AMD wants us to breakfree

[spaztic7]

Wow... we just went from debating if AMD has a legit reason to pursue a course of action against Intel, to DDR3. Whoever did that, great job on the thread highjack!

 

[DJ_Jumbles]

Try to find a faster 939 in stock, somewhere, yet another “new” AMD socket that was “supposedly” going to be supported for awhile (amid wildly varied rumors of AM2+ and AM3) or a socket 775, which already had proven longevity.

Yeah... I think that you're confusing the fact that the socket is the same, but if you wanted to upgrade to the C2D, you had to upgrade your chipset. or wait... DDR3, had to change the chipset, or WAIT!!! PCIe2- GOING TO UPGRADE THE CHIPSET!!! What does it matter if the socket stays the same if the functionality is crippled? You're talking apples, yet seeing oranges. All you have to do is swap out the word SOCKET with CHIPSET and *poof* it's the same argument turned around again. The fact that AMD changed socket designs doesnt mean that the upgrade path wasn't the EXACT same as the 775 in essence because if a new chipset feature came out, you had to get a newer 775 processor that yould utilize it and vice-versa. Want to use a newer 1300FSB C2D? Gotta buy a new P35 motherboard. Want to use DDR3? Have to buy a new P35 motherboard. Why the hell would you buy a processor and cripple it on an older chipset? That's just foolish on your part.
The 775 socket has been around a while now and is starting to show it's age, frankly. Intel just won't innovate past it. It's not like as if it's the pinnacle of technological advancement in the CPU socket race. I can't wait for Intel to start integrating the memory controller and off-loading some of the CPU overhead off the FSB.
So... you mean to tell me that it's a good investment to buy a Pentiac D 775 with a P35 chipset? Or a e6850 with a 955 chipset at 1066FSB? See? Granted, the older chips will work on the newer boards, but what's the point?

 

[lcandy]

turpit, really well said. I read through this entire nightmare of a thread to get your last few posts to realise I would not have much to say.

Many many times have I sold CPU's for AMD or Intel, depending on who was at the top at the time. Consider thousands and thousands of enthusiasts, with varying levels of influence and yes they do make a big difference. Enthusiasts really helped AMD big time when they were not advertising much. But hey, I am just repeating what you said effectively.

What amazed me about six months ago, a couple of friends were going to buy computers. After the usual advice on what to buy, I use the phrase, "... and remember, whatever you do, you must get a CPU by..." and was interupted, "...yes I know, get AMD...". After a few years of ramming it down peoples throats to get AMD because they were superior in just about every way, I now had to undo the work and convince them now Intel was superior. Sometimes I wonder why I bother lol.

As for the legal wranglings, Intel are in a spot of trouble, but how much? There is no doubt they are not whiter than white, it just depends what law, when and where was broken. Regardless, it would not have stopped AMD getting their butt's kicked today. In fact, AMD just annoys me now because of the way they are approaching this. The only thing they are doing is proving no matter how bad Intel may have been in the past when their products were not good, AMD can be just as bad when they have an inferior product. Neither one of these companies deserve brand loyalty IMHO.

 

[Mousemonkey]

@ spaztic7 The thread wasn't hijacked but it was threatening to turn into a playground type slanging match.

@ lcandy I agree with you that neither AMD or Intel deserve brand loyalty, but AMD had lots nonetheless most of which they have most likely lost due in no small part to the way that they conducted themselves over the past year or so, which makes this upcoming lawsuit all the more interesting, can you imagine Hecter telling the Judge "I'll have the papers with you on the 24th, no make that the 28th, oh, can we do it on the 17th of next month" [:mousemonkey:1]

 

[wolverinero79]

http://breakfree.amd.com/en-us/default.aspx
* Provided below-cost microprocessors for servers to prevent rival solutions from being chosen by strategic customers.

So that's the argument with the least amount of truth. Below cost means that basically you operate at a loss in order to sell your chips cheaply (for the purpose of hurting your competitor). Intel hasn't had a quarter where they have lost money in...well...I can't even remember - it's been a verrrry long time. AMD has been selling their chips "below cost", so I guess Intel should sue them. 😀


Intel could drive down ASPs (Average Selling Prices) much farther and still make money (AND IT WOULD BE 100% LEGAL!).

