duo? No! But look at the price! I'd say the price to performance ratio is way up there!
What they originally wanted to do was merge with nVidia, it made sense at the time because nVidia was producing the best chipset for AMD CPUs. Anyway the communications between the 2 companies went sour, so AMD, still hot to do something picked the number 2 choice, ATI.
Now a merger between nVidia and AMD would have produced a powerful company. nVida has 3DFX tech, Telsa, chipsets and the 2 companies had already done a lot of joint work on the original X-BOX design (intel was a late entry). AMD brought CPU tech, flash and some other tech into the mix. However it was not meant to be.
So buying ATI was just a plan B, and not really optimal.
The Intel Core architechture is impressive. It's powerful enough over the Athlon that they can take shortcuts. Gives them more headroom for later, whereas the Athlon is reaching its maximum efficiency of instructions per clock so they have to be more thoughtful with their engineering.
Australian running a company that does C# / C++ / Java / SQL / Python / Mathematica
In the past, AMD had an architecture advantage over Intel and Intel had a slight process technology advantage.
Now the situation is different:
-Since the introduction of the Core 2 Duo Intel has the better architecture (minus memory controller though).
-Intel is smoking the rest of the industry with 45nm high-k/metal gate in therms of process technology. Compared to what has been published by IBM about their hkmg technology IBM/AMD has a long way to go to catch up.
And let me say this: Intels technology is extremely clever, they did one fundamentally different thing (gate first) against conventional wisdom which took them onto an entirely different path. Getting the fundamental flaws out of this approach enables a flurry of additional optimizations that IBM/AMD will not be able to apply in their technology. (full metal gates, not using any exotic materials for the gate)
The only disadvantage for intel could be higher cost/lower yield associated with the hkmg process. However they have the benefit of scale (in therms of volume) on their side. In addition they went go through the painful hkmg transition two years earlier and hence things will be much easier for them at the 32nm node. IBM/AMD will be in even more trouble than they are now. I predict that Intel will have a very quick 32nm ramp around the time IBM/AMD managed to get their 45nm hkmg process to manufacturable yield.
They spent oodles of money to R&D the new quad core architecture to really be a seamless 4 core proc that shared caches etc. Intel just slapped two dual cores together and shipped that. Turns out that in benchmarks for consumer programs, intel's stuff works quite well. AMD's cache sharing and topology of memory access that seems better for true multithreaded applications is irrelevant and occasionally a hinderance when you're running multiple single threaded programs.
When you are designing architectures for 7 or so years out, you need a powerful crystal ball, but no such thing exists. AMD just guessed wrong about the nature of future applications. Intel guessed wrong with the Itanium also. Maybe the common thread is you have to fit existing apps instead of the other way around. But, betting against app change has risk also.
Perhaps AMD should focus on the low end rather than guess what the high-end app technology of the future will look like. This may be a better bet for them because they cannot absorb the kinds of gambles that Intel can, being a smaller company. Thus, if they focus on the low-end, they don't have to predict the future of the high-end apps, reducing their risk. They just have to make existing apps run faster and/or cheaper. This would essentially force Intel to be the pioneer (of app change guessing) and take the arrows so that AMD doesn't have to. Of course there are the arrows of internal technology changes, but at least having to guess what *apps* of the future will be like is out of their court.
Table-ized A.I.
A caveman dreams of being us, the incalculable power and riches. We dream of being Q, then what?
I'll admit I don't know much about the matter, but they seem to be fairly Free Software friendly, in terms of their releasing of documentation for both their CPUs and the ATI GPUs.
Does anyone have any detailed information on this? Perhaps the Free Software community can support AMD's openness by buying AMD hardware, *and letting them know this is the reason*.
It is interesting to note that this article is dated February 17, 2003. In other words IBM made this prediction literally 2 months before AMD introduced their first 64-bit processors, the Opteron, in April 2003. Little did they know the impact the AMD64 architecture would have on the industry (Intel cloned the architecture) and on AMD itself (it helped them stay afloat for the past 5 years).
A caveman dreams of being us, the incalculable power and riches. We dream of being Q, then what?
Excellent points sir. ... ...
I should like to take this opportunity to introduce you to a friend of mine.
The Paragraph break.
For full disclosure, I should add that last I checked was twenty years ago.
full metal gates
Some kind of Microsoft-themed manga?
sic transit gloria mundi
Let us all hope they don't die, I'm almost an Intel fanboy but my god if AMD dies! Intel would rape us all. Competition is always healthy. I think AMD has good low priced CPUs though and they sure do the job.
That good high end technology often gives you a good low end too. That is the current case with Intel's Core technology. You take a Core 2, but instead just make a single core version with less cache and clock it way down. You then have a chip with extremely good performance per watt, and good yields (and thus low price) to boot. The Core Solos, as Intel calls them, are extremely competitive on the low end. They've got ones with a TDP as low as 5.5watts.
So it can be hard to try and just compete on the low end of things, since you can't charge as much, and often the people doing the high end things get killer low end products as a side effect.
