Slashdot Mirror
K8 Details
Urban Dragon writes "Cnet has a story on how AMD will be giving details of it's K8 chip next week. The K8 will be competing with Intel's Merced chip. It should be interesting to see which comes out first.
" Maybe it won't run as hot, either. I mean, I'll want one regardless, but...
This is just pure speculation, but if the K8 is going to be a 64bit chip it'll need a new instruction set. Isn't the Alpha a pretty damn good 64bit chip with and instruction set that's already supported by Linux, NT 4.0, VMS, and Tru64? Why doesn't AMD either use the same instruction set or start making Alphas (or RISC chips that can also run x86 binaries)? Basically, we have x86, RISC (and many RISC platforms, by the way), and Merced/McKinley. Why are we going to need another instruction set? Maybe AMD will just highly optimize their x86 processors. I'd rather see that than yet another platform.
By the way, wouldn't it be nice if an EV7 motherboard could handle either a K7 or Alpha? We could buy one motherboard and choose between CISC or RISC. Imagine upgrading from an Athlon to an Alpha. There are so many cool things AMD could be doing right now. I just wish we could see some action.
I would wonder how AMD would name it's next gen chip. K9?(I know it's off-topic but....)
Following its architectural triumph with the Athlon chip, Advanced Micro Devices next week will detail the K8, a 64-bit chip that will compete against Intel's Merced.
Very nice, unfortunateley it's not a question of having a good CPU architecture, but to a much larger degree a marketing question. I don't doubt that AMD can design a good K8 chip, but in order to do that, they first have to make the K7 a success. They are pretty strapped for cash and unless they can stop bleeding red ink, they might not even be around long enough to see the introduction of the K8. So they can design decent CPUs. This is good, but hardly news. In the past they had good chip designs falter due to manufacturing problems. Lets hope they can ramp up K7 production fast enough so they have a product to sell.
Having said that with a light undertone of sarcasm, I should probably note that I am/was a satisfied AMD customer. I wish them well because Intel deserves some competition, but they need to be careful not to repeat past mistakes. They have to become profitable soon, which is no easy task when you face a giant like Intel.
I would have liked a much more in-depth discussion of the motherboard support required - can any of you solder-heads out there enlighten me?
Strong data typing is for those with weak minds.
I'm taking it as something of a given that AMD hasn't really gotten a decent break on the Athlon. CNET's article points out that despite the fact that the Athlon chips have regluarly beaten the pants off Pentium IIIs, but they haven't been able to make a comperable dent in the market, largely due to their inability to procure parts. Hardly anyone is making Athlon mobos, which is a damn shame.
That being the case, what are the chances for the K8? They'll be intended for a much more limited market than the Athlons to begin with. I can't see a mobo manufacturer nervous about ponying up resouces to make Athlon stuff being sold on the K8. Throw in the tendency of corporate users to by the safest solution, and AMD may be screwed.
I really hope we're not in for yet another round of the "superior tech can't get an even break" game.
It seems to me that a year ago I was reading how the K6-3 and eventually the K7 would be bring AMD out of the low-end desktop market into the high-end desktop market. They could be *gasp* faster than Intel on the desktop. The K6-3 did that and has been selling well.
Now the line is that the K8 could bring them into the server and multi-processor market. Never quite good enough for the journalists, eh? No one seems to be noting that this company has gone from making 486 clones after the 486 was being fazed out by Intel to creating a chip that was cheaper and faster than Intel's best offering (excluding the Xeon's which are only overpriced Pentium III's with tons of L2 cache).
On top of that, they are selling! AMD beat Intel in retail sales for a quarter! Big guys like Compaq, Gateway, and Intel are selling them in their systems! If you would have told me 2 years ago that AMD would beat Intel in sales and that you could buy one in a Compaq, I would have tried to sell you some nice swamp land in Florida.
As far as AMD bleeding red, look at any company playing catch up or expanding as quickly as AMD and you will always see a trail of red.
The Athlon is testing at 900megahertz and 1000megahertz. AMD has boxed and ready to go large shipmets of 800megahertz Athlons. I'd say that any investment in an Athlon is a safe one. Just think, buy that Athlon 600 Monday when the 700 is released, get a great deal on a processor that outperforms a PIII650 and have a system that you can upgrade to 800 or 900 by the end of the year or early 2000.
One strange thing I've seen with Athlons..
Several companies round here actually have the Athlon chips in stock.. They're advertising them across the web by mail order...
Yet not a one of them advertises a mother board...
As far as I can see, this is going to kill enthusiasm faster than anything else... People going out to buy the chip and board, seeing the chip present, but not being able to do anything with it other than use it as a paperweight...
Womething has to be afoot to keep the motherboard manufacturers at bay like this... It seems AMD have the chips there, but the M/board manufactures are holding back..
I'm still drooling here, and waiting, but feeling more disenchanted as the days pass, and still not board to be had...
So, any motherboard makers out there.. Get into gear guys, there's easy money to be had...
Malk.
