A 10th of FAB 30 only used for A64's ?

[clarkey01]

http://www.overclockers.com/articles1107/


"True, this doesn't include Opterons, but toss them in and you only get about 750,000 Hammers sold. This accounts for roughly a tenth of AMD's production capacity at 130nm (even after figuring in additional die size). "



There either not selling enough, yeilds are crap or they choose not to ? Hmmm take a pick.

 

[bradley]

For some reason this scenario reminds me of the way audiophiles greeted both DVD-A or SACD; they stayed away in great numbers, and opted for the CD instead. SACD and DVD-A introduced lots of new *features* that audiophiles didn't need, like copy protection, multi-channel sound etc., and featured them on a line of fairly weak, but expensive converters. You needed to spend a lot of money on equipment and converters to fully realize the sound quality benefits. But the average person doesn't want to spend lots of money for insignificant gains.

Even the computer enthusiast is asking him/herself, why bother with upgrading to Socket 939, PCI-e, SATA2, or 64-bit computing, when the price is high, and the benefits non-existent until some future undetermined date? Perhaps SLI will draw some people into the current equation. But until Windows x86-64 officially arrives, I bet they won't move many Athlon 64s. In all fairness, AMD met almost all their deadlines, and many moons later Microsoft failed to deliver on their promise.

Btw, I'm hearing rumors that this week AMD will be introducing 90mm chips to replace their higher end S939 130nm chips. These new 90nm chips, or San Diego and Venice, will be 15-20% more power efficient, and SSE3 instructions will be included. So maybe this is what AMD needs to jumpstart sluggish sales, especially since they are about to announce significantly lower profits from last quarter.

 

[Markfw]

That article is 4 months old. They even mention "when the 4 ghz P4 comes out....."

 

[mickles]

Good call.

 

[bradley]

Originally posted by: Markfw900
That article is 4 months old. They even mention "when the 4 ghz P4 comes out....."

Fair enough, but then many projects were promised during the span of that article, including that 4 GHz Pentium 4 - only to fall short. I believe computer technology is in a sort of holding pattern, making many points of that article still relevant; I guess one has to wonder just to what degree.

For instance, Athlon 64 availability (vs. Athlon XP,) at many retailers the last few months has been low, which seems to indicate AMD's trouble with yields at the high-end. Ever notice just how many retailers have a per-customer limit on purchases of Athlon 64 90nm processors? And right now AMD has nothing available beyond 3500+ for Socket 939.

So even if there suddenly was high demand for 90nm parts, who knows if AMD would have the ability to adequately supply it. Although, it does appear AMD has greatly refined their automated material processing and are about to introduce second-generation 90nm parts. (On a side note, AMD is currently building fab 36 in Germany, a transition over to third-generation 65nm wafers -- with dual-core right around the corner.) But have yields improved substantially enough from the last batch? I guess we'll see.

 

[miketheidiot]

Originally posted by: bradley

Originally posted by: Markfw900
That article is 4 months old. They even mention "when the 4 ghz P4 comes out....."

Fair enough, but then many projects were promised during the span of that article, including that 4 GHz Pentium 4 - only to fall short. I believe computer technology is in a sort of holding pattern, making many points of that article still relevant; I guess one has to wonder just to what degree.

For instance, Athlon 64 low availability (vs. Athlon XP,) at retailers the last few months seems to indicate AMD's having trouble with yields at the high-end. Ever notice just how many retailers have a per-customer limit on purchases of Athlon 64 90nm processors? And right now AMD has nothing available beyond 3500+ for Socket 939.

So even if there suddenly was high demand for 90nm parts, who knows if AMD would have the ability to adequately supply it. Although, it does appear AMD has greatly refined their automated material processing and are about to introduce second-generation 90nm parts. (On a side note, AMD is currently building fab 36 in Germany, a transition over to third-generation 65nm wafers -- with dual-core right around the corner.) But have yields improved substantially enough from the last batch? I guess we'll see.

hmmm i had heard their yeilds were wonderful. Most, if not all 90nm a64 processors reach 2.4 (3800) speeds easily. Demand for retail A64 cpus has been tremendous BTW, lots of people made the upgrade in the past few months. Unfortunately they aren't selling (or even present for that matter) at best buy and whereever else people buy computers these day. If amd wants to gain market share against intel they will have to go several performance levels above intel. A little tiny speed advantage over intel isn't going to get people to buy a processor they've never heard of. A huge gap would do it though, at least imo. And with what i've heard of the next run of 90nm, they should be able to do it.

 

[DAPUNISHER]

Hector isn't wnning awards for nothing He already stated some time ago that they would no longer place emphasis on trying to directly compete against Intc for market share *like Sanders did* but would "focus on building a firmer financial footing through higher-value products". AMD has picked up from 8.5 in Q2 to 9.9% of the x86 server market in Q3 even though the average processor price rose *ten per cent quarter on quarter :Q Capturing 10% of the server market is no small feat and given profits are higher there, I'd say Hector has the company properly focused.

AMDs sktA platform is still in demand in the 2nd and 3rd world countries becasue of its strong price to performance ratio. Futhermore, without a 64bit windows OS to help drive 64bit desktop sales, I doubt AMD is suffering from demand far outpacing the supply of 64bit microprocessors yet. Even if they are, it appears their strategy is to maximize profits as oppossed to maximizing production :light: IIRC AMD just over 2yrs ago had so many sktA chips stuck in the channel that they cut back production and pricing to sell through existing inventory, and their stock was nearly bottomed out back then. I think their current situation is preferable, but that is just my perspective of course.


* = According to IDC