Slashdot Mirror
AMD Announces A Shift In Focus From PC Processors
slughead writes "Forbes Magazine is reporting that AMD will no longer compete with Intel to make faster, smaller, and more efficient processors. Just as Mac users would be worse off if Windows didn't exist, Intel users will be much worse now that AMD will no longer compete. You see, there's this thing called demand, and when there are no competing products in a market, a good or service will always increase the price to the economic equilibrium, unless forced not to by the state (forget that right now, communists!!). In English: you're going to get less new technology, and higher prices on existing technology." On the other hand, AMD is definitely not exiting the chip business -- they're just trying to branch out from chips for microcomputers.
I'm an AMD fan... their processors often offer more bang for the buck compared to intel.
They do run hotter, but so what? (and how else will I heat my server room in the wintertime?)
competition is always good; free markets demand it, and consumers will suffer when choice is reduced.
Does anyone know some more specifics? C'mon you AMD employees out there... I know you read slashdot... Please tell me this is some kind of sick joke.
Even if a man chops off your hand with a sword, you still have two nice, sharp bones to stick in his eyes.
ya gotta wonder with all the rumors of Apple sending out AMD boxes running OS X if all the rumors were wrong to conclude x86 was involved at all....
Maybe AMD is 'branching out' by manufacturing PPC chips for Apple. No evidence is conclusive but this will definitely add fuel to the rumor fire around the AMD/Apple connection.
A fool throws a stone into a well and a thousand sages can not remove it.
AMD ... said that making semiconductors smaller, cheaper and faster was no longer the key for an effective strategy.
I'll give you a hint, processors fall under the "semiconductors" category. So unless you think they plan to continue an ineffective strategy, it seems clear that they will cease competing.
It doesn't say they'll stop making processors, just that smaller, faster, and cheaper is no longer the goal. Instead, they intend to focus on what their customers actually need.
*buzzer sound* Wrong. Both IBM and Motorola make PPC processors. From what I learned in school, that is two vendors, not one. But you probably knew that, which is why you trolled as AC.
If you hadn't been paying attention, memory bandwidth has gotten a huge shot in the arm lately. Graphics cards are up in the 20GB/sec and in a couple of months, main memory will be at 6.4 GB/sec (dual channel DDR-400). That's a *huge* jump in memory bandwidth over what was the case only a year ago. No, CPU speeds definately need to go up, because as of late, memory bandwidth has been keeping pace. As for why home users need 3.0 GHz, I can rattle off a list of several things:
1) Not everyone uses just Word. A large chunk of the population actually does demanding work on their PCs, and this includes video/audio editing, 3D, programming, scientific computing, desktop publishing, etc.
2) Gaming!
3) Natural language processing/artificial intelligence.
4) Windows Longhorn and KDE 3.1!
A deep unwavering belief is a sure sign you're missing something...
The reason so many x86 based servers run Intel CPU's is that Intel has managed to keep AMD out of the major PC manufacturer's server lines. Take any server oriented AMD based system (w/ a Mobo from, say, Tyan or Serverworks) and compare its stability with the stability of a comparable Intel system. Result: both tend to be rock solid. I worked in a testing lab for a maker of RAID systems and storage contollers. We rarely saw a difference in overall stability between AMD and Intel based server platforms.
Again, verbatim from my site:
Attention Slashdoters:
Don't get all your news from Slashdot
I should really start out by saying that I have never owned and Intel chip, my PC is an AthlonXP. I'm about to attempt to put an end to any dispute, and I'm going to use Forbes as my main source (because I think the man might know a thing or two about money, because he's got a lot more than me).
After the PC market's recent downturn, AMD was facing tough times. However, unlike Intel, AMD has little diversity in chip manufacturing. Recently, Intel announced their 3ghz processor with hyperthreading (a way of fooling software into running more efficiently). So AMD cut 15% of their employees shortly after.
