Friday, May 18, 2007

Barcelona schedule in flux

It seems like the Barcelona schedule may again be pushed out to August or September with bugs still being uncovered:

AMD Barcelona launch schedule up in the air

This is not good news for AMD. And if this happens then the confidence of not just the ODM guys in Taiwan and China but also the OEM manufacturers who are building their server SKU line ups for 2H will be shaken. Which would result in cancelled orders for AMD and most probably those orders being moved to Intel. Who it appears are doing a bang up job of managing their product launch schedules and have been beating those schedules over the last 9-12 months.

It now seems highly unlikely that Barcelona will have any kind of impact for AMD in 2007 and with Penryn most likely to launch in Q407, it appears they are giving away any window of opportunity they may have had with Barcelona.

Right now, it seems like Intel's tick-tock is the sound of impending doom for AMD.

Thanks to Anonymous and Sheepshagger for the link.

14 comments:

Roborat, Ph.D said...

i suppose this is the REAL reason why we haven't seen Barcelona running. Because it can't?

Anonymous said...

Hey guys, let's give them a couple more years, I mean months.

Anonymous said...

Come on folks - was this not as obvious as the nose on your face?

When AMD is saying "mid-2007" for Barcelona when it is MARCH 2007, that means they don't have a clue and there are issues... Most liberally interpreted mid 2007 could mean Q2/Q3 or Apr-Sept. When it was March 2007 and AMD is still sticking with the "mid-2007" story you know there is another "delay for hard launch spin coming"...

When AMD is saying we "don't do soft launches" and they spec Barcelona at up to 2.6Ghz and then say only 2.3GhZ at launch (whenever the hell that ends up being), you now there are problems...

Expect continued FUD around all future AMD releases/announcements with 6 month launch windows with the appropriate CYA in the schedule...(fusion anyone? 45nm?)

Now take the next logical step - AMD says customers don't buy 45nm, what might that tell you about their 45mn process? Are they hedging their bets so to speak - they had previously been making a big PR deal about "catching up" to Intel, now it's not that important?
(I use the ctaching up phrase loosely as AMD seems to think catching up on schedule is the only thing their is to match process technology)

I wish AMD would focus on one area and get it right (product, timing, pricing) - trying to match Intel across all areas seems to be Napoleanic. It seems like if they focused on server or mobile they could give Intel a run for its money...

Anonymous said...

I feel really bad for AMD employees who are being kept in the dark about Barcelona. Only a few employees really know the real story which is that there will be no chance in hell that Barcelona, the quad core version, the one that is supposed to have high margins for the company, will be DOA because of what Harpertown will do. This is why you see AMD rhetoric in Barcelona beating Clovertown but you don't hear squat about Harpertown which is the 45nm version. AMD are trying their damnest to not be shot at Computex: the rumor is that Intel will let the general press bench Harpertown at Computex whereas AMD will not take that approach.

I am sad for the hard working employees at AMD because they are the ones tht are going to ultimately suffer the same fate that the Intel employees went through last year. when Intel chopped it's 10,000 people it did so with a winning product. Here the situation is different for AMD.

180 Sharikou said...

Anonymous - if what you say is indeed true, about AMD still not allowing Barcelona benchmarks by the tech press at Computex and Intel permitting Harpertown then AMD is toast. Everyone and their uncle is expecting AMD to use Computex to put the spotlight on Barcelona. However, if that doesn't happen and the schedule does get pushed, I will probably have to re-consider whether I'm being conservative in projecting another 500 million USD of losses only over the next 2 quarters.

Anonymous said...

180 sharikou,

Please note that I didn't say that AMD would not try their best at putting their best foot forth at Computex: What I said was allowing the general press to up and bench machines...they will be controlling quite a lot to avert more pad PR that what they already have with respect to Barcelona.

AMD is expected to have a cherry picked setup on the desktop side; the server side is up in the air especially any cores clocked and running _stable benches_ at 2.6GHz or above, which is what AMD needs at a minimum to compete with Harpertown.

Anonymous said...

Speaking of Harpertown

http://www.fudzilla.com/index.php?option=com_content&task=view&id=1054&Itemid=35

Anonymous said...

"However, if that doesn't happen and the schedule does get pushed, I will probably have to re-consider whether I'm being conservative in projecting another 500 million USD of losses only over the next 2 quarters."

Barcelona will HAVE NOTHING TO DO with AMD's profitability over the next 2 quarters...

Henri Richards (or Ruiz?) acknowledged this in an interview (and didn't you blog on this?). Barcelona will not materially impact AMD's revenue/profitability until H1'08. While their stock price might suffer if things aren't demo'd at Computex, I don't see how you can argue revenue and/or profitability in the next 2 quarters will be seriously impacted by a product that even if it is on schedule will account for <5% of total shipments.

What will drive AMD revenue / profitability for the rest of the year is:
A) overall CPU market growth
B) impact of Intel's upcoming price cuts (across all segments)
C) AMD's ability to more accurately forecast product demand (getting the right product mix)
D) Mobile sector - AMD has been making inroads here and this really is what has kept them going. If Intel cuts back into this share than ASP's will drop further
E) Ability to win back business in the channel (better margin than Dell)
F) Ability of graphics segment to either lose less money or get back into the black.

180 Sharikou said...

Anonymous - to clarify, what I meant was allowing the tech press to do their own benchmarks as opposed to doing sealed benchmarks. Because if AMD doesn't allow the tech press to touch the systems or give them samples for their own benches by Computex, there will be accelerated scepticism around the schedule.

On the profitability issue. In my post I said that if there is schedule doubt, this might result in cancelled orders for AMD. Why - because the OEMs/channel have to trust their vendor to deliver the product they need when they need it. If AMD has committed Barcelona server parts in X month and that slips, it ruins the OEM's SKU line up plans. Which means they have to re-plan their SKUs and guess what...they aren't going to hand those SKUs over to AMD easily. Either AMD will have to discount Opteron heavily, or those orders will go to Intel. This also has a potential ripple effect on the overall relationship and the OEMs will use this to leverage further discounting from AMD into as many of their orders as they can.

So while all your points are valid, there is a hidden risk here for AMD.

Anonymous said...

http://news.com.com/AMD+shows+sunny+side+of+Barcelona/2100-1006_3-6185526.html?tag=nefd.top

Still no hard performance details and core frequency info.

Anonymous said...

Some reader was able to calculate the frequency of the POVray demo by AMD at 1.9->2.1GHz.

Read below link and the reply by grimreapercomesforamd

http://messages.finance.yahoo.com/Stocks_%28A_to_Z%29/Stocks_A/threadview?m=tm&bn=893&tid=1556782&mid=1556782&tof=3&frt=2

Ho Ho said...

the original povray article said:
"The company ran a demo comparing the performance of two four-socket servers, one using the quad-core Barcelona chip and one using a dual-core Opteron chip"

So was that 4k pixels per second for 4P/16 core Opteron?

Scientia from AMDZone said...

Yes, we've been trying to figure this out on my blog as well. This late in the game a processor can't really slip by one month; it has to slip by one entire production cycle which should be at least 2 months. In other words, if what was supposed to be the B0 stepping had to be redone then you lose at least 2 months.

So, if AMD planned for a June launch then slipping to August would be possible. Although, this does seem to contradict earlier claims that B0 was ready. However, the claims mean nothing until launch. It does look like Barcelona's schedule has slipped and that AMD is trying to compress the desktop and dual core version releases.

Anonymous said...

Interesting to know.