The fact is, that people trusted Intel and guess what...the P4 did what was promised. It was a CPU that gave decent performance with good durability. People who bought a P4 used it in a computer and it did what it was supposed to. Intel is still producing and people are still buying P4s, regardless of the better chips in existence. Don't blame Intel for people enjoying their products when there was a better product out there (and that's still arguable - there were definite strengths in the P4 line that couldn't be touched by AMD - the idea that the Athlon 64 was just fantastically better everywhere is a joke - many times any lead was marginal at best). People are quick to demonize the large organization without paying attention to facts.

Germany (big influence in the EU) has spent tons of money in subsidies for AMD - do you really think the EU will be impartial in a legal case b/w Intel and AMD? It'd be like having the case in Israel (major center for production and design for Intel).

 

[turpit]

Yeah... I think that you're confusing the fact that the socket is the same, but if you wanted to upgrade to the C2D, you had to upgrade your chipset. or wait... DDR3, had to change the chipset, or WAIT!!! PCIe2- GOING TO UPGRADE THE CHIPSET!!! What does it matter if the socket stays the same if the functionality is crippled? You're talking apples, yet seeing oranges. All you have to do is swap out the word SOCKET with CHIPSET and *poof* it's the same argument turned around again. The fact that AMD changed socket designs doesnt mean that the upgrade path wasn't the EXACT same as the 775 in essence because if a new chipset feature came out, you had to get a newer 775 processor that yould utilize it and vice-versa. Want to use a newer 1300FSB C2D? Gotta buy a new P35 motherboard. Want to use DDR3? Have to buy a new P35 motherboard. Why the hell would you buy a processor and cripple it on an older chipset? That's just foolish on your part.
The 775 socket has been around a while now and is starting to show it's age, frankly. Intel just won't innovate past it. It's not like as if it's the pinnacle of technological advancement in the CPU socket race. I can't wait for Intel to start integrating the memory controller and off-loading some of the CPU overhead off the FSB.
So... you mean to tell me that it's a good investment to buy a Pentiac D 775 with a P35 chipset? Or a e6850 with a 955 chipset at 1066FSB? See? Granted, the older chips will work on the newer boards, but what's the point?

The changes you listed to the socket 775 are all factual, and would be germane to the debate if it was about things Intel did to PO its consumer base. But it wasn’t. It was not about changes to mobo specs, but about what AMD did to PO its own consumer base by stating a course of action which caused people to spend money, then failing to adhere to that course of action. In short, AMD pulled an Intel. AMD made a statement. They stated would support the 939 for a significant period of time after the introduction of AM2. They reversed that decision and didn’t. They left everyone who bought the 939s, based on AMD's promises, swinging in the wind and POing them in the process.

Frankly, if your argument is that AMD & AMD mobo manufacturers don’t change mobo specs or chipsets then I’m disappointed and your argument is incontestably wrong. As proof, I offer socket A. AMD's long lived socket that went through many significant chipset and bus changes, maturing from EIDE 66 all the way through EIDE 133 and then to SATA, from PCI to AGP, then AGP 2x, then 4x, then 8x, and ATX to ATX 12, along the way incorporating many new chipsets to support those changes, for ex in VIA chipsets from KT 133 all the way up to the KT 880s. This is just to name a few of the changes socket A went through, in the process limiting upgradeability of that particular socket, just as socket 775s were limited. In the particular case of 939 however, the type of changes you reference really didn’t apply as 939 wasn’t around long enough to undergo a lot of significant changes.

As to whether or not Intel will or won’t "innovate" past 775 is debatable. One theory is they haven’t changed to socket to avoid POing their own customer base as they did when they promised everyone that socket5 would support Pentium I, and then jumped to socket 7 for Pentium instead, leaving those people who bought socket 5s based on the ability to upgrade hanging in the wind. They did it again when they made socket 8 for P Pros and went from socket7 to slot 1 to socket 370 in such a short period of time.

As to every change to the C2D line forcing a new chipset/mobo that jsut isnt true. Every change has not necessitated a new mobo. The specific case you use of going to an E6x50 1333FSB chip does not mean a new mobo with a P35 chipset as you claim. Most all 975X and a fair number of 965P chipset mobos support 1333FBS with a simple bios upgrade, meaning most Intel chipped mobos produced over the last 18 months are still upgradeable. Don’t confuse this with older chipped mobos that have been out of production but are still in retail stock and supported by the manufacturers. For the few people who want to pay the premium for the underperforming, over priced DDR3, yes, that does require a new chipset, but I for one, have no desire to move to DDR3 anytime soon. Especially with both prices and latencies on DDR2 still falling while a higher clockspeed recently became avalable. At this point in time, DDR3 is still mostly pointless.

