AMD Roadmap

http://www.c627627.com/AMD/Athlon64/
http://www.c627627.com/AMD/AthlonXP/

(link dead, being fixed)

Thanks bro, I was just gonna start a thread asking about the upcoming FX chip ..

very nice! Q1 2004 at 2.8Ghz...

I wonder how soon Intel's going to be above 4Ghz...

shiyan said:

very nice! Q1 2004 at 2.8Ghz...

I wonder how soon Intel's going to be above 4Ghz...

Click to expand...

Where did you see the Q1 2004 thing?

on the chart, under San Diego:

Doh, I guess I'm so tired I don't read the "smaller letters"
Thanks

The dates are a bit weird, how come the FX57 is supposed to be released Q1 of 2k4, and the FX53 only in Q2. While the FX55 is only known to be released in 2k4, nothing more than that?

What's up with the "Athlon XP 2800+ Thoroughbred" rated at 2.25 GHz stock?

Have any of these been seen in the wild? This processor would be the highest MHz clock processor that AMD currently sells - making it strange that they PRed it @ 2800+, when the Barton 3200+ stocks out at 2.2GHz. Do the red letters mean that it was never released?

Man, is it ever nice to see this chart .

Now if only someone could get a complete stepping chart with what each stepping "means" straight from AMD!

2800+ does seem a little low for a 2250Mhz Tbred... but I guess AMD is trying to adjust its ratings now that P4's are running with 200fsb and have dual DDR?

That doesn't look like it is a real AMD document. You can see what the offical AMD roadmap looks like here: http://www.amd.com/us-en/Processors/ProductInformation/0,,30_118_608,00.html


This looks like compiled data and some educated guesses. But it is missing some cores that are listed on AMDs offical roadmap from May.

The 2800+ Thoroughbred chip was nowhere to be found for a long time after its announcement in Oct 2002. Word was that some OEMs got it and that it made its way into retail desktops in late 2002/early 2003. It may have showed up briefly in retail for hundreds of $ at the very time 2800+ Barton was announced in Feb 2003.

2800+ Barton is the 2800+ that everyone has for sale. It’s hard finding a 2800+ Thoroughbred even on eBay.

There are no other Desktop cores.

Of course all future dates are speculation, if you have a problem with any of them, clearly state why and if your view makes sense, future dates will be changed.

The Coolest, yes i noticed that was a little messed up. I think its a typo, and they most likely mean Q1 2005 and not 2004. That would be great if it was 2004 but IMHO highly unlikely. And im ALMOST completely sure becuase according to the chart, the FX-53 comes out a whole Q later than the FX-57, and AMD would not do that. Unless its one of those things where they come out with the fast procsessor and sell for really expensive and later come out with lower models selling for cheaper.

KCT2: its not the official AMD roadmap, butI almost trust it. Its from some other website, i used to know the name but I forgot. From the begining these roadmaps have been pretty acurate. I pretty much trust them.

2800 tbredB: Yes, i do remember the 2800 coming out and being a tbred, and yes as i recall it ran at 2.25 Ghz. I think someone here actually might have had one, but soon after it came out, the model 10's, "barton" was released with the 3000,2800,and 2500 all at once as i recall. Im not 100% sure though.

-f1

Just so he doesn't have to make himself look like an a$$,

KCT2: its not the official AMD roadmap, butI almost trust it. Its from some other website, i used to know the name but I forgot. From the begining these roadmaps have been pretty acurate. I pretty much trust them.

Click to expand...

c627627 made this roadmap himself, and the website you saw it on was his.

It doesn't appear that the San Diego will be coming out for a while now; more like Q3 2004. http://www.overclockers.com/tips00478/
Unfortunately, I think that Ed is right here. A die shrink in Q1 doesn't make much sense for anyone other than overclockers.

There are no other Desktop cores.

Click to expand...

True - but you included the sledgehammer (a server/workstation core) in your list, so it didn't seem unreasonable to expect future server/workstation cores to be there also. The Athlon64-FX's may be marketing the sledgehammer core in the desktop market, but they were not intended for it, AMD's own roadmap, and its design (Registered RAM especially) show that - but that gets to a bigger point. What good is an AMD roadmap? They change what they are going to do from day to day to try and keep up with intel.

I have no problems with any of your dates (except Q1 2004 for 90nm San Diego, just doesn't make sense, but that doesn't mean it isn't true I guess), but what is the point? AMD is careful to avoid officially committing dates, and regularly change product release, and marketing. AMDs success in the business sector can be closely tied to a lack of confidence in their roadmaps, and the fact they don't follow what they do publish. Maybe you have a magical crystal ball (or a well placed AMD source), and can see what will actually happen, but if this is just AMD releases and rumors put together into a pretty graphic, I don't put too much faith in it.

c627627 this isn't an attack on your work, I'm not saying you didn't put this together with the best possible information, probably information most people/websites don't have, but AMD doesn't have a history of reliability in this area. I can't give them the benifit of the doubt, I don't believe they know what they will be doing a year from now, when that has never been the case in the past.

I HOPE I'm wrong, I hope this isn't the case, I would love to see AMD hit all their release dates at the speeds promised. Then they would be able to build a larger market share and be able to be proactive towards intel in the future, instead of reactive like they have been in the past. Sledgehammer/Clawhammer are a good step in the right direction - but that is a mess also because of the number of sockets - , finally Intel is being reactive with the extreme edition P4, i doubt we would have seen that if they weren't concerned about Athlon 64-FX. But AMD needs to prove then can produce them, and make people confident in the company.

