• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

AMD in Danger

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
You guys know the 4000+ part is a 939 1mb cache chip running at only 2.4? Not that the rating is a problem, it's exactly a FX-53 after all, which blows away a 3.6E in the bechmarks. The problem is 2.4. Where's 2.6? or 2.8? It's getting real late in the 130nm stepping process and I worried for them that's all. We've had a 2.4 part for almost a year already in 53... Intel will strike back.
 
Originally posted by: DrMrLordX
Sandorski. the main reason to produce something that would, as you said, "own" anything available from Intel is to

a). prove that they can do it
b). attract the attention of OEMs

Simple as that. A stable, low-supply, high-end CPU available for benchmarks would do the trick, as long as its clearly labeled as a non-enthusiast part. In other words, if they rest on the FX-53 for too long, they'll be sorry. They're going to need something more out of the 4000+ than what they're going to provide in order to stay convincingly on top.

And, let's face it, Intel is still winning some benchmarks, namely some encoding benchmarks along with 3DSMax benchmarks. AMD has yet to gain complete victory.

If they can, they should.

I agree that the need to meet demand, but they should be able to meet demand as long as they're selling all their different chips(and not just selling a ton of their theoretical uber high-end cpu). The point is to get across to people that "AMD chips are the fastest" so that demand for all their cpus will rise.
Click to expand...

The problem is that "Proving" they can do it is pointless if they can't supply the parts. It would be like Rolls Royce trying to topple GM from the most car sales list by producing the best $20k econobox ever, but only being able to handmake 500 cars a year. Everyone's mouth would water, but only 500 could actually own one.

It will do AMD no good to have everyone want an AMD processor when they can't provide everyone with one. It is widely believed this is a primary reason why Dell has never seriously considered AMD, AMD just doesn't have the ability to provide all the processors that Dell needs in a reliable fashion. Slow and steady.
 
Originally posted by: sandorski
Originally posted by: DrMrLordX
Sandorski. the main reason to produce something that would, as you said, "own" anything available from Intel is to

a). prove that they can do it
b). attract the attention of OEMs

Simple as that. A stable, low-supply, high-end CPU available for benchmarks would do the trick, as long as its clearly labeled as a non-enthusiast part. In other words, if they rest on the FX-53 for too long, they'll be sorry. They're going to need something more out of the 4000+ than what they're going to provide in order to stay convincingly on top.

And, let's face it, Intel is still winning some benchmarks, namely some encoding benchmarks along with 3DSMax benchmarks. AMD has yet to gain complete victory.

If they can, they should.

I agree that the need to meet demand, but they should be able to meet demand as long as they're selling all their different chips(and not just selling a ton of their theoretical uber high-end cpu). The point is to get across to people that "AMD chips are the fastest" so that demand for all their cpus will rise.
Click to expand...

The problem is that "Proving" they can do it is pointless if they can't supply the parts. It would be like Rolls Royce trying to topple GM from the most car sales list by producing the best $20k econobox ever, but only being able to handmake 500 cars a year. Everyone's mouth would water, but only 500 could actually own one.

It will do AMD no good to have everyone want an AMD processor when they can't provide everyone with one. It is widely believed this is a primary reason why Dell has never seriously considered AMD, AMD just doesn't have the ability to provide all the processors that Dell needs in a reliable fashion. Slow and steady.
Click to expand...

Thats what banks are for. If Dell was serious about carrying AMD they could have the money by tomorrow. I don't know how long it would take to build the fabs but it's a non-issue since OEMs are bribed by intel to stay intel. Only consumer demand for AMD processors can change OEM attitude and only overwhleming performance leads can percipitate such a demand.
 
Well,Zebo(as ever) and iwantanewcompuetr have hit the nail on the head, but id like to add a few points?..

FABS

AMD has one fab for CPU?S, FAB 30,it?s the saving grace ,a miracle ground for AMD, but it is the only one FAB. Sure FAB 36 is on the way but once its fully operational FAB 30 will be converted into flash. With one fab AMD cant really aim for high market share numbers or maintain them. Dell ?might? take opteron onboard but its only cos it knows AMD can make enough of them, remember the server market is not as big as the desktop market,so AMD don?t have to sweat as much worrying about making huge numbers.