Finally, it was never a good idea to buy a Pentium D, on any chipset.

 

[turpit]

turpit, really well said. I read through this entire nightmare of a thread to get your last few posts to realise I would not have much to say.

Many many times have I sold CPU's for AMD or Intel, depending on who was at the top at the time. Consider thousands and thousands of enthusiasts, with varying levels of influence and yes they do make a big difference. Enthusiasts really helped AMD big time when they were not advertising much. But hey, I am just repeating what you said effectively.

What amazed me about six months ago, a couple of friends were going to buy computers. After the usual advice on what to buy, I use the phrase, "... and remember, whatever you do, you must get a CPU by..." and was interupted, "...yes I know, get AMD...". After a few years of ramming it down peoples throats to get AMD because they were superior in just about every way, I now had to undo the work and convince them now Intel was superior. Sometimes I wonder why I bother lol.

As for the legal wranglings, Intel are in a spot of trouble, but how much? There is no doubt they are not whiter than white, it just depends what law, when and where was broken. Regardless, it would not have stopped AMD getting their butt's kicked today. In fact, AMD just annoys me now because of the way they are approaching this. The only thing they are doing is proving no matter how bad Intel may have been in the past when their products were not good, AMD can be just as bad when they have an inferior product. Neither one of these companies deserve brand loyalty IMHO.

Well said.

My post comes off as if I might be trying to say Intel is clean. Im not. Some people didnt get that, some people, like you did. My point was that the changes in market share have demonstrated that both AMD and Intel sales have waxed and waned based on performance, value and advertising, not any dirty dealing.

It annoys me that people say AMD was held back......as if claiming almost 30% of the overall market was being held back. It belittles AMDs products and corporate accomplishments. AMD wasnt held back and they were succeeding as well as anyone could have hoped....right up to the point they let Intel regain the high ground in performance and value. All the people who cry foul at Intel fail to realize that had Intel not released C2D, or had C2D failed, based on the rate AMD was consuming marketshare, they probably would have achieved greater than 50% by the end of CY08 if they could have expanded or contracted out sufficeint manufacturing capacity to supply the demand. That and AMD X2 prices would still be high....competition works both ways 😉

 

[vylo]

The OP is hilarious. "AMD wants us to break free!"

Yeah I'm sure they all want to give us handjobs and free beer too.

AMD and Intel want one thing.

YOUR MONEY.

 

[Mousemonkey]

The OP is hilarious. "AMD wants us to break free!"

Yeah I'm sure they all want to give us handjobs and free beer too.

AMD and Intel want one thing.

YOUR MONEY.

AMD & Intel simply exist because they 😍 you.

 

[MU_Engineer]

The changes you listed to the socket 775 are all factual, and would be germane to the debate if it was about things Intel did to PO its consumer base. But it wasn’t. It was not about changes to mobo specs, but about what AMD did to PO its own consumer base by stating a course of action which caused people to spend money, then failing to adhere to that course of action. In short, AMD pulled an Intel. AMD made a statement. They stated would support the 939 for a significant period of time after the introduction of AM2. They reversed that decision and didn’t. They left everyone who bought the 939s, based on AMD's promises, swinging in the wind and POing them in the process.

I think that part of AMD's reason for dumping 939 and moving to AM2 so quickly had to deal with the rise of multi-core chips and the fact that the fastest common DDR memory was DDR-400. Dual-channel DDR-400 worked fine for the X2s on up to 2.6 GHz or so, where the line ended. AMD would have needed DDR-500 or faster to become a JEDEC spec to push the speeds up farther or else performance would have been bottlenecked by the memory, according to tests of the FX-60 vs. the AM2 chips. They had to move for a reason. Do I think they could have done it better? Sure, they should have never have made 939 and kept the FX-55 and 57 on Socket 940, and all of the rest on Socket 754. The dual-core chips would then have been on AM2 as DDR2 was certainly around at that time as Intel had been using it for some time already. It wasn't at the sweet spot for price vs. performance like DDR was at the time the X2s debuted, but it wasn't that far off.