Not that anyone needs to be reminded, but when a company cuts jobs, it doesn't just mean that they will have less employees, it also means that they will pay taxes. In addition to their current financial problems, just 5 days ago, AMD converted $300m of debt into stock, which will hurt their economic standing in the future, and by extension, the present (the news left AMD's stock in shambles at $5.90).
Converting debt into stock on such a large scale has consequences. Like, for instance, S & P could cut your credit rating. Of course, when the S & P does that, you have to convert more debt into stock.
Just a bad time for PC CPU mfg.ers? Well Intel's doing great, so how about that?
So whether the article said it directly or not (I read it 5 times and I think it pretty well did), AMD will not be competing in the PC market for a while. Their 64bit chip might help bridge the gap between 32bit and 64, but it has to come out first, and then it has to beat Intel's benchmarks on 32bit applications (which I could presume it will not). It will beat it in UT2003 though, hopefully that'll be enough!
Latewire
Also, I don't think it's fair to assume Intel's primary domain is purely in PC Chips. Their work in Communications chips is nearly as important, and is becoming more so. I think they're trying to develop it into a stronger market force in the next few years.
All circuits busy.
Didn't AMD just mention that it planned to surpass Intel as the world's number 1 chip maker?
And I quote:
AMD BOSS HECTOR RUIZ says AMD is "dead serious" about ousting Intel to become the number one player in the "computational processor market".
"We're not just trying to be a good number two," he said.
Ruiz claimed its "competitor" had done "everything possible" to keep it from competing in new segments of the market but, despite Intel's best efforts, AMD was on course to make significant progress in a number of areas.
Surely AMD didn't change its entire business direction and core corporate strategy in a matter of days. It seems to me that there is a misunderstanding, and seeing as how the Forbes article quoted not one single comment from AMD brass stating that they "will no longer compete with Intel", I think it's Forbes, and the story submitter.
I seem to recall rumors back when AMD was kicking ass that Intel planned to leave the PC CPU business to pursue more "long term profitable measures." Well, what sure doesn't seem to be the case.
Computer Science is no more about computers than astronomy is about telescopes. --E. W. Dijkstra
How is it that this forbs article got an entilrely different spin on what AMD is doing from everyine alse who reported on the same event and was sitting in the same rooms for the same lecture?
Again here is the headline. I will decode some of what may not be obvious.
1) The first thing is the title which you could maybe read to mean what the slashdot article makes this story out to be, but technically they dont say that AMD is moving out of PC processors or anything remotely like that.
2) The story actually comes from rueters, not forbes, so they just picked a meaty article to post on their site, not necessiarily paying much attention to any possible fud or slant. Again not the most responsible, but not as bad as the slashdot article.
3) Dont know who this guy is, but probalby a reuters hack.
4) "LAS VEGAS, Nov 19 (Reuters)" - this was from the same stuff on the same day as all of the other Ruiz (AMD CEO) articles, but this is the most creatively different article (besides slashdot) that I have read so far.
5) The leading paragraph is true as is technically the rest of it.
6) The other two paragraphs I show here, when mixed with the title, is where people might begin to get the wrong idea and where I think they could have done a better job. I dare say that this is the part where it could seem slanted.
7) The last sentence in the excerpt that I show above is the "hook" for the article. Its where an actual statement is made as proof of the "moving away" title. It is this sentince that makes it sound like AMD is not doing PC things at all for these guys.
I'll try to counter a few of these:
- Cray is going to be made of commodity PC processors. Many/most of the super computers these days are made with commodity procesors (and many of the other parts are commodity as well).
- The Lucus pre-rendering was done on PCs and there was never a mention of *not* user PCs for any other work. Here is a real quote form a real reporter: "Using the Athelon processor, JAK Films was able to help realize George Lucas' vision of his fantastic world," Ruiz explained after a brief onstage chat with Star Wars robot character R2D2. http://www.showgo.com/storytest.cfm?story_id=2054
- And as for gibson: It features a "hexaphonic" pick-up that transfers the analog sounds of the strings to a digital format. The signal is then moved via Ethernet to an AMD Athlon 64-bit digital audio workstation (DAW), where it can be amplified, modified or blended with other musical instruments. The technology will allow musicians to collaborate and jam with others even if they're in different cities around the world.
http://www.showgo.com/storytest.cfm?story_id=2054 (same as other link).