As Zebo stated, they can only aim for performance, that?s it. So Ramp up?.

I just wish IBM would buy AMD and pump new into it( as well as funds), I was going to write a hell of a lot more but I got go to lessons.
 
yeah i know i was kind of nitpicking and they should start cranking out faster chips(still at 2.4 for the 4000 is getting rediculous), or get to the point where they can get faster ones. i think they are delaying higher clocks so they just don't run out of upgrades in a year. it's going to be like mid 2006 before they start 65 nm production, and way things with 90 went probly later. they can only add so much speed to the hammer core (which is a lot like the previous athlon xp cores with upgrades, not design changes) without a major redesign (ie northwood -> prescott). i think they are timing releases so they don't release like a 3 GHz now and have nothing they can upgrqade that to in the next year. they have to slowly introduce upgrades for the sales in the future. they might be holding back on revisions that will allow hamers to get to 2.6-2.8 easily, but i doubt it.

they probly are playing their cards slowly to stay just ahead of intel, but couldn't play them much faster if they had to.

btw just saw this posted in response to that same article. pretty much same thing i said
overclockers.com

good points all. amd only does have 1 fab, not capable of much more, they could get huge loans and build another if they had the sales, but this would take a long time remember
 
Originally posted by: Zebo
Thats what banks are for. If Dell was serious about carrying AMD they could have the money by tomorrow. I don't know how long it would take to build the fabs but it's a non-issue since OEMs are bribed by intel to stay intel. Only consumer demand for AMD processors can change OEM attitude and only overwhleming performance leads can percipitate such a demand.
Click to expand...
It takes about 4-5 years (and $2.5-$3 billion) to build a fab from the ground up. If Dell were to commit to using AMD cpu's in five years, that certainly wouldn't guarantee that AMD could secure a few billion dollars. (Or that it would be a smart business decision on AMD's part.)

AMD currently has Fab36 under construction in Dresden. It's a 300mm wafer fab, and is slated to begin production on the .065µ process in 2006.

AMD has had a few agreements in the past (UMC, IBM, etc) to begin producing AMD chips in their factories as a foundry. None of those has come to fruition.

And yes... If AMD could absolutely guarantee Dell the supply they would need, I have no doubt Dell would begin using AMD.

I realize those facts aren't as fun as conspiracy stories and circumventing anti-trust laws... 😉

 
Wingznut good post, but I bet Intel is on dells back watching them look over into AMD'S garden when it comes to opteron....

" No you dont wanna play with them over there.......There different to us"
 
The problem is that "Proving" they can do it is pointless if they can't supply the parts.
Click to expand...
Very good point. If you ever need proof that building a high end part that you can't deliver in quantities angers your customers, just pop into the video forum. The last thing we need from AMD or Intel is a 6 month paper launch of each new cpu.
 
Originally posted by: clarkey01
Wingznut good post, but I bet Intel is on dells back watching them look over into AMD'S garden when it comes to opteron....
Click to expand...
It wouldn't surprise me at all if Dell started selling Opteron. Again... Presuming that AMD can guarantee Dell the supply (and priority) they require.

 
Originally posted by: Wingznut
Originally posted by: Zebo
Thats what banks are for. If Dell was serious about carrying AMD they could have the money by tomorrow. I don't know how long it would take to build the fabs but it's a non-issue since OEMs are bribed by intel to stay intel. Only consumer demand for AMD processors can change OEM attitude and only overwhleming performance leads can percipitate such a demand.
Click to expand...
It takes about 4-5 years (and $2.5-$3 billion) to build a fab from the ground up. If Dell were to commit to using AMD cpu's in five years, that certainly wouldn't guarantee that AMD could secure a few billion dollars. (Or that it would be a smart business decision on AMD's part.)

AMD currently has Fab36 under construction in Dresden. It's a 300mm wafer fab, and is slated to begin production on the .065µ process in 2006.

AMD has had a few agreements in the past (UMC, IBM, etc) to begin producing AMD chips in their factories as a foundry. None of those has come to fruition.

And yes... If AMD could absolutely guarantee Dell the supply they would need, I have no doubt Dell would begin using AMD.

I realize those facts aren't as fun as conspiracy stories and circumventing anti-trust laws... 😉
Click to expand...


