Barton to have 128k L1, 512 L2 cache...no faster FSB

[ST4RCUTTER]

http://www.theinquirer.net/


I like most people here would have preferred a faster FSB matched to current DDR standards as opposed to more cache. One thing should be true though, with so much more cache the die should be a good deal larger than the Tbred, and able to dissipate heat more efficiently. Does anyone know when AMD will begin to use Si28? Supposedly it allows for greater heat transfer from the lower substrate levels to the thermal interface w/the heatsink.

 

[Zugzwang152]

this is the inquirer? why are we believing it this time where they've blatantly lied before?

 

[ST4RCUTTER]

Well, besides the fact that every other computer hardware site is saying the same thing, and the fact that the Inquirer quoted an AMD employee...Mike has proven reliable as much as he's been wrong. And when he's wrong he admits it and corrects his statements. Call it what you will...I call it decent journalism.

 

[Zugzwang152]

All right, as long as you're not basing it solely off of the inquirer. What other sites did you get your info from? I'd like to take a read at a more dependable source.

 

[Adul]

I call it a rumor mill

but it is nice to hear it never the less

 

[mfavin]

The Inquirer is pretty reliable.

The real question though is whether AMD will survive much longer. Their revenue is dropping by 1/3 this quarter alone, from an expected $900 million to $600 million. That is a MASSIVE hit. How can AMD keep financing the development of the Clawhammer, etc when their revenues are dropping so dramatically and unexpectedly?

The Athlon appears to be in deep trouble, unable to match the Pentium 4 anymore. Nobody knows right now why AMD can't keep up with Intel right now. Is it problems with the .13 process or is the Athlon architecture hitting its maximum speeds? If the architecture is running out of room, AMD is in real trouble, and their entire existence is bet on the Clawhammer coming to market successfully and on time.

Intel is rumored to have moved up the 2.6 AND 2.8 Ghz P4 release to mid August now. The P4 2.26 will probably be reduced to <$150 when this happens. If AMD doesn't have 2400+, 2500+ processors out soon, they will be selling Athlons at a loss.

 

[Zugzwang152]

You do have a point there, mfavin. Is it possible Intel is seeking to snuff AMD out right now? I don't see Hammer having a big as impact as the Athlon did when it first came out. AMD won't be grabbing acres of market share at a time this round. If anything, they've proven how unreliable they are.

This of course, will not stop me from supporting their efforts and buying whatever they have, provided it's cheaper and still at a decent performance level

 

[Ben50]

The real question though is whether AMD will survive much longer. Their revenue is dropping by 1/3 this quarter alone, from an expected $900 million to $600 million. That is a MASSIVE hit. How can AMD keep financing the development of the Clawhammer, etc when their revenues are dropping so dramatically and unexpectedly?

I wouldn't worry about AMD "surviving" as you put it. They have been around for a long time and will continue to be around. As far as financing, there are these things called loans that companies can get that magically get them more money. Sorry for being so sarcastic, but come on, you don't actually think they will go out of business or lose financing because of one bad quarter. They have a long track record of being a successful company so lets be a little more realistic here.

AMD knew it would be a bad quarter for them from the beginning. They have been somewhat stagnant on releasing new speed grades lately so its no big surprise that their ASPs are down and so are sales. They know that the Athlon is nearing its end, so they are trying to hold on as long as possible while putting most all of their resources into hammer development.

As far as the FSB improvements, there have been benchmarks done that show very little increase in performance, on the order of about 2-3%. It wasn't worth bothering with defining a whole new standard for motherboards when the gains would be so small. By improving cache sizes, they maintain backward compatibility and will continue to have good speed.

 

[Soccerman]

AMD's been through alot worse.. they ended up selling their K6-2 400 CPUs for extremely low prices becuase they over-estimated demand and overproduced it. they had to clear out warehouses of these things to make room for the original Slot-A Athlon. They're in much finer condition now than they were back then.

now, as for Barton, I expect that won't be the only modification to the core (cache increases that is). I would hope they move to a 256 bit bus to the L2 cache as well as other modifications. we all knew that AMD said it won't move to a faster FSB for the K7 core simply becuase they really want their Hammer to look that much better. hopefully we'll have chipsets available to us that will allow for perhaps a 200mhz fsb when the Barton comes out (boy that release date is coming up soon).

 

[MadRat]

Lets just get Barton to market before we jump all over AMD's decision to stick to the 133fsb.

Seriously, folks, the 133fsb is much less of a bottleneck than the 64-bit path to external memory.

 

[Zugzwang152]

When was the official release of barton?

 

[glugglug]

Bigger cache will make a much larger difference than a faster FSB in probably > 85% of all applications.

Personally I'm surprised they never made a 1MB cache version. That was around back in the Pentium MMX/K6 days, why did it go away? (except high end Xeons)

 

[ST4RCUTTER]

When was the official release of barton?

The roadmap says 2nd half of 02'. My guess is end of third quarter-beginning of the fourth. The thing is, Barton will be made exclusively at UMC so it depends on how well their .13um process is ramping, and how high their yields will be.