This is something companies have found out with graphics cards. There have been a number of companies who have tried to compete with nVidia and ATi in the lower end market. Their idea is that while they don't have the R&D to produce a top flight graphics card, that's ok because most people don't buy one of those anyhow. They'll make midrange and lower end cards and sell those.
Great idea, it seems, until you consider that ATi and nVidia get great midrange cards as a side effect of their high end cards. Graphics cards are highly parallel beasts so all they do to make a lower end card is cut some of the units off, put on less memory, maybe clock it down a bit to improve yields and they are good to go. An 8800 GTX and an 8600 GT are the same beast at heart. The 8600 basically just has 25% the number of shader units the 8800 does, and other things like a smaller memory bus. End result is nVidia has and extremely fast $100 card that cost them very little in terms of R&D that wasn't already done for their high end card.
So the companies that have tried have thus far met with little success. Their offerings just haven't been able to compete with the big boys and it is no surprise. You can pour a lot more in to R&D when you are going to sell graphics cards at $500+ and then make use of that very same technology in midrange and low end cards.
I've used several AMD processors (couple of Durons and now an Athlon X2). I chose them on a value for money basis. I never buy the fastest chips that command a heavy price premium, so the arguments over who has the top chip of the moment are irrelevant to me. I considered an Intel for my current PC, but the price difference was minimal and I know the AMD-based chipsets a bit better so I knew it should work for me. I do like to support the underdog, but not if it exceeds my budget.
Even in the last few years I have met people who consider AMD to be inferior or less reliable than Intel chips. Intel's marketing millions must be doing something for them, but I find their jingle intensely annoying when it crops up in the middle of an ad.
Ironic that IBM isn't manufacturing popular CPU's anymore.
FRAUD ALERT? First, for me this story crossed a line. It looks like stock manipulation. Was KDawson paid to post this story? Who at Slashdot or its parent company has recently sold AMD stock short, betting that the price will fall? Are any Intel employees involved?
... through dumbness rather than
malice."
I would like to see a statement added at the end of this Slashdot story that KDawson took no money for this story, and that no one at Slashdot or its parent company took money or will benefit from a drop in price of AMD stock. I'm not accusing anyone of anything; I am just concerned that this story is worded in a way that seems sleazy and possibly fraudulent to me.
Second, in response to the parent comment. ATI is the premier video CPU provider now. nVidia is so lame that there is an entire web site devoted to fixing nVidia driver issues: LaptopVideo2Go. I spent hours trying to get one of my laptops, which has an nVidia chip, to work correctly with an external monitor. It works well now, but I could never have done the work without the help of LaptopVideo2Go.
Third, Intel is suffering from very bad management. For example, see the comment I posted to an earlier Slashdot story, responding to someone saying, "Intel's behavior regarding the OLPC is reprehensible."
Fourth, AMD seems to be the more technologically dedicated company. Intel has a history of dumb mistakes. For example, see this December 2000 article about the Pentium 4, which calls Intel "Chipzilla": Pentium 4 Linux problem all Chipzilla's fault, apparently. Quote: "Intel... failed
I seem to remember that the entire Pentium 4 architecture was abandoned in favor of the Pentium 4 Mobile architecture, which is what Intel is shipping now.
Both AMD and Intel make VERY sophisticated processors. It's amazing that a product that is so tiny it is affected by quantum physics is cheap enough for everyone to own. When one is temporarily ahead, it is simply silly to say that the other is dying.
Stock prices are often affected by hysteria. This is especially true of prices of technical stocks, which are often owned by people who don't really understand the technology of the company they partly own.
Dell is finally selling PC's with AMD processors right along the Intel offerings.
They finally, now, have the platform.
Not just that - the difference between Intel is hubris vs economics. As nerds, WE have the responsibility to show people where they're wasting their money. If you're shelling out $6000 to get something bleeding-fucking-tomorrow-edge, yes, you want Intel. If you want something you can use for the next 3 years, but not top of the line (which most people don't need), then an AMD chip will cost you less than half as much as an equivalent-powered Intel.
My hope is that AMD continues to grow and gets their chips into lines from a few other commodity manufacturers. The best thing for the consumer would be two companies competing on approximately equal footing.
Intel's failings on Itanium and Netburst were common corporate faults.
When the competitive marketplace isn't driving you, you have to drive yourself. Once that starts to happen, the directions can become bizarre, with Itanium and Netburst being to very good examples.
Itanium: The problem Itanium was designed to handle was cloning. First and foremost, they sewed up the I.P. so that it was not subject to any existing cross-licensing agreements. Second, the architecture was sufficiently different that they were outside of the realm of existing art ahd cross-licensing, so their I.P. was "strong." Notice that I haven't said a word yet about performance, cost, or any of that normal stuff. When mere technical and marketplace concerns are that low in the priority schemes, guess what happens.
Netburst: It seemed like someone in marketing got overly focused on clockspeed as the Ultimate Metric. The rest falls from there.
The reality is that ANY corporate product, will turn to junk without a competitive marketplace to keep it focused on delivering value to customers. Once competition is gone from a specific marketplace, the company will either focus its development budget in other areas where it needs to respond to competition, or it's development will be driven by motivations internal to the company, that are likely irrelevant or even negative to customers
The living have better things to do than to continue hating the dead.