Hammer delayed til 2003! From AMD Conference Call

[KenAF]

In its conference call today, AMD announced the following:

* Mobile markets were very weak.

* Hammer samples are meeting ALL design expectations.

* Clawhammer product announcements in Q1 2003.

* AMD won't be participating in highest performance PC segment this year.


If you didn't listen to the conference call today, a few more details are available over at the AMD board at The Motley Fool.

 

[socketman]

Which hammer... the claw or opteron. Its been known for a while that the Opteron (sledgehammer) wasnt coming until 2003. But claw should be here in 2002. Ive heard nothing to contradict that. In fact, I have heard AMD is doing everything to make damn sure it comes out this year.
Im gonna say someone got some details mixed up.. nothing more.

 

[KenAF]

Clawhammer. AMD was very clear on the point in their conference call for analysts today.

If you want to hear it straight from AMD, listen to the audio webcast right here.

 

[Rectalfier]

Oh man. This is not good news at all. AMD's stock price is going to keep falling untill they get Hammer out and making money. I hope AMD can last that long, they are probably going to lose 200+ million dollars this quarter. This is starting to remind of 3dfx. Rampage was going to be so kick ass, but they didn't last long enough for it, and I lost my money. AMD should have skipped thoroughbred and gone straight to Barton.

 

[Chadder007]

Argh!
Yep, another Barton before Thoroughbad vote here.
Didn't AMD say before too that their Hammers performance wasn't living up to expectations too (probably just the 64 bit performance)?...like at the recent shows ..(computex?)

 

[KenAF]

> Didn't AMD say before too that their Hammers performance wasn't living
> up to expectations too (probably just the 64 bit performance)?...

I don't think so. Everything I've read indicates they believe the Hammer is meeting design expectations. They reiterated today that Clawhammer is meeting "design expectations," whatever that means.

There were several mainboard oems at Computex that indicated Clawhammer would debut at a 3000+ rating rather than 3400+, but this is likely the result of a reduced debut frequency, rather than reduced per clock performance. According to some oems at Computex, Clawhammer will debut with 256Kb L2 cache and 1600MHz rather than 512Kb L2 cache and 2000MHz that was initially expected. The German magazine, c't, did confirm that Clawhammer would feature 256Kb L2 cache in its latest issue, rather than the 512Kb L2 (like Barton) that some had expected. Perhaps with the delay into 2003, AMD will be able to get the clock speed up closer to 2Ghz.

 

[pm]

One thing about AT that I have noticed is that people seem to over-react. In this case the news is primarily that people aren't buying computers lately. It is a problem that Intel announced last week and AMD announced today. It is not a catastrophe, it's merely an extension of a slowdown that people had optimistically hoped was ending. On the plus side, Flash is starting to solidify and AMD, unlike Intel, are not primarily a CPU design company but are closer to 50/50 with memory and CPU's.

As far as Clawhammer, this merely makes official a rumor that has been circulating for a while. It's not like this announcement was entirely unexpected. Historically speaking based on my experience designing CPU's their original schedule was an "if all goes perfect" schedule. Based on their tapeout, it was a stretch goal at best to get it out by Q4. Now it's merely slipped a quarter, but this is not the end of the world. It will sink or swim on it's own merits when it arrives in the marketplace. Don't prejudge it now. Judging shipping frequency or shipping performance based on pre-release parts is a mistake. About all you can judge based on these is how far off they are from release - not the performance of the final design.

I don't mean to cheerlead AMD's designs. I work for Intel and believe firmly that whenever Clawhammer arrives, Intel will have a better solution. But I don't like to see people getting too negative on an announcement that is not that grim. And, yes, I have listened to the conference call.

 

[socketman]

thanks for the link. Dont think I would have found it otherwise.
Initial shipments in q4... customer product shipments in q1-03. I still dont buy it. I think they are gonna have a reasonable volume by december. Perhaps not a lot, but enough... Im not sure how much financial punishment they can take though. AMD can only blame themselves though. T-bred is decent but not great (AMD's fault) and the longer Hammer is delayed, the worse their postion gets. They should have executed Hammer sooner.
Unless Hammer is the the CPU equivolent of oral sex, AMD is gonna have serious probs.

 

[Rectalfier]

I don't think that I am over reacting. It's not only this quarter, AMD is in for a rough ride untill Hammer comes out. I wouldn't be so worried had thoroughbred been a good product. Thoroughbred is garbage, how can it not overclock any better than the palomino core? It's a bloody die shrink, it should run a lot cooler, and clock to much higher speeds. Look at how good the Northwood is.

If Hammer does not ship on time, and does not live up to expectations, AMD is doomed. I am quickly losing faith in AMD. I have already lost half of the money I invested in them, and even more since the huge drop after the conference call.