Thanks for reading
Yes--I agree that AMD is making a smart move. The only segments of the PC market that are going anywhere are the budget PC market (think under US$500), the mobile market (notebooks, tablet PCs, etc) and new form factors and "non-pc" consumer devices(ultra-compact desktops, set-top boxes like PVRs and game consoles). These all have a few common elements: the edge goes to the one who can produce low-cost, low-power, low-heat designs. The high-end desktop/workstation market is flat at best (and most likely in decline). Few in the market for a new PC are geeks who want to compile their Linux kernels faster than all their friends or render sophisticated 3D videogames at framerates higher than their eyes can perceive them (or most monitors can display them).
BUT---hold on a minute. Who said AMD was just going cheap. I got the idea that they were diversifying their market strategy. Until now, their marketing has been very unsophisticated to be polite. Hell, in an effort to win a petty pissing match over who is the king of the crap-pile they have gone so far as to give their processors "model numbers" instead of labelling them by clock speed. Joe blow on the street gets very confused when you try to explain what the 2100+ means on his "Athlon XP 2100+" (no it isn't REALLY a 2.1 GHz chip--it's only 1.733GHz but it's just as good...huh?). time to shape up.
Now, AMD is promising to get a bit smarter in selling their bleeding edge technology. Why languish in a stagnant high-end PC market with razor-thin profits against a giant comptetitor? Instead of trying to find, retain and support thousands (millions?) of geeks and "keep-up-with-the-joneses" types, each with 1 to maybe 8 processors each, perhaps it would be better for business to land a deal with hundreds of customres like Cray, or NASA, or JPL, or governments and big corporations, who each need thousands of processors to meet their needs?
Big customers don't need their hands held--they order in volume, making production runs easier to manage and cheaper and margins higher. They engineer and support their own products. These big guns are also much better at showcasing the latest technology. Think about it. What would be more impressive to the average person (and the mainstream press):
* If AMD powered supercomputers rendered the latest CG animated hit movie, an AMD-powered cluster of servers made Google search 500% faster, or predicted storm movements on Jupiter or repeatedly beat Kasparov at chess...etc
* AMD scored higher on some obscure benchmark on Andtech and Tom's Hardware, the pizza-faced kid next door got his 3D gam to go 120FPS at 32-bits and some insanely high resolution (ultimetely more a testament to his video card's performance anyways), and the computer salesman made and Athlon XP system boot faster than a similar but higher priced P4 system
I'd say the former would garner more respect and a higher profile than the latter.
Shouldn't that be "Windows users would be worse off if the Mac didn't exist"?
My first PC was 386 SX 16. It had Intel CPU of course. Then I bought 386 DX 40, then AMD 486 DX (DX4? something like that) 100, then Cyrix 6x86 166+, then K6-2 500, then Athlon XP 1800. As you see I haven't used Intel for a long time.
If AMD is going to increase prices and stop active development - there is a place for a new competitor. You probably don't remember Cyrix CPUs from pentium times. Thanks to Cyrix - Intel and AMD put new CPUs faster on market. Of course they also created very aggresive marketing (I still hear "Cyrix is unstable" sentence). If AMD stop now - whole market will slow down.
We need Transmeta in Desktop PC. Is it possible?
what about Crusoe? what's the status of the Crusoe processor and why don't they take advantage of this opportunity and jump into the market?
"The ones who dont do anything are always the ones who try to pull you down" -- Henry Rollins
The submitter may have been patronizing in tone, but the spirit of his/her message is more or less accurate given the barriers to entry in chip fab market (plant/equipment + design = BIG$$). The problem with your statement is that I am not sure that the market will create this "more aggressive supplier" -- unless those rumors about the ultra cheap and fast Chinese chips are true, but in that case we are dealing with a state created (or at least heavily subsidized) chip producer and not one created by the competitive market.