Frankly, if your argument is that AMD & AMD mobo manufacturers don’t change mobo specs or chipsets then I’m disappointed and your argument is incontestably wrong. As proof, I offer socket A. AMD's long lived socket that went through many significant chipset and bus changes, maturing from EIDE 66 all the way through EIDE 133 and then to SATA, from PCI to AGP, then AGP 2x, then 4x, then 8x, and ATX to ATX 12, along the way incorporating many new chipsets to support those changes, for ex in VIA chipsets from KT 133 all the way up to the KT 880s. This is just to name a few of the changes socket A went through, in the process limiting upgradeability of that particular socket, just as socket 775s were limited. In the particular case of 939 however, the type of changes you reference really didn’t apply as 939 wasn’t around long enough to undergo a lot of significant changes.

There were two significant changes: AGP -> PCIe and the rise of dual-core CPUs. The former certainly required a new board to take advantage of, but the latter only did in the case of Intel CPUs. If you had an early 775 board with a 915 or 925 chipset for the Prescott, you were hosed and could not put a Pentium D in there, while any 939 board could take an Athlon 64 X2. If you remember, Intel *purposefully* sabotaged the ability of the 915/925 to work with dual-core CPUs as they were pissed that many board makers took advantage of the fact that the 865/875 worked with multiple CPUs and made low-end dual Xeon boards out of them without paying for a much more expensive E7300/E7500 Xeon chipset.

As to whether or not Intel will or won’t "innovate" past 775 is debatable. One theory is they haven’t changed to socket to avoid POing their own customer base as they did when they promised everyone that socket5 would support Pentium I, and then jumped to socket 7 for Pentium instead, leaving those people who bought socket 5s based on the ability to upgrade hanging in the wind. They did it again when they made socket 8 for P Pros and went from socket7 to slot 1 to socket 370 in such a short period of time.

Well, as far we know that the 775 will end with the Wolfdale/Yorkfield CPUs as LGA 715 will replace it. (LGA 771/Socket J will also be replaced with LGA 1366.) Intel did change from Socket 5 to Socket 7 and I am sure it PO'd a few people, but there were Pentiums that were socket 5- the Pentium OverDrive series. These worked fine in Socket 5 486 boards. I can certainly understand why Socket 8 wasn't used for anything after the Pentium Pro as it was a very weird socket with a rectangular PGA with two different pin grid densities, one for the CPU and one for the MCM's L2 cache. However, Intel could have done without 370 and 423 and just gone straight from Slot 1 to 478. Slot 1 was originally intended for Intel to be able to put large amounts of cache just off-die on a daughterboard, but later chips had all of the L2 on-die and just used the daughterboard as a BGA mounting place. There weren't many chips that were 370-only- the Tualatin PIIIs were the only ones I recall- so those chips could have been on Slot 1 at the end of its days. The 423 should have been skipped as it was inadequate for the P4, so a Slot 1 -> 478 transition would have been fine.

 

[lcandy]

Yes, all true turpit.

The thing is though, I clearly remember that time not so long ago - logging onto Dell's website, thinking about buying a new PC with a wonderful AMD chip inside and I did feel annoyed at them for not stocking it. I felt annoyed that I couldn't get a ready made machine from a lot of well established retailers. I felt further annoyed that when considering building an AMD based machine that the chips were so expensive I was put off anyway. I wanted dual core the second AMD made it a success - I had to wait for Intel to hit the market right.

So right back on topic, someone, not saying it was Intel for sure, but someone, was placing barriers on AMD from getting into some of the big retailers. I just can't help but feel that if AMD's chips were priced more aggressively they would have blown away most if not all barriers placed on them.

Maybe this is what it comes down to? Knowledge of the market? Whether it is: advertising, pricing, supplying the right areas, producing enough at the right time or simply knowing the right people to sell to - Intel often seems to get this right and AMD does less so (neither all the time). I can't be sure of the facts and I am not sure anyone can, especially those already prejudging Intel of making billions of $/£ worth of illegal practices under EU law.

 

[ImajorI]

The Intel AMD flames are the best.

the whole point of this is that intel were preventing you the consumer from even having a chance of buying AMD CPU's. The whole point of this is that mainstream buyers and the business world COULD NOT buy AMD because of intel's anti-competative practices. have i drilled it into you yet, do you have even the slightest comprehension of the seriousness of these allegations.

I have owned both AMD and Intel stock. I generally like your posts and value your opinions, but feel your very wrong here.

You know, this thread, in various guises, has been popping up over and over again for months and months. Of course, as many people keep trying to insist, AMD’s current situation is the result of those ‘evil’ machinations. AMD is the virginally pure, wholesome underdog that does no wrong….the democrats of the CPU industry….never to blame for their own situation, yada yada yada.