4-5 years.:Q

No ones consprititizing... Dell gets massive volume discounts for MS OS, Intel chips and a ton of other things too.. Perfectly legal too. Otherwise known as a "volume discount." AMD can not match these prices since they produce less processors but still need the engineering and leadership just like Intel has therefore they have higher fixed costs per unit because they sell so much less than intel.

 
Intresting article with some good points.

One part I didn't agree with esspecially, however:
The author stated that when AMD was sitting with it's 1Ghz Athlon, the original P4 came along and completely outperformed the Athlon. That's not how I remembered it. The original P4 wasn't that great of a processor in the beginning. It relied on a high clock speed more than any other non-Celeron processor I can think of. The 1400mhz P4 outperformed the 1Ghz Athlon in some cases, but the Athlon had wins as well, so I wouldn't say the P4 really dethroned Athlon. It wasn't until the Northwood that the P4 really started becomming a power house. But it wasn't long after that the the T-bred was released.

See? If history does repeat itself, when the P5, or whatever is released, AMD will already have or quickly create a processor that is very competitive. Yet this time AMD has an advantage. In the past few years, their following has increased exponentially, esspecially with enthusiests.

Just my 2 cents
 
The author is way wrong on that point, AMD t-bird's and pal's dominated the P4A's all the way. Then the B's came giving a slight lead to intel but AMD was still hanging in there with T-breds. Then the C's came and totally destroyed AMD's PR rating BS. I mean look at some tests a P4C 2800 on 865PE beat even a A-XP 3200 in majority of tests. AMD credibility was hurting bad until the A64 came out...and may be about to hurt again if they don't scale more.😉🙂
 
Originally posted by: LTC8K6
Is AMD going to support DDR2 soon? It looks promising if this thread is any indication.


DDR2
Click to expand...

Expected when DDR2 hits 667MHz (around April next year) on Socket 939 (not 754)...
 
Originally posted by: LTC8K6
Intel has dual core up their sleeve as well, so that is probably going to work out to a tie. Which is just what we all need to keep prices reasonable.
Click to expand...

Intel's dual core (except for Itanium) probably won't show up till the end of next year. AMD should have a good 6 months lead time on them...
 
Originally posted by: clarkey01
Well,Zebo(as ever) and iwantanewcompuetr have hit the nail on the head, but id like to add a few points?..

FABS

AMD has one fab for CPU?S, FAB 30,it?s the saving grace ,a miracle ground for AMD, but it is the only one FAB. Sure FAB 36 is on the way but once its fully operational FAB 30 will be converted into flash. With one fab AMD cant really aim for high market share numbers or maintain them. Dell ?might? take opteron onboard but its only cos it knows AMD can make enough of them, remember the server market is not as big as the desktop market,so AMD don?t have to sweat as much worrying about making huge numbers.

As Zebo stated, they can only aim for performance, that?s it. So Ramp up?.

I just wish IBM would buy AMD and pump new into it( as well as funds), I was going to write a hell of a lot more but I got go to lessons.
Click to expand...

Actually, Fab 30 will remain CPU production even after Fab 36 is in production.
However, the capacity of Fab36 will be HUGE by comparison. FAB36 will be able to have 13,000-20,000 Wafer Starts Per Week (WSPW) at 300mm, compared to Fab 30's current 5,500 WSPW at 200mm.
Fab 36 will be 65nm as well...at dual core.
If the size of the dual cores (my guess) at 65nm are around 99mm2, then the capacity for production at 50% yield will be ~4.3 - 6.6 million dual core cpus/week...223.4 - 343.7 million/year.
 
Intel doesn't even have the 3.6 out in the wild in any real numbers. AMD Fx-55 will be 2600 remember. Basically AMD usually scales from the bottom of the line to the top in a die shrink so the 4000 will be insurance of scaling in case there are real issues. The 4000 will defeat the 3600 from intel (and probably the 3800 easily) and then the AMD 4200 (2600) comes out in quarter 1 and the FX-57 (2800) rolls out with semi mature yields. Seems smart enough to me. If the 4000 cost $700 then Amd will have problems however. I still believe that AMD's forturnes reside in the server market and flash market. The Opteron is the key for AMD's success
 