The Q1 San Diego date was changed to Q3 the moment Gautam posted his suggestion.

I missed that one. I used to read Stroligo's articles religiously but there was so much negative energy in them lately that I started skimming through them and so missed the Q3 prediction.

SledgeHammer replaced ClawHammer after extensive debate with member OC Detective who proved to me that Socket 754 Athlon 64 3200+ was NOT Paris and consequently Athlon 64 FX-51 was not ClawHammer.

AMD unexpectedly made a server chip into a desktop chip: There is no difference between the 940 pin Opteron 1xx and the 940 pin Athlon 64 FX.

Athlon 64 FX-51 is a de facto SledgeHammer.


What good is an AMD roadmap? There are virtually no clear AMD CPU tables being posted on any web sites discussing CPUs.

The map changes all the time. OC Detective and Gautam pointed out things that were wrong and they were changed.

What good is an AMD roadmap, you say? What good is the weather man, I say.

holy crap these are your roadmaps? OMFG!!! OMG!! I never knew that!
Awsome work! I have that website bookmarked on my dads comp i think. They ceratinly give alot more info than the AMD ones.

I had one question though, I am really quite confused as to what paris and victoria are. Just the A64 but no 64bit? and cheaper? Is this kinda gonna be like a model inbetween the duron and the A64? If so, I think its a smart move for amd becuase they will have a processor that fits just about every price range, the duron, this socket 754 XP, the A64, and the FX.

As for the 2800 tbred B, i have a pic from [H] of one.

EDIT: this is as close to posting a pic i can get, dumb macOS9.2

http://www.hardocp.com/image.html?image=MTAzMzQ0MTY1M0lobFlKZWZKRWpfMV8yX2wuanBn

What good is an AMD roadmap? There are virtually no clear AMD CPU tables being posted on any web sites discussing CPUs. That's frankly mind boggling to me. The map changes all the time. OC Detective and Gautam pointed out things that were wrong and they were changed.

Click to expand...

That's the problem with AMD roadmaps, they are constantly changing, how can you make decisions when the information is always changing?

I'll say it agian, I was not attacking your work - it's good someone is doing it, because God knows AMD doesn't publish reliable information. I was more of mentioning the futility of it, and you said it yourself, "THE MAP CHANGES ALL THE TIME", is that the way it should be? Obviously there are good reasons for changes, but does AMD ever provide these reasons?

AMD unexpectedly made a server chip into a desktop chip: There is no difference between the 940 pin Opteron 1xx and the 940 pin Athlon 64 FX.

Click to expand...

I know, but it was not meant to be that way. It is just another example of AMD being unpredictable, therefore making maps somewhat useless.

What good is an AMD roadmap, you say? What good is the weather man, I say.

Click to expand...

At least the weather man knows what city is under the cloud. The sky may be ever changing, but the MAP of the ground stays the same.

well, once it changed and announced then he goes and changes the roadmap. like look at it now, all the info is correct IMHO, except for the future coming CPU's but we will never know exactly until that happens, but the map gives us the best idea of what is coming. And except for maybe the name, its all been pretty good so far. IMHO

Keep up the good work c627627!

modenaf1 said:

....As for the 2800 tbred B, i have a pic from [H] of one.

EDIT: this is as close to posting a pic i can get, dumb macOS9.2

http://www.hardocp.com/image.html?image=MTAzMzQ0MTY1M0lobFlKZWZKRWpfMV8yX2wuanBn .....

Click to expand...

Thanks for the picture. Weird. I wonder how well they clock

(seriously, that is the first thing that I thought when I saw that picture - followed by "[H]? yuk." )

It has a weird stepping:

AXDA 2800 DKV3D AIUAB 0231 MPM

The bolded parts are the weird parts... I know that the AIUAB means that it is a T-Bred "B", but the DKV3D part, which I assume is related to the stock Vcore, I am curious about. Does anyone know what the stock Vcore would be by looking at this stepping?

c627627 - Keep up the good work, I think that this evolving roadmap idea is a very good one. Have you considered adding in all the known steppings, just for eye-candy ? It is much better than anything else I have seen, and of course it will be constantly changing as more information about unreleased products comes to light! Would you (kct2) rather have nothing to look at? If you don't like it, than don't look at it, and just pretend that it wasn't made. Or make your own. No insult intended.

Would you (kct2) rather have nothing to look at? If you don't like it, than don't look at it, and just pretend that it wasn't made. Or make your own. No insult intended.

Click to expand...

First of all, no insult taken. I just think I am being misunderstood.

I think it's great that he is putting these roadmaps together.

BUT...

The problem is on AMD's side, they NEED to produce a roadmap and stick to it (within reason). If they ever want to be taken seriously in the business market they will need to do that. What I want is an official roadmap that doesn't change every week, doesn't have new things coming out of left field, doesn't have products moving into different market segments, and most importantly, something that AMD can stick to and customers can rely on.

Is that unreasonable?


AXDA - Standard Athlon XP Model 8
2800 - 2800+
D - Organic PGA
K - 1.65v
V - 85c
3 - 256kb
D - 333mhz fsb

Kool - thanks alot for the stepping breakdown! Where did you learn the meanings of the letters?

And yes, your point is not unreasonable at all - and I see where you are coming from now. Unfortunately for us all educated predictions like this are all we can really claim to have. I agree that AMD needs to solidify their production plans.

Sorry if I sounded rude there, it wasn't worded very politically I know.