Bah. I'm a Mac guy. Where's my IBM 970?
Here is the actually press release from AMD. Theres nothing in there that says they are stopping consumer chips (infact, they talk about the 64bit chip, and unreal). they do mention that they are 'branching out' Microsoft branched out from Operating systems (everything from keyboards, to chairs, to crappy networking hardware) but they still make the same great os!
I guess that people just miss the point that nowhere in the linked story does AMD ever say they're going to stop producing PC chips, it says that they are seeking to diversify. This is a good thing. It certainly doesn't say that they aren't bailing on the market, but it's pretty obvious that they're not going to bail on the PC market so soon after developing the Athlon XP and the Hammer.
Then again, it's 5am, I could be reading a story about chickens taking over New Zealand and not about AMD market share.
If there is a God, you are an authorized representative. - Kurt Vonnegut Jr.
If I remember correctly, Intel just released its 3.0 GHz chips. If I were AMD, and I wanted Intel to slow down a bit, I'd want them comfortable and happy in their current position so they didn't have quite the same sense of urgency about developing newer, better, faster technology.
;)
In fact, I might just start off with such a press release. I might continue by quietly starting up subsidiary firms, owned by the AMD corporate head office, and moving people and tech over to those companies, while making a big show about how the AMD CPU processor focus is being back-burnered, production factories are being retooled for different things, the corporate vision is changing (frequently), etc... All the signs of a firm that is visionary, bold, courageous - in other words, about to show up at fuckedcompany.com.
And when Intel was sitting pretty, regaining market dominance and feeling pretty good about its position vs. "former competitor" AMD, AND the market is starting to boom and demand is increasing, AMD could release something that blows Intel's doors right off.
Yes, it's risky. Long term profit in the face of a short/medium bear market usually is.
With Transmeta still trying to push into the PC CPU market I'd think AMD backing off a little would be good news for them. Two names is about all most customers can keep track of. This will let the Transmeta name have a chance at becoming known outside the geek society. If the recent news of Transmeta's new much faster and even more effecient CPU is true and that CPU is cheap and faster (and more energy/heat effecient) than a 1.8Ghz Intel CPU then they might grab some decent market share.
Just to make a wild prediction I'd say handheld wireless devices will be a big boom over the next decade or so.. faster CPU's probably will matter less than extending battery life. If they can make them cheap enough and so that they don't need massive cooling then they also should work well for parallel designs.. for the power users.
At what price learning? At what cost wisdom? The price is a man's peace of mind, and the cost is his life.
That's the funny thing about personal computers - they usually only need maximum power for particular applications, and even then only for so long. The rest of the time, they're sitting idle. So people *do* need faster processors, but only for an hour or two every day while they do something resource-intensive.
This is why I find the Playstation 3 cell processor idea amazing. But I think it's being used in the wrong area. The thing with console games is that they adhere to a strict set of system requirements, 'cause it's only one system the game is running on, and it's the same system every time. So, I would assume that every developer would try and max out the available resources at any given moment. I wouldn't think there would be lots of free resources available, but I don't know much about it.
If you applied something like this to the PC market, it might have a chance. Or maybe a different approach to the modern mobile CPUs which only use as much CPU as they need, upping and cutting the power as they see fit.
Now, I see those ideas as possible, but I don't want them to happen. Call me selfish, but I want my resources to stay on my system so that I *always* have them when I need them. Besides, a lot of people run distributed applications that are always running in the background, using up the free resources. I never have any free CPU power to spare, except when a program *I* am running needs it.
Smarter devices instead of raw computing power.
Realtime rendering is sorta nice, but not really in the general interest of computing. Look at a human brain: A single neuron transmits information much slower than a microchip. But as a whole, the system simply wrecks a intel (amd, motorola, etc etc) in computing power.