Actually if they want to be compatible with the Merced (IA-64), then instead of using Alpha technology they will need to clone HP PA-RISC, which is what the 64 bit part of IA-64 is largely based on. If AMD chooses for the K8 not to be IA-64 compatible, then they may have a lot of problems competing with the IA-64, since it is unlikely that Microsoft will be able to support the K8, and they will still be in direct competition with other 64-bit RISC architectures in the *nix market (UltraSparc, PA-RISC, Alpha, MIPS, etc).
Might have been more interesting if it were the actual detail announcement. Things like number of simultaneous real-world instructions per clock, support for predicated instructions, register counts and windows, demands upon compilers...
Pure clock speed ain't enough, folks. Judging from (dated) info about the IA-64 architecture, there's a lot of nifty stuff that AMD has to at least match if they want to claim any sort of lasting advantage.
Only the dead have seen the end of war.
The slowness of release of Athlon-based systems appears to be related to, surprise, surprise, a dearth of availability of motherboards. I wouldn't want to be accusative of Intel for formenting this, but I'm sure they're very grateful at the inability of AMD to sell massive quantities of Athlon chips...
Every time a new CPU comes out, the real insight comes from looking to the motherboards...
If you're not part of the solution, you're part of the precipitate.
Just a quick comment on that, AMD doesn't have a problem with finance management, at least not any more than your typical company. Their problems largely rooted in the following two factors:
1) AMD's sixth generation processor design was put together decently, but with a very shallow pipeline. This means that with your typical ramping schema, it should be at about the same MHz level as the Cyrix chips (300MHz) or the WinChips (250MHz). As it is, AMD has an immensely aggressive ramping team which has managed to bring AMD's K6 family to just under Intel's P6 family in MHz, which has a couple effects:
(a) Because the K6 family has been historically about two clock bins lower than the P6 family, and because Intel's pricing schema involves tremendous gulfs between the top two clock bins and all below it, AMD's cpu Average Selling Prices could not help but drop lower and lower as time progressed.
(b) Due to the K6's low pipeline and the fab team's uncomparable (and absolutely necessary) aggressiveness, the bin split of the K6 family parts are HORRENDOUS. Before AMD's recent jump to their cs44e7 hybrid process (quarter micron with some 180nm features), the top bin being produced was 475MHz and the bottom bin was still way down at 333MHz or so, with over half the parts still binning below 400MHz. This added more shame to their ASPs, as anything below 400MHz was under a hundred bucks, which means something like only fifty dollars profit per chip, at best.
(c) As a result of the aggressive ramping they needed (to compete with Intel's more easily rampable design), yields were kept lower than comparable Intel parts (though for the most part not horrendous, save for the little "incident" in February). This means that they get lower quantity to sell than they could have gotten otherwise, which means that, in addition to ASPs, they're making very low amounts of revenue.
2) There really is no way to get past problem 1a without making a newer cpu core with a deeper instruction pipeline. And to get past the problem in 1b, while that newer cpu core will help, it'd really be the wiser choice to expand your capacity, so AMD has forced themselves to spend a whopping, Intel-like amount of money (in R&D and in building a whole new megafab) so that, while they hurt in current quarters, they can thrive in future quarters. Would this strategy work? It's not guaranteed, but it's a hell of a lot cooler than the old "play it safe" mentality. If AMD had played it safe and not done all this fab or R&D stuff, then they'd have easily made profits (I believe) off the K6 series in every quarter of 1998 and 1999. The only problem is that they'd be lagging in clock speed at this point and they'd have no real future technology with which to compete. In effect, though they'd be profiting, they would be writing their own tombstone. The way they're doing it now, they've lost lots of money but they *finally* have superior technology to work with. Even without that newer fab, as soon as they ramp K7 to at least 60% capacity, they'd be making a pretty solid profit. With the newer fab, they'll be able to profit very nicely and retroactively fund these projects that they so unharmoniously dumped cash into all these years. They'd also be able to afford their future plans, which is a nice byproduct.
-JC
PC News'n'Links
http://www.jc-news.com/pc
PS: This stuff is largely my opinion, though I believe it to be largely based on fact. It isn't merely a pipe dream that leads me to believe that the K7 is the first design since the 486 that offers everything AMD needs to absolutely thrive in the market.
You are quibbling about details. I am sure that IA-64 goes beyond PA-RISC (otherwise why bother doing anything more than sticking a 32-bit x86 core in for backwards compatibility and calling it good). However, if AMD wants to clone IA-64, they would have to start by basically cloning PA-RISC and adding the same extentions to that architecture as Intel is. My point was that cloning PA-RISC is far more logical starting point for them in building an IA-64 than licensing Alpha would be. I did not mean to imply that it was everything they had to do, nothing is ever quite that simple.
I would agree about 64-bit x86 being an abomination (and I think it would be a marketing flop as well). I don't, however, think that splitting the market with an Alpha variant would be successful for AMD either. If for no other reason than I doubt that AMD could do much differently in getting long-term cooperation from Microsoft for an Alpha derived architecture than Compaq did.