Originally posted by: Viditor
Actually, Fab 30 will remain CPU production even after Fab 36 is in production.
However, the capacity of Fab36 will be HUGE by comparison. FAB36 will be able to have 13,000-20,000 Wafer Starts Per Week (WSPW) at 300mm, compared to Fab 30's current 5,500 WSPW at 200mm.
Fab 36 will be 65nm as well...at dual core.
If the size of the dual cores (my guess) at 65nm are around 99mm2, then the capacity for production at 50% yield will be ~4.3 - 6.6 million dual core cpus/week...223.4 - 343.7 million/year.
Click to expand...
There's just no way it would be able to run 20k WSPW. That would require somewhere in the neighborhood of 400,000 sq ft of cleanroom space, ~3000 engineers and technicians, and probably ~$7 billion dollars.

Since Fab 36 is ~125,000 sq ft of cleanroom, will employ ~1000 people, and is costing AMD ~$2.4 billion.... I'm guessing wafer starts will be in the 6k/wk range.

As for what will become of Fab 30... The market will dictate that. I'm sure there are no plans set in stone for 2 years from now.
 
To take the title on a literal basis...

When has AMD ever NOT been in danger? A small company taking on a giant and his castle. Granted AMD may have held the upper hand for sustained periods over its history and it is in itself massive compared to what it used to be... its still pretty small compared to Intel.

Until AMD close the gap even more, its still going to be in danger. Its R&D department seem to be pulling off miracles that Intels cant.... early reports suggest that AMD's 0.09 venture may be significantly better for the Athlon line then 0.09 was for the P4 line and how big is the difference in R&D budgets?

I think the fact that AMD are continuing to innovate and are able to stay reasonably close to Intel in their dogfight is what is keeping them going and I for one enjoy the competition. Everyone who understands the situation fully always likes the underdog better 🙂
 
Originally posted by: Wingznut
Originally posted by: Viditor
Actually, Fab 30 will remain CPU production even after Fab 36 is in production.
However, the capacity of Fab36 will be HUGE by comparison. FAB36 will be able to have 13,000-20,000 Wafer Starts Per Week (WSPW) at 300mm, compared to Fab 30's current 5,500 WSPW at 200mm.
Fab 36 will be 65nm as well...at dual core.
If the size of the dual cores (my guess) at 65nm are around 99mm2, then the capacity for production at 50% yield will be ~4.3 - 6.6 million dual core cpus/week...223.4 - 343.7 million/year.
Click to expand...
There's just no way it would be able to run 20k WSPW. That would require somewhere in the neighborhood of 400,000 sq ft of cleanroom space, ~3000 engineers and technicians, and probably ~$7 billion dollars.

Since Fab 36 is ~125,000 sq ft of cleanroom, will employ ~1000 people, and is costing AMD ~$2.4 billion.... I'm guessing wafer starts will be in the 6k/wk range.

As for what will become of Fab 30... The market will dictate that. I'm sure there are no plans set in stone for 2 years from now.
Click to expand...

I shall enjoy this crow lunch...sorry, you're right about the capacity on the floor space (I need more sleep).

The cost of Fab36 is indeed $2.4 billion, but AMD isn't paying for all of that...$500 million is from grants, $700 million in extremely low cost government (EU) loans, and $200 million from equity partners in Europe...
As to Fab 30, it will certainly be in CPU operation through 2006 and most likely well beyond (just as Fab 25 was).
 
Originally posted by: Elcs
To take the title on a literal basis...

When has AMD ever NOT been in danger? A small company taking on a giant and his castle. Granted AMD may have held the upper hand for sustained periods over its history and it is in itself massive compared to what it used to be... its still pretty small compared to Intel.

Until AMD close the gap even more, its still going to be in danger. Its R&D department seem to be pulling off miracles that Intels cant.... early reports suggest that AMD's 0.09 venture may be significantly better for the Athlon line then 0.09 was for the P4 line and how big is the difference in R&D budgets?

I think the fact that AMD are continuing to innovate and are able to stay reasonably close to Intel in their dogfight is what is keeping them going and I for one enjoy the competition. Everyone who understands the situation fully always likes the underdog better 🙂
Click to expand...