 

[Mrburns2007]

Doom and Gloom....the sky is falling......Oh wait no the Hammer will rock and I figured it wouldn't be out til March 2003.

 

[christoph83]

Rectalfier your looking at it all wrong. Do you really think 1.6A northwoods are helping intel's bottom line? Not really. The overclockability of the northwood isnt making Dell buy more chips. The Thoroughbred is still a good chip, it may not be the fastest on the market but still OEM's will buy the chips and build computers with it. Basing buying a stock on a superior desktop chip is definatly not the way to go.People thought Intel was doomed when the P4 didnt live up to expectations but still Intel was making billions a year. Its all about demand like pm said. This current quarter has been softer than expected so both intel and amd are feeling it. As long as AMD and Intel continue to make competitive chips that Boxmakers can use, neither company will be doomed. Basically...both companies are feeling the pain of a computer buying slowdown. If you bought intel 3 months ago you'd be down just as much on intel as AMD. This news may hurt AMD a little bit on the balance sheet, but their defintely not doomed.

 

[Barnaby W. Füi]

i could care less, i want cheaper athlon XPs!

 

[Sunner]

Thanks for the(as usual) insightful post Patrick.

As for the people who doubt the performance of K8, remember when the first benches of K7 arrived, on beta silicon?
The benchmark of choice back then was Quake2, the crusher demo in particular, and the K7 performed worse than a K6-III in those benches.

Then once the final product showed up, it devastated the Katmai P-III's.

We'll just have to wait and see if history repeats itself.

 

[Degenerate]

One thing about AT that I have noticed is that people seem to over-react.

Doom and Gloom....the sky is falling......

Hmm..... indeed well said.

 

[Athlon4all]

To be quite honest, this is a relief because it should put to rest all these darn rumors. Not be big deal tho, I've felt for a while that the earliest it would be launched is late december. Turns out I'm rite, now the question is will it be launched about the time the Northwood core was, cause then I mite get a nice BDay present. Christoph well said. To be quite honest, every time I've gone to dell's website this year to look at PCs for somebody, I've looked at whether they use a Northwood in the lower end Systems, and lol, they last I checked were still using the Williamette on every P4 that is slower than 2GHz!!! LOL And now they thrlow the Celeron in there. lol. EDIT:

I don't think that I am over reacting. It's not only this quarter, AMD is in for a rough ride untill Hammer comes out. I wouldn't be so worried had thoroughbred been a good product. Thoroughbred is garbage, how can it not overclock any better than the palomino core? It's a bloody die shrink, it should run a lot cooler, and clock to much higher speeds. Look at how good the Northwood is.

Well, I think you need to compare price for price because price is everything in the OEM market with T-Bred. There are Athlon XP's avaialble for as low as $80 now the pricing is somewhat different for OEMs but anyway, and even at $80, those $80 CPU's destroy the Celeron, and the $100 1800+ totally defeates the $130 1.6A. There is definately money to be made in the value sector and its not totally dead year yet for AMD. Is it a disapooinment yes. But oh well

 

[Jayllo]

Lol,

You guys make it seem like AMD lost is cutting edge on Intel. Pre-Athlon they got their ass kicked by Intel with every K6 release. (K6-III +/-)
I don't think theres a real problem, if they rushed out a flakey chip like Intel did, their x86-64 confidence might go out the window. With a flakey processor release, their goal of getting into the server market goes up in smoke.
Just because you can't clock your T-bred several hundred MhZ doesn't mean its no good. Lastly, I wish Intel and AMD made a true lower power chip that only required a heatsink without the crappiness of a C3.

 

[acejj26]

publicly, AMD has always been saying Q1 2003 for Clawhammer

it's just been our speculation that it would be out earlier.

they're saying Q1 '03 to save their butts. if they get it out beforehand, everyone gets excited, OEM's are happy because they'll have 64 bit athlons for the holidays, blah blah blah. if they don't, and they meet their "public" goal of Q1 '03, then they met their deadline and investors are happy

however, and this is a big however.....if somehow it slips to Q2 '03, AMD will be seriously hurting. investors will lose faith, OEM's will lose faith, and the enthusiast community will lose faith in AMD. they can't let this happen

personally, i see a late october/early november launch with product availability to OEM's in late november/early december.....us little guys will be able to get our paws on 'em early january

 

[Insane3D]

"Hammer delayed til 2003! "


OMG!!! OH NO! AIIIIIEEEEEEEEEEEEEEEE!!!!! NOOOOOO!O!OO!O!!!!

