OMG! AMD socket 1207 Opterons delayed!

Action_Man.....You trying to steal 9 INCH'S thunder?

Did you color that T-Rex pic. Since most of the computer news suckz today I am moving on to more important matters. Coloring.

Since you didn't realize that it is intended as satire, your making yourself look pretty dumb right now.

i do my own thing for ppl who start threads with OMG, so here u go

OMG!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! not again......................

You do know you sound like a 12 year old girl right?

OMG! That is sooooooo true!

Delay could only mean one thing and it is tweaking or improving the socket design. Things are heating up between the two major semiconductor manufacturer and Intel want to get the crown as AMD own it for the last 3 years. So don't OMG about it!

Hey Action Man what happen to your cool avatar?

I think its clever... Word.

lol, I think it'd be something more like this:

lol, I think it'd be something more like this:




F'n ROFL !

You forgot to make the subject line all in CAPS...at least give someone else the chance to post kb pix...

Oh come ON, they totally wouldn't make blue ones, they'd only make green ones, duh!

Copy exact is relatively expensive with the brunt of the cost front load into the project by nature. JumpingJack, as you've stated, it takes lot's of long term planning, practice and training before a FAB is opened using this method. Since AMD doesn't have money like Intel, it may find it necessary to cut corners where possible to make ends meet. However, after the FAB is up and running using this method, the long term savings associated with a new FAB are realized through reduced thrash and technicians being able to step right into a familiar environment.

Perhaps if AMD had something worth copying, this might be a better candidate method. Ok, that was harsh...I'll chill.

Ya know -- something funny going on in Dresden ....

a) Fab 36 was originally going to startup 65 nm --- pushed out, 65 nm was going to ship in Q2...oops, now Q3...ooops 2H 2006, ooops, Q4...ooops, sample product in Q4 announce in 2007 (this is rumor)....so now it will be late Q4 (12/7/06 as I recall) and they are right on track ---- rrrrrrrrriiiiiiiiiiiight....

b) Fab 36 was suppose to ship revenue 90 nm parts in Q1, a week or so after Q1 ends AMD announce revenue shipment from Fab 36 started in Q1 on plan and with flawless execution, Q1 earnings AMD reiterated this position, then analyst day AMD states Q2 is when 90 nm revenue material shipped. Hmmmmm..... depreciation behavior also suggests they didn't ship in Q1, wwwwwwweeeeeirrrrd.

c) It took at least 4 spins of silicon just to tweak the IMC for DDR2 -- with final release silicon just barely eaking a gain, first sampled material in January generate "massive bug" rumors. Ssssssstttttrrrrrraaaaangggggge.

d) Now rumor has it, and IT IS A RUMOR as it came for Charlie D. that the Socket F will be pushed a month, perhaps longer. As Charlie says "Bad news...."

Doesn't sound like flawless execution to me I think the analyst/investors are starting to pick up on it.



Fab36 is producing 65nm. I would bet the delay is so Dell can get TONS of em while fulflling OEM contracts. Dell may actually get the first K8Ls ahead of Sun and HP, especially since AMD procs are the top three out of five at Newegg

I am being kinda harsh here actually -- but I am also trying to read between the lines.... what AMD says publically and what the data is suggesting are two different things, as I enumerated above -- Fab 36 doesn't appear to be going as well as what I think AMD is leading us to believe, the 4x4 is beginning to look more and more like a last minute scramble, and 65 nm dates are thrown around than a King Kong temper tantrum.

Add on top of that that the leaked 65 nm info is suggesting a lower bin introduction, and AMD has nothing for Conroe until next summer, that is 1 year against a superior product -- in 1 year (2005) AMD went from 16 % to 22% market share, in 1 year they can easily go back.... and if they do this is trouble for them.

Agreed, especially with your statement regarding the ease of going from 22% market share back to 16%. The biggest problem with AMD right now that I see is the fact that they've tasted some success and it's gone to their heads. I understand that they want to monopolize on their current momentum, but they are setting themselves up for failure. Instead of following a more reasonable slow steady growth plan, they are heaping more operating costs upon themselves in the likes of Dresden upgrades, possible ATI purchase, possible Big Apple fab, etc. Granted, some of these moves are speculative, but if they don't materialize, the affect will come in the form of negative perception...and they can't afford too much of that with such a short history of success. Should be interesting to watch though.

I got to go through a CE training session. One of the students made the same observation you did. The Intel instructor kinda went off on him and told how much it costs when things go wrong at fab startups or fab mods. Sorta like, yea it costs alot but it saves big too. And they live and breathe it.

Good think.. I was not ready to buy anyway...

Ya know -- something funny going on in Dresden ....

a) Fab 36 was originally going to startup 65 nm --- pushed out, 65 nm was going to ship in Q2...oops, now Q3...ooops 2H 2006, ooops, Q4...ooops, sample product in Q4 announce in 2007 (this is rumor)....so now it will be late Q4 (12/7/06 as I recall) and they are right on track ---- rrrrrrrrriiiiiiiiiiiight....

b) Fab 36 was suppose to ship revenue 90 nm parts in Q1, a week or so after Q1 ends AMD announce revenue shipment from Fab 36 started in Q1 on plan and with flawless execution, Q1 earnings AMD reiterated this position, then analyst day AMD states Q2 is when 90 nm revenue material shipped. Hmmmmm..... depreciation behavior also suggests they didn't ship in Q1, wwwwwwweeeeeirrrrd.

c) It took at least 4 spins of silicon just to tweak the IMC for DDR2 -- with final release silicon just barely eaking a gain, first sampled material in January generate "massive bug" rumors. Ssssssstttttrrrrrraaaaangggggge.

d) Now rumor has it, and IT IS A RUMOR as it came for Charlie D. that the Socket F will be pushed a month, perhaps longer. As Charlie says "Bad news...."

