When will AMD take the crown again?

[BitByBit]

Probably never. The problem with always being atleast one manufacturing process generation behind is that they can never hit the clock speeds necessary to match or exceed Intel, certainly at acceptable TDP. When Deneb is released, AMD should have a chip that performs nearly as well as Penryn (though perhaps not as clockable), but Intel will be fielding Nehalem at that time. As I see it, this is how the future will pan out for AMD: always a process generation behind and only competitive with Intel's previous architecture generation.

Their graphics division is, thankfully, a different story.

Originally posted by: v8envy
The next shot at the high end is with Bulldozer, some time in 2009 or 2010.

I wouln't be surprised if Bulldozer has been canned entirely. Even if it hasn't, it certainly won't see the light of day before 2011.

 

[Keysplayr]

Originally posted by: BitByBit
Probably never. The problem with always being atleast one manufacturing process generation behind is that they can never hit the clock speeds necessary to match or exceed Intel, certainly at acceptable TDP. When Deneb is released, AMD should have a chip that performs nearly as well as Penryn (though perhaps not as clockable), but Intel will be fielding Nehalem at that time. As I see it, this is how the future will pan out for AMD: always a process generation behind and only competitive with Intel's previous architecture generation.

Their graphics division is, thankfully, a different story.

Originally posted by: v8envy
The next shot at the high end is with Bulldozer, some time in 2009 or 2010.

I wouln't be surprised if Bulldozer has been canned entirely. Even if it hasn't, it certainly won't see the light of day before 2011.

Manufacturing process has little to do with attainable clock speeds. This depends entirely on the architecture. If you recall, the fastest Pentium 4 reached 3.8GHz (570j. I had one of these.) on a 90nm process.
Doesn't really matter how bad or good it performed. It reached those clocks because of the architecture.

 

[Nemesis 1]

Originally posted by: KingstonU

Originally posted by: taltamir
When I saw the charts describing larabee and its future the first thing I said "hey, that is fusion"... the second thing I said was "hey this is gonne be here a long time before AMDs fusion"

Yah It's ironic that AMD brags about a great new concept that they are working on, and from hearing about it Intel decides they need to be copying this idea too and throw their Massive R&D budget at it and actually get to produce and sell it before AMD gets too. Yet it was AMD's idea... crazy...

You should try a little research. Intel had 3 projects going for GPU's at the same time . The winner of the 3 was Larrabee. It was in development way befor AMD bought ATI or announced or fusion.

 

[Nemesis 1]

Originally posted by: taltamir

Originally posted by: harpoon84
The problem for AMD is that they simply cannot survive on 'value' CPUs alone - unless they restructure and become a '2nd tier' CPU manufacturer like VIA, which would mean drastically cutting down R&D costs and would basically make Intel a monopoly in anything but the absolute low end.

well... the chinese are about to unveil their own x86 cpu....

Go dragon go. Lets see who is in second after 09. Sales volumn. AKA Turtle 1

 

[Idontcare]

Originally posted by: keysplayr2003
Manufacturing process has little to do with attainable clock speeds. This depends entirely on the architecture. If you recall, the fastest Pentium 4 reached 3.8GHz (570j. I had one of these.) on a 90nm process.
Doesn't really matter how bad or good it performed. It reached those clocks because of the architecture.

Manufacturing process (i.e. process technology) determines your absolute max operating speed capability versus a given power consumption.

Architecture determines how close the final product comes to reaching that absolute max allowed or provided by the processes in manufacturing.

 

[taltamir]

you know, i wonder if the chinese will actually be selling their CPU outside of china...

 

[Sureshot324]

Intel is on a roll right now, so AMD needs to cut R&D and just focus on budget CPUs for now. It worked fine for them before the first Athlon came out, and they're in a better position now then they were then because of all the OEM deals they have. Go to any major OEM's site other then Dell and a lot of the machines will have AMDs. The debt from purchasing ATI is a problem, but I think this is finally starting to pay off. ATI is fully capable of competing with Nvidia, they just fell behind because they devoted so much resources in making a custom GPU for the Xbox 360, whereas Nvidia just put a plain 7800GTX in the PS3.

AMD should just hang back and wait for Intel to drop the ball again (and they will).

 

[Idontcare]

Originally posted by: Sureshot324
Intel is on a roll right now, so AMD needs to cut R&D and just focus on budget CPUs for now. It worked fine for them before the first Athlon came out, and they're in a better position now then they were then because of all the OEM deals they have. Go to any major OEM's site other then Dell and a lot of the machines will have AMDs. The debt from purchasing ATI is a problem, but I think this is finally starting to pay off. ATI is fully capable of competing with Nvidia, they just fell behind because they devoted so much resources in making a custom GPU for the Xbox 360, whereas Nvidia just put a plain 7800GTX in the PS3.

AMD should just hang back and wait for Intel to drop the ball again (and they will).

AMD was uniquely blessed with being able to pick up the NextGen team and their early K6 design (helped them compete with the Pentium and the P2) in addition to DEC falling apart and making available numerous key talent for hire (such as Dirk Meyer, their current CEO and lead architect of the K7).