 

[BlvdKing]

Maybe this is AMD's way of telling us that the slump in technology will continue until early 2003. Why should AMD release a groundbreaking, flagship product that will undoubtedly raise ASP's when nobody is buying? I think Christmas 2002 will be very disappointing sales wise for computers; early 2003 sounds just about right for a Hammer release.

 

[KenAF]

> hey're saying Q1 '03 to save their butts. if they get it out beforehand,
> everyone gets excited,
> if they don't, and they meet their "public" goal of Q1 '03, then they
> met their deadline and investors are happy

They can't do that. In a conference call, they can't knowingly lie/mislead like that. If they believed there was a realistic possibility that a Clawhammer announcement was forthcoming later this year, they would be bound by law to say it (rather than indicating that it would occur 1Q 2003).

 

[Rectalfier]

Okay, maybe it's not the end of the world, but it really will only get worse for AMD. Right now they are two speed grades behind Intel, this will only get worse, as the T-bred does not look like it will offer a higher speed grade anytime soon. AMD has a lot of debt, so I hope they can last through this tough period.

 

[Soccerman]

heh I knew it was a bad move for AMD not to add something to the T-Bred core..

hey if they just added a faster FSB like 333mhz then it would have be more competitive, plus they could still take the Barton up to 400mhz if they really needed to and in this case, they might need to, except for the fact that the Barton really adds alot of transistors to the T-Bred core, so whatever they've done to it, it probably will be faster clock for clock anyway. I just wish they didn't have to cripple it with the 266mhz fsb! oh well, at least we can bump it up to 333mhz on todays boards, and if we're lucky, go all the way to 400mhz with the KT400..

another thing they could have (SHOULD HAVE) done for the T-Bred was implimented a heat spreader like the P4 (that's the only reason you don't hear of heating problems with the P4, cause it consumes a hell of a lot of power). that would also allow for naturally higher speeds to be reached. that T-Bred core might not be consuming as much power as the Palomino, but to get the heat out to the heatsink is harder with it, becuase it's a tiny-ass core.

as for the lack of overclocking for the T-Bred, I wouldn't get TOO worried. we've only seen the first couple of weeks of this core!

 

[dullard]

Clawhammer product announcements in Q1 2003.

Every official AMD roadmap has had the Clawhammer release sitting on the Q4 2002/Q1 2003 border. Every AMD comment until this conference call has stated limited shipping starting December and much more starting Q1 2003. There is no surprize at all. Only silly rumors (like the one I saw started just recently of an August release) have said otherwise. To reach the commonly rumored October release, AMD would have to break a speed record in the shortest time from first silicon to release. Given AMDs release date record (including the t-bred) and given the complexity in Hammer it is rediculous to think AMD will do that. Even if AMD could have it available sooner, it is in their interest to test it, retest it, and retest it again. One slip here and they will lose any chance of ever selling to the server market (which is already under the impression that AMD is less reliable).

I hope AMD can last that long...AMD should have skipped thoroughbred and gone straight to Barton

AMD will last until Hammer arrives. They have gone through many years of money losses, one more won't kill them. However if I were a stockholder I would have been quite worried when the 2200+ XP was released since AMD set its $241 price lower than any other AMD top-end-chip release. When they have been losing money right and left, placing a record low price on their top chip won't get them out of the red. I guess AMD is going back to their value image for this year. Value and Barton don't mix at the moment. The extra cache won't help the speed too much, but it will double the T-bred chip size (back to the original Athlon XP size). The Athlon XP 2100+ was released at $420. Thus Barton will have been about $420 - the exact opposite of AMD's value position. As a consumer I'd rather pay $241 than $420 for a very slightly slower processor.

AMD won't be participating in highest performance PC segment this year

That is certainly no surprize. We all knew that the Athlon can't keep up with the P4's speed ramping. It has just about reached its end. AMD has been saying for quite some time that the Athon and its successors (T-bred/Barton) will follow a 21 month speed doubling curve. That is much slower than the typical 18 month Moore's law. This highest performance market statement extends beyond just desktop processors. The Athlon MP 2100+ release today worsens AMD's price/performance ratio over the Intel Xeon (using 1000 lot prices). The 1.8 GHz Xeon and 1800+ Athlon MP both cost $192. The 2.0A GHz Xeon and 2000+ Athlon MP both cost $224. Unfortunately for AMD the 2.2 GHz Xeon and 2100+ Athlon MP both cost $262. The performance hasn't kept up, but the price is just as expensive as Intel - AMD really needed a 2200+ MP chip today that costs $262.

 

[Soccerman]

holy crap, since when could you buy a Xeon for so damn cheap?? they used to be near $900!

 

[mchammer187]

Its not late its just because all the rumors that stated it was being launched in october

i dont think AMD EVER intended on releasing it year and this is right on track with their projections