WRONG.

What is happening right now with AMD has absolutely nothing what so ever to do with any of the allegations against Intel. Too sum up.. AMD’s current situation is of their own making, not Intel’s.

I agree with Turpit. AMD is in trouble because of their current products stink, not because Intel is a bully. What do you AMD guys think, life is always fair, it ain’t. These companies aren’t playing wiffle ball , its hardball out in the real world. All the "I want AMD to sue and win" talk sounds like whining to me.

As Sean Connary said “You tried your best? Losers try their best, winners go home and screw the prom queen.”

 

[turpit]

I think that part of AMD's reason for dumping 939 and moving to AM2 so quickly had to deal with the rise of multi-core chips and the fact that the fastest common DDR memory was DDR-400. Dual-channel DDR-400 worked fine for the X2s on up to 2.6 GHz or so, where the line ended. AMD would have needed DDR-500 or faster to become a JEDEC spec to push the speeds up farther or else performance would have been bottlenecked by the memory, according to tests of the FX-60 vs. the AM2 chips. They had to move for a reason. Do I think they could have done it better? Sure, they should have never have made 939 and kept the FX-55 and 57 on Socket 940, and all of the rest on Socket 754. The dual-core chips would then have been on AM2 as DDR2 was certainly around at that time as Intel had been using it for some time already. It wasn't at the sweet spot for price vs. performance like DDR was at the time the X2s debuted, but it wasn't that far off.

I've never debated AMD's need to shift to DDR2 nor would I. They did need to, at some point in time. If you chose to believe their statements, they deliberatly choose the timeframe rather than just chasing blindly after Intel. If thats true, then, IMO, it was a good choice. But regradless of AMDs descisions regarding the implementation of DDR2 via AM2, the statement was never about whether they needed to go to AM2 (DDR2) or the whys, but that they had stated they would continue supporting 939, then reneged on that statement leaving those who had purchased based on it in a lurch.

There were two significant changes: AGP -> PCIe and the rise of dual-core CPUs. The former certainly required a new board to take advantage of, but the latter only did in the case of Intel CPUs. If you had an early 775 board with a 915 or 925 chipset for the Prescott, you were hosed and could not put a Pentium D in there, while any 939 board could take an Athlon 64 X2. If you remember, Intel *purposefully* sabotaged the ability of the 915/925 to work with dual-core CPUs as they were pissed that many board makers took advantage of the fact that the 865/875 worked with multiple CPUs and made low-end dual Xeon boards out of them without paying for a much more expensive E7300/E7500 Xeon chipset.

Im not sure what you point is here. That Intel deliberately forced a change? If so, that is what it is, but Im not sure how it applys to the original debate that AMD cut the life of 939 short.

Well, as far we know that the 775 will end with the Wolfdale/Yorkfield CPUs as LGA 715 will replace it. (LGA 771/Socket J will also be replaced with LGA 1366.) Intel did change from Socket 5 to Socket 7 and I am sure it PO'd a few people, but there were Pentiums that were socket 5- the Pentium OverDrive series. These worked fine in Socket 5 486 boards. I can certainly understand why Socket 8 wasn't used for anything after the Pentium Pro as it was a very weird socket with a rectangular PGA with two different pin grid densities, one for the CPU and one for the MCM's L2 cache. However, Intel could have done without 370 and 423 and just gone straight from Slot 1 to 478. Slot 1 was originally intended for Intel to be able to put large amounts of cache just off-die on a daughterboard, but later chips had all of the L2 on-die and just used the daughterboard as a BGA mounting place. There weren't many chips that were 370-only- the Tualatin PIIIs were the only ones I recall- so those chips could have been on Slot 1 at the end of its days. The 423 should have been skipped as it was inadequate for the P4, so a Slot 1 -> 478 transition would have been fine.

I only vaguely remember anything about pentium overdrives. I do remember that when the hubub over the socket 5 to socket 7 issue arose, one of the solutions to quell the furor was an adapter that plugged into socket5 making it pentium capable...Again, Im very vague, but was that the P overdrive or something different?
On socket 8, it was short lived and didnt seem to have any significant impact on the market other than to muddle and confuse, adding another short lived socket to the mix, POing at least some consumers in the process. Even back then, I remember not being sure that Intel had any idea where they were going with socket 8.
Concur on the 370/478 transition.