Its specialized, sure. But in terms of speed, microchips already fulfill 80 - 90 percent of our daily computing demands - the stuff where a mass of neurons is bad at, e.g. all sort of clulating. Difficult, but not real smart stuff.
On the technical side, the current hype is phones, not pc's - think Europa & Asia. And the next big thing will require smartness. Something a single general purpose chip, even in its multi giga hertz form, can't handle. Raw Speed alone is losing it's selling point.
AMD seems to get it.
The only skin on a computer should be porn.
True enough and point well made. Missed that one.
Mind you, the highest min. req. spec. on a game I know of right now (could be slightly out of date here) is UT2003 with a massive graphics card requirement of about 32Mb GeForce level or something like that and a processor of about 733 MHz. Even if this is a bit low on the current scale of things, the top of the range gamers PC is at least 2.3 GHz with a really beefy graphics card and memory sticking out of every oriface. My guess is that it'll take about a year (maybe two) until this level of system is too slow to run most new games. This is including advanced AI and physics models.
I could be a bit biased here since my own computer use tends towards anything new and shiny and processor intensive. But it's against the game manufacturers to make their products SO high spec that only the top of the PC system crop will be able to play it.
But, I'm rambling now...
If you can't think of something nice to say then don't say anything at all. No, REALLY.
Back in the day, AMD was a chipfab. Intel made processors, but most of them were contracted out to independent companies, who could do a better job. Then one day, AMD realized that with their efficient factories, they could make a pretty good 486 clone. Intel sued, and in the end, AMD was forced to [a] stop directly copying Intel, and [b] use a different name.
The rest of the story's history. AMD started making the K5 and K6; then they made the K7, their first processor to not be compatible with Intel-standard mobos (remember when every processor used Socket 7?) And their custom architecture, in the end, almost netted them 25% of the chip market.
Then, for whatever reason, AMD started doing badly. And they said to themselves, maybe competing with Intel isn't such a good idea after all.
I expect AMD to release Barton and the like, simply because they're already developed. I expect them to release Opteron and future x86-64 processors, but only with cache/speed/price configurations designed for servers, because there's still money in people switching from proprietary Unix to Linux. I expect Apple to soon make an announcement that AMD is its new supplier; whether that means AMD buys Motorola's desktop PPC chip business, or whether it simply becomes a fab for Apple/IBM-produced designs, I have no idea. But I don't expect AMD to announce any new desktop x86 processors from now on. If this article means anything, it's that Barton's it.
Those fabs are likely to be foundries for companies that design marketable CPU's. And where will these designs come from? The same places that are currently designing marketable CPU's with the addition of some additional startups right here in the US.
The culture of the far east doesn't have what it takes to forge ahead into brave unknowns and consistently produce risky and excellent designs. That culture excels at honing, or incrementally improving existing designs and products. It's as if the whole culture was run by committee on every scale.
Back to the point, if China builds huge fabs, the most profitable things they'll make are chips designed right here in the US whose companies use the Chinese fabs as outsourced fabrication (foundries). Worry about them? I'm counting on them. Those huge fabs will make the chip startup company a possibility again by dramatically lowering barriers to entry.
Regards,
Ross
But you seem to be neglecting the story that led to this all here with your startup dreams. AMD probably shared your enthusiasm a few years ago. You seem to assume there's this huge market for cutting edge CPU designs, but in fact the huge market is for yesterday's bargain basement cutting edge CPU designs, not tomorrow's glittering hype. 90's designs can now be had cheap and when these existing IP cores are combined with current silicon design techniques (130nm) that the foundries already have, they are more than enough for watching movies, playing MP3s, downloading whatever and sending e-mail. CPUs are going commodity like toilet paper and toothpaste.
In case you haven't noticed, there's already a substantial body of free IP cores at OpenCores.org and the competition among i386 architecture companies with existing designs that never made it into mass production is fierce.