AMD X399 ！！！！！

[lobz]

Why do you think they keep losing so much of it?

you're looking at the wrong quarter though

 

[lobz]

Why would it not be possible to have 10C and 14C with 2 8C dies in MCM.

10C = (3+3) + (2+2)
12C = (3+3) + (3+3)
14C = (4+4) + (3+3)
16C = (4+4) + (4+4)

I think its definitely possible to have 10C and 14C along with 12C and 16C. The flexibility with AMD's MCM approach is really very good.

exactly, thank you for sparing me the time

 

[lobz]

And again, you cannot have 5 or 7 core configs with Zeppelin.
Both CCXs must have the same number of cores enabled, or the other CCXs must be completely disabled.