Not only do I think that the abandonment of the desktop and server MPU market by AMD would be disasterous for the industry (competition for Intel has been a very good thing for consumers and system vendors), I think it would be disasterous in the long run for AMD, as it would permanently relegate them to being only a niche player. I think AMD's only option in the long run is to take Intel on directly with an IA-64 clone. There are ways of getting around the patent issues, although they certainly won't be easy and probably not cheap either. There are a lot of features of the current and previous generations of x86 designs that are the subjects of patents, and it hasn't stopped companies from building clones (AMD, Cyrix and IDT).
It may at some point become not so advantageous for Intel to have all of the clone chip builders drop out of the market, as it could start or re-start a move against them on anti-trust grounds. There are other ways that having a certain amount of healthy competition is good for a company too.
A year and a half ago, all we heard was "K7 is going to put Intel out of business. They are going to be for sale signs on all of Intel's fabs and they stock is going to tumble".
A year and a half later AMD losses are at an all time high, K7 hasn't made any dent in sales, and I can't even see one at Best Buy.
This is largely due to the pre-announcement effect: everybody heard about K7 and delays purchases of AMD chips(when they would have bought a K6).
Now they announce the K8 when the K7 is barely ready for production. This is going to have the same effect. Consumers are going to say, "Why should I buy K7 now when K8 is coming out Really Soon Now?" They'll probably just buy Intel.
Willamette/Foster will be out by the time K8 is. What has the world heard about that? Very little. Almost no details have been made public.
Intel never makes the details of a processor public to the industry until it it ready for VOLUME production. Often their published figures are lower than expected, so the compeititors feel comfortable and slack off, then they grab the crown from out of nowhere (P6 is the prime example, and Willlamette/Foster will do the same).
This is the reason Intel is more successful than AMD - they don't preannounce, so they can sell chips they can fab in volume NOW, and not tell customers "don't buy what we have now just wait a little while longer and we'll have this whiz-bang part" (OK, they didn't do this with Merced, but they have with everything else)
The Athlon needs a minimum of a 300W peak PSU
:-) Even my personal machine, a full tower, has a 300W PSU.
Okay, I know that a CPU cannot possibly be drawing that much power. I imagine this specification is due to the fact that AMD expects the Athlon to be in big, hefty machines, with lots of drives, memory, etc. Throw your average (i.e., total crap) PC case at that, and yes, you'll undervolt.
and midi/full towers tend be around 230-250W
Maybe the ones you buy.
High power PSU's are pretty tricky to find as well.
PC Power & Cooling
A/Open (I have the HX-08 case)
dragonhawk@iname.microsoft.com
I do not like Microsoft. Remove them from my email address.
As of Oct 4 price cut Athlon 600s are going to be much cheaper than Intel 600s. Numerous reviews have showed that Athlon is 10 to 40 % faster than Intel's whimpy P3.
Not to mention the fact that the P5-III/600 and 650 lines seem to be having heat and power problems due to Intel rushing them out the door to compete with Athlon.
Reminds me of an old FidoNet tagline...
Pentium = (P)arts (E)xist (N)ow (T)hough (I)nvariably (U)ndergo (M)eltdown
dragonhawk@iname.microsoft.com
I do not like Microsoft. Remove them from my email address.
The Merced has always been thought of by Intel as a way for them to transition into the 64bit RISC market while still keeping the 32bit CISC market open. The McKinely will be the real ass-kicker, while Merced is more of a warm-up bout. I've heard rumours about Intel not even actually producing the Merced and skipping right to the McKinely because everyone who currently uses the IA-32 is well on their way to porting their stuff to IA-64 which would mean backwards compatibility would be awaste of core space. I think skipping Merced would be more beneficial to them than releasing it them a year later come outwith the McKinely. Sure AMD might be able to kick Intel's ass with the K7 for now Intel could easily pop out the McKinely and leave AMD waaaay behind since the K8 wouldn't be ready for production any time close to when Intel could release McKinely.
I'm a loner Dottie, a Rebel.
IA-64 is not PA-RISC 2.0 plus some stuff added on; some aspects of it may be inspired by PA-RISC, but PA-RISC didn't, last I looked, have 128-bit bundles containing 3 instructions plus template bits.
As such, it's not clear to what extent AMD, were they to try to create an IA-64 implementation, would benefit from first doing a PA-RISC implementation, rather than just going straight to IA-64.
And your evidence that this will be the case, rather than that just being a possibility (Intel has, of course, done that in the past), being?
(No, the fact that they currently aren't documenting the stuff needed to do an OS kernel doesn't count; I think they've said that All Will Be Revealed, at least to the extent that the source to ports like the Linux port will be available, by the time they ship.)
Intel have quacked and barked in the past; they may be more likely to quack than bark, but that doesn't prove beyond the shadow of a doubt that they will.
I don't think Intel has any plans to make it easy for anybody other than themselves or HP to build IA-64 implementations; I have no reason to believe that this necessarily means that they'll keep stuff secret forever - I suspect they'll make heavy use of patents, instead.