In other words the saving grace for AMD in the K5/K6/K7 era was that they acquired pre-existing talent thru M&A (nexgen) and hiring design teams in block (DEC and Meyer's group). Where are they going to acquire that talent in 2010? IBM is a viable possibility (if they folded their Power7 and Power8 teams) I suppose, but the bottom line is that AMD doing a re-do of their prior business cycles will be difficult because the semiconductor industry is by no means similar to those by-gone days.

 

[taltamir]

right... and fireing more R&D people is the worst thing to do right now, they actually need to hire more of them.

 

[Idontcare]

Originally posted by: taltamir
right... and fireing more R&D people is the worst thing to do right now, they actually need to hire more of them.

Well I don't know AMD's HR policies in specifics but having come from an R&D team in the same industry I can say that it is healthy to have some kind of a purge process so you do get rid of those individuals that are actually the 1 in 10 people in the room that derails progress meeting after meeting and just doesn't get jack done on their own either (the Wally's of the work group).

It definitely did not hurt Intel to purge 10k employees back in 2006 (or was it 2005?).

If it is a poorly implemented HR policy though it can have the undesired effect of deflating moral in the workforce that was never under consideration for layoffs.

I don't personally know any of the people who were let go by AMD recently, but I do personally know many (talented) R&D folks who are still there so it can't be all that ridiculous a layoff policy they have been implementing.

Consider that at current going rates it AMD only gets about six (6) R&D engineers/designers for $1M/year of cost (salary/taxes/401k/bonus/options/etc). With that kind of economics working against you in this industry it doesn't take long to realize you can't just throw 50-60 engineers at 4yr long R&D projects willy-nilly, especially if you are losing money and waaaaay in debt.

It's a downward spiral. Texas Instruments (my former employer) elected to get out of R&D because of the financials and they were no where close to being in AMD's predicament.

 

[Dazed and Confused]

It should be illegal to post "What if" or "which is better" threads when AMD and Intel are in the same sentance.

 

[fire400]

well back in 2007, Intel claimed years before that it would release 3D motherboards, named, Trigate.

pci-SIG will also be advancing with JEDEC DDR-4 desktop solutions in motherboards, making AMD shift gears from it's late DDR-3 boards and SCSI express should be an interesting picture to look at if SATA 600 doesn't come out fast enough, but Intel with it's SSD's might make both future HD projects go bankrupt.

the Pentium 4 wasn't as fast as Athlon 64's, like Pent. D and X2's compared again, but then again, Intel still outsold AMD because it had better marketting, way more companies that just stuck it out with Intel. but that might change as pricing becomes more competitive. USA's economy is suffering badly, so anything that's cheaper and still decently fast, will do for just about anybody. as far as the crown goes for processors, AMD is still only worth around 5 billion, prolly a bit more with the ATI merge, but Intel could buy them out any day, if the US gov. let's them, that is.

 

[XBoxLPU]

Originally posted by: Idontcare

Originally posted by: bX510
What do u guys think?

Innovation takes money, and in the semiconductor world it takes billions upon billions which AMD simply doesn't have

And this is where I strongly agree with you. Intel's Dunnington is an example of where Intel is ahead of AMD in manufacturing process. They increase at least 10% to 20% better performance per core (upgraded core) , two extra cores per CPU package, and an upgrade that only requires a BIOS update. It has a massive die size and a great increase in transistor count but it actually pulls less power than its predecessors:

http://it.anandtech.com/IT/showdoc.aspx?i=3414&p=9

Intel has been moving strongly to lower power consumptions on their CPUs almost parallel to increasing performance along the way. AMD is still stuck with higher power consuming CPUs and slower performance.

 

[Gillbot]

I hope AMD gets the ball rolling myself, I want cheaper CPUs! Intel charges a premium because they know they can. AMD keeps cost down because they must if they want to compete.

 

[OCGuy]

Originally posted by: Gillbot
I hope AMD gets the ball rolling myself, I want cheaper CPUs! Intel charges a premium because they know they can. AMD keeps cost down because they must if they want to compete.

Actually Intel has been surprisingly kind to us relative to what CPUs used to cost.

Sure, they have a $1499 SKU. But the great thing is, you can buy a $180 CPU and beat out the $1499 CPU with a little knowledge of BIOS.

Hell, a $130 dual E7XXX OC'd can beat out a stock 9770.

 

[Gillbot]

Originally posted by: Ocguy31

Originally posted by: Gillbot
I hope AMD gets the ball rolling myself, I want cheaper CPUs! Intel charges a premium because they know they can. AMD keeps cost down because they must if they want to compete.

Actually Intel has been surprisingly kind to us relative to what CPUs used to cost.

Sure, they have a $1499 SKU. But the great thing is, you can buy a $180 CPU and beat out the $1499 CPU with a little knowledge of BIOS.

Hell, a $130 dual E7XXX OC'd can beat out a stock 9770.

Oh I know, but with more competition comes better products and everyone will benefit then.