 

[MU_Engineer]

I've never debated AMD's need to shift to DDR2 nor would I. They did need to, at some point in time. If you chose to believe their statements, they deliberatly choose the timeframe rather than just chasing blindly after Intel. If thats true, then, IMO, it was a good choice. But regradless of AMDs descisions regarding the implementation of DDR2 via AM2, the statement was never about whether they needed to go to AM2 (DDR2) or the whys, but that they had stated they would continue supporting 939, then reneged on that statement leaving those who had purchased based on it in a lurch.

Yeah, that did sort of happen. But if AMD had made new 939 chips, the memory limitation would have kept 939 as a mid-to-low-end socket and there wouldn't likely have been anything faster than the FX-60 released for it. That would leave basically the only new chips for 939 as 65 W EE variants of the Mancehster/Toledo/Denmark chips and possibly the 65 nm chips. That would give 939 new chips, but unless you are concerned with power consumption and don't undervolt yourself, there's not much new as far as features or performance. Most of AM2 just duplicates what was in 939 anyway- the only AM2 chips with higher performance than the FX-60 are the X2 5200+ Brisbane, and X2 5400+, FX-62/5600+, 6000+ Windsors. Four chips that are at best 400 MHz faster don't seem like that big of a deal. The Really Big Deal with AM2 is that it supports quads, which there's no way in heck that 939's DDR could. I understand why you feel shafted, but the truth is that you really didn't get shafted all that much. Neither did I as I have an X2 4200+ for 939 as well as 4 GB DDR-400. The $250 in RAM that's not usable in the future kind of sucks, but it'll be far more than obsolete when I replace this unit and as such won't matter much.

Im not sure what you point is here. That Intel deliberately forced a change? If so, that is what it is, but Im not sure how it applys to the original debate that AMD cut the life of 939 short.

I was trying to say that the future upgrade compatibility of 939 was really no worse than their competitor's and it seems like everybody rags AMD about it but doesn't say much about Intel doing the same thing. That may just be perception, but I do hear it a lot.

On socket 8, it was short lived and didnt seem to have any significant impact on the market other than to muddle and confuse, adding another short lived socket to the mix, POing at least some consumers in the process. Even back then, I remember not being sure that Intel had any idea where they were going with socket 8.
Concur on the 370/478 transition.[/quotemsg]

The Pentium Pro was designed to be a server chip and thus wasn't intended to enter the consumer market. The Socket 8 was deigned to handle the MCM L2 cache and Intel decided to axe that for the daughterboard cache on the Pentium II and the PIII Katmai after finding out that the way they made the MCMs was very expensive. Perhaps they intended Socket 8 to be used instead of Slot 1, but ran into trouble. I think trying to kill the Super 7 supply was probably pretty large on their mind as well.

 

[spaztic7]

@ spaztic7 The thread wasn't hijacked but it was threatening to turn into a playground type slanging match.

Oh ok.. I just found it funny how all of a sudden it turned toward a totally new topic.

 

[Mousemonkey]

Oh ok.. I just found it funny how all of a sudden it turned toward a totally new topic.

Yeah no worries mate, the longer you play on these forums the more you notice that these things happen , on topic though I think one of the key things that some, not all but some, contributors to this thread seem to be overlooking is the the term allegations, AMD are alleging that Intel have done wrong and to that end must now provide evidence to the judge backing up their accusations, lets just hope they can do this otherwise they will get one big kick in the corporate goolies.

 

[techguy911]

The problem i see is that amd really needs to have top management shakeup, i can't sell ideas only product i run a computer store and see that amd product sales drop off as far as gamers go, they drive the market none are interested in amd pipe dream of 4x4, not to mention amd wants to change there sockets yet again which means buying a new mother board.

4x4 is too expensive for consumer level applications sure it would be good for server applications, amd has not created a product to compete with core 2 duo and now quad cores.

Example:
Customer is an avid gamer needs cutting edge gaming rig nothing amd has can touch the performance of intel's cpu's, i have tried this already bought best amd had to offer put the system together and customer was not satisfied with the performance, rebuilt changing only mb & cpu and customer was very happy with performance of intel core2 duo.

OEM they will buy what people consider to be fast for gaming purposes its what drives the market since intel beats amd in the gaming market the oem builders will use intel.

Its product and nothing more until amd HAVE a product that can beat intel in the gaming market its game over for amd, ideas are nice show me product, its been 2 years still waiting............