Initial AMD Hammer will have a PR rating of 3400+. I'll take two.

[SexyK]

Originally posted by: MadRat
I'm thinking that AMD can sell their chips at those low margins because of their ability to build so many per wafer.

The real money losers would likely be P4-Celerons, not Durons or Athlons.

The Palamino core is 128mm^2, while the newest Northwoods are 131mm^2, thats not enough of a difference to warrant the huge loss they're taking. Sure, now they've got tbred which is only 80mm^2, but Intel will be on 12" wafers soon, so that's another mitigating factor. There's no two ways about it, AMD should be charging a lot more than they are now if they want to be profitable.

Kramer

 

[MadRat]

The Palominoes seem to be incredibly-shrinking cores, probably smaller now than that original 128mm^2 size. The Palomino-like Duron-Morgan is larger than the Palominoes I've been seeing at the local shop. When Palomino first came out it dwarfed the Thunderbird. Now they seem to be around the same size.

 

[Valinos]

Originally posted by: Zugzwang152
Agreed, but back on topic, I doubt Hammer can get a PR rating of 3400+ to begin with. Their test sample debuted at only 800 Mhz correct? Plus I think moving directly up to 2.6Ghz would hurt their sales. I want to see Hammer at 3000+, and closer in Mhz to the last Barton or T-bred that comes out. At that point, we'll be able to see benchies showing just how much better Hammer is than equivalently clocked t-breds, or a p4 near 3Ghz.

More than likely we'll be seeing (Claw)Hammer before Bartons arrive. I'm also fairly sure we'll see the hammers debut at 2GHz+ to compete directly with the P4. I've heard differing things, but supposedly the Hammer will be able to ramp up in real clockspeed similar to the P4 once they are in full production. Who knows, AMD may be slightly behind Intel in the MHz department for awhile, but offering a 64 bit processor with the x86 architecture is the biggest benefit the chip offers. I could care less how good the 32-bit performance of the chip is. 64-bit may be near worthless to us at the moment, but software will surely take advantage of the larger instruction set and we'll see a bigger benefit in the longrun...and AMD will be leading the charge. I think that gives AMD the biggest advantage. 64-bit may be marketing at first, but so was the original Pentium, MMX, and the Pentium 4.

 

[drogue]

Originally posted by: Rand


As for my opinion of the Model rating- I view it as a necessary evil. Yeah it stinks, and it's far from my ideal measurement of performance.
But at this point it's an absolute necessity.

wow, its about time someone realized that. amd isnt trying to play the clockspeed game, they are simply trying to conform toa system where their processors dont fit.

AMD, Cyrix, IDT etc... in the past they've all tried to teach consumers that MHz does not equate to performance, all attempts have failed.

consumers rarely want to be educated enough to make informed buying decisions. look at cars for example....mustangs are one of the lowest performing sports cars on the market, yet they are perenial record sellers.

Their mainstream AthlonXP 1800/1900+ processors would be laughed off the market if they were forced to compete on the merits of clockspeed alone.

finally, there is some sense here. if you dont like the pr ratings, then ignore it. they arent trying to cheat anyone...and most people that arent educated enough to see through a pr system, prolly wouldnt even ever find out that the chip is running physcially slower than they thought. for that matter, most people dont even know what a Mhz/Ghz is.


you must remember that most of the people in these forums are techies, geeks, dweebs, goobers, engineers, and studs (like me)....we spend our time speculating on the next processor, vid card, monitor...while most people cant remember what brand computer they bought.

In the end, I view the Model rating as an unfortunate necessity.... so long as the Model rating remains realistic compared to Intel's highest performance X86 processor at a comparative MHz/Model I will accept it.
Right now, I do view it as completely realistic. If the day comes that AMD markets a Model rating that is NOT on par with Intel's competing solution then I will be among the first to speak against AMD.

sing it soul-sista!
amen

 

[MadRat]

Hammer will begin its life by competing with Xeon, not P4. *dons flame suit*

 

[drogue]

no......
if you read the anandtech article on the hammer architecture, it says clearly that the hammer will debut as a high end workstation part, not a server part a la xeon. the server parts and lower end hammers are supposed to appear sometime in the future
i dont have a link to that article, but i'm sure someone can find it

 

[MadRat]

Xeons are for workstations, too. Xeons and multi-processor P!!!s are Intel's main professional workstation choices. Intel is really pushing Itanium in the server market.

 

[nmcglennon]

The PR Rating is all a numbers game.

AMD is trying to attain a level amongst common people who only measure processors by "MHz", not their features and architecture. The common user barely knows anything about a CPU's architecture, nor researches it. When they buy something, they go by what they know... in this case, clock speed. Since AMD wants a share of that market, they have instituted this PR system to show that they can attain close to the same performance levels as the competing processor.

Basically with the PR Rating, AMD is trying to compare their apples to Intel's oranges. Now this trick does work for the most part, as I have been caught offguard thinking the by the 2100+'s were 2.1 Ghz and the 1700+'s were 1.7 Ghz, etc.

It isn't that hard to use the PR system when comparing 2 things over a short period of time. It is a terrible idea to use the same PR system, cross platform, over a longer period on more than 2 different processors (ie Athlon, P4, AND Hammer). THAT is when it gets confusing, because you forget what you are actually comparing. There should be some sort of a standard that can adapt to different architectures and different clockspeeds over time. I guess we should just use benchmarks of some sort from now on. I am against the PR System, as it is confusing in the long run. But it is not fair to AMD to lose sales because their chips because their clock speeds are lower when in some cases they can be just as fast. I am a fan of neither AMD nor Intel, all I want is a standardized benchmark system between ALL CPUs over a long period of time. That way we don't have to compare Ghz and Mhz when clock speed isn't everything.

It is awfully hard to change the common people's mind though from what they already know, so (unfortunately) AMD will probably stick with the PR rating for now...

 

[tart666]

I think the time has come to adopt Geforce type nomenclature for CPU ratings:

AMD Athlon Ti/Ultra/Super/Duper
and just keep the fastest chip as Ti, updating every few months.

Works for Apple, right? They have been facing the MHz problem for decades now, so they switched to names and MFLOPs instead of MHz.

Just a quick 2 ¢

Toni

PS: come to think of it, why doesn't AMD use FLOPs instead of their abstract PR rating? GFLOPs are measureable, standardized, and meaningful. Top500.org uses them for the supercomputers for chrissakes. this will be a simple scheme similar to rpm / horsepower / torque set everyone seems to understand. (rpm=mhz, horsepower=gflops, torque=something else)

 

[SexyK]

Seems to me that Xeon's are designed for workstations while the Xeon MPs are more server oriented. I could be way off though.

Kramer

 

[DirtylilTechBoy]

I sell computers, and when a customer asks why the AMD is called an 1800+ when it may only run at 1.4 or whatever, I tell them nevermind the clockspeed its just that the AMD runs about as fast as an 1.8GHZ P4. This may not be totally correct, but for someone who doesn't know a hard drive from a monitor, it gets the point across that even though it may not be as fast as the P4 in terms of MPH it may still be a better chip, which I think it is. Anybody remember those benchmarks anandtech did where the 2 x 1.2 Athlon MP outdid a 2 x 2.2 P4 Xeon?