Doesn't sound like flawless execution to me I think the analyst/investors are starting to pick up on it.



Fab36 is producing 65nm. I would bet the delay is so Dell can get TONS of em while fulflling OEM contracts. Dell may actually get the first K8Ls ahead of Sun and HP, especially since AMD procs are the top three out of five at Newegg

Yeah right....next I bet you have great beach front property in Albuquerque to sell....

If AMD could do 65 nm today, we would have it ...


Look at what rushing DDR2 would have gotten them. Everyone craves perf so why not wait until it's ready.....look at Vista. Intel feel into the "too many" competing chip families trap and obviously AMD doesn't want to do that. Intel's ASP (average selling price) is in the crapper (Google --- Intel ASP) and AMD knows that if they have a chip that's too fast iyt will make it hard to get rid of current inventory so an evolutionary path is better in a volatile market. Intel of course has the advantage of being able to lose 7 billion dollars - not that that's a good thing - and AMD is not so they have to introduce products carefully IMO.

As you and others have said FABs cost 2-3 billion - masks for new chips cost millions - so it's not like you can just call your local banker for a small loan to change over to new transistor sizes in one quarter while still getting rid of your "older" inventory.

AMD is now probably doing what the Inquirer said they were doing months ago - piling up chips for Dell. You don't think they will b pushing low end out of Chartered/FAB 38 and high end out of FAB 36 but want to pile up a little of the new stock before discontinuing old stock?

I am being kinda harsh here actually -- but I am also trying to read between the lines.... what AMD says publically and what the data is suggesting are two different things, as I enumerated above -- Fab 36 doesn't appear to be going as well as what I think AMD is leading us to believe, the 4x4 is beginning to look more and more like a last minute scramble, and 65 nm dates are thrown around than a King Kong temper tantrum.

Add on top of that that the leaked 65 nm info is suggesting a lower bin introduction, and AMD has nothing for Conroe until next summer, that is 1 year against a superior product -- in 1 year (2005) AMD went from 16 % to 22% market share, in 1 year they can easily go back.... and if they do this is trouble for them.

Agreed, especially with your statement regarding the ease of going from 22% market share back to 16%. The biggest problem with AMD right now that I see is the fact that they've tasted some success and it's gone to their heads. I understand that they want to monopolize on their current momentum, but they are setting themselves up for failure. Instead of following a more reasonable slow steady growth plan, they are heaping more operating costs upon themselves in the likes of Dresden upgrades, possible ATI purchase, possible Big Apple fab, etc. Granted, some of these moves are speculative, but if they don't materialize, the affect will come in the form of negative perception...and they can't afford too much of that with such a short history of success. Should be interesting to watch though.


In short APM (Automated Process Management) - AMDs brand of FAB control is worth the time. It beat Intel for 2-3 years.

Ok, what ever you say.... I don't have the energy to argue

I'm feeling argumentative (shocked, aren't you?) so I'll make this quick comment: If AMD is stockpiling 65nm, then I'd expect them to have released engineering samples to reviewers already. They wouldn't want to give up a bunch of sales to Conroe. And don't doubt for a nanosecond that if a 65nm AM2 product rumored to be competitive with Conroe was available at about the same time, it would hurt Conroe sales. The engineering sample benchmarks would only need to be in the ballpark.

It is simply really, going 65 nm automatically doubles your output and improves your performance, and lowers your costs. AMD would but utterly stupid not to be selling 65 nm if they did in fact have it running. The finanical incentive alone is simply too much.

We know this is not the case, because Hector himself said in the April 5 announcement paraphrased of course, "We are targetting Q2 2006 to install fab kits to BEGIN printing 65 nm die...", that's likely happening right now...then there is debug, yield, etc.

Add on top of this that AMD still needs to squeeze another 10-15% in performance to meet ITRS based on IEDM 2005 and they still have some work to do.

Jack

Word.


Look at what will happen when COnroe comes out. Intel won't be able to give away P4/PD systems. AMD is taking an evolutionary approach. It takes 3-6 months to qualify and engineer systems. If they want K8L out to OEMs in Dec, they have to be at least at the first spin.

SInce when has AMD EVER been forthcoming with info on unreleased products. I figure we'll start hearing about 65nm AND K8L samples within the next 2 months.

I mean no one believed when I said their may be a socket added for coProcs or extra procs for high end desktops/workstations then a few weeks later AMD intros 4X4. And now ATi has the opportunity to make X1600s to fit in a coProc socket or HTX slot for physics. nVidia ought to make some noise about it soon too. They already have flip chips.

Anyway, we'll all see soon enough. I'm not being paid for this so it doesn't matter THAT much but obviously you don't understand current stock vs. new stock(SKUs). AMD should not REACT to Intel they should act on the momentum Opteron/X2 has given them. That means stick to the schedule. That maybe why they have delayed shipments of high end AM2, so that a new rev could be in place. Besides the current Opteron is still owning the server market in perf/watt so they can afford to be "late." I'm actually surprised that they haven't qualified for 1066 DDR2. That could be the delay also. Again we'll all see soon enough.

Thx for listening