There is a huge Philosophical difference between Intel and AMD, though for AMD it's not entirely by choice. That difference is that AMD relies on others much more and is willing to work with others(IBM for eg) for developement of key technologies. OTOH, Intel likes to do almost everything In-House.

This diffeence allows AMD to concentrate on the Design of their CPUs more and let others work on the Manufacturing Processes and Materials Research. It is the Philosophy that has allowed AMD to be the first to use Cu Interconnects, SOI, and even recognize(so it appears) that MHZ speeds were hitting a wall at approx 3ghz. At the sametime AMD was adopting these Manufaturing Technologies, Intel wasn't adopting them, but rather trying to develop their own version or variations that they could Patent as their own. As long as these 2 Philosophies are adhered to by AMD/Intel, I think AMD will continue to be competitive even though Intel would appear, at first glance, to have the upperhand in R&D.
 
Originally posted by: sandorski
Originally posted by: Elcs
To take the title on a literal basis...

When has AMD ever NOT been in danger? A small company taking on a giant and his castle. Granted AMD may have held the upper hand for sustained periods over its history and it is in itself massive compared to what it used to be... its still pretty small compared to Intel.

Until AMD close the gap even more, its still going to be in danger. Its R&D department seem to be pulling off miracles that Intels cant.... early reports suggest that AMD's 0.09 venture may be significantly better for the Athlon line then 0.09 was for the P4 line and how big is the difference in R&D budgets?

I think the fact that AMD are continuing to innovate and are able to stay reasonably close to Intel in their dogfight is what is keeping them going and I for one enjoy the competition. Everyone who understands the situation fully always likes the underdog better 🙂
Click to expand...

There is a huge Philosophical difference between Intel and AMD, though for AMD it's not entirely by choice. That difference is that AMD relies on others much more and is willing to work with others(IBM for eg) for developement of key technologies. OTOH, Intel likes to do almost everything In-House.

This diffeence allows AMD to concentrate on the Design of their CPUs more and let others work on the Manufacturing Processes and Materials Research. It is the Philosophy that has allowed AMD to be the first to use Cu Interconnects, SOI, and even recognize(so it appears) that MHZ speeds were hitting a wall at approx 3ghz. At the sametime AMD was adopting these Manufaturing Technologies, Intel wasn't adopting them, but rather trying to develop their own version or variations that they could Patent as their own. As long as these 2 Philosophies are adhered to by AMD/Intel, I think AMD will continue to be competitive even though Intel would appear, at first glance, to have the upperhand in R&D.
Click to expand...

Well said. I knew AMD outsourced/collaborated with a few companies to cover some of the issues that they couldnt fully fund but I did not realise it was to that extent.

Although Intel alone make AMD look small, AMD's business associations are actually giving them signifigant leverage in their CPU battle? Impressive that such a strategy is being pulled off. It must really show that AMD are quite serious about becoming the dominant force and for large companies to share support, it must show that AMD are showing some good initiative if only in their technological developments.

Personally, Im glad of Intels 'downtime' currently. AMD need to become a bit stronger in the market place with respect to the customer knowledgebase and being pretty much the top-dog by way of performance is a good way to get there imo.

I always get the feeling that Intel has something decent in the works but only time will tell.

Thank you for the enlightenment.

But i still think AMD are in some sort of danger. What they do now could seriously affect their future.
 
Originally posted by: Viditor
Originally posted by: LTC8K6
Intel has dual core up their sleeve as well, so that is probably going to work out to a tie. Which is just what we all need to keep prices reasonable.
Click to expand...

Intel's dual core (except for Itanium) probably won't show up till the end of next year. AMD should have a good 6 months lead time on them...
Click to expand...

I think that dual core is going to be where AMD finally makes the big leap ahead of Intel. I don't think that Intel will get their own on-die memory controller technology going before the AMD dual core launch, so they may have to slap together a dual core chip that is using a shared FSB, and to make up for that bottleneck they'll heap on a bunch of cache (like the 24MB used in the Montecito dual core design), which will drive up the cost of the chip and heat dissipation - maybe even to the point where they have to paper launch a chip with issues (like the first 1.13 GHz PIII). The AMD version, which will get away with less cache because of their on die controller, being cheaper and running cooler will have a distinct advantage as far as cost and availability, letting AMD gain the performance crown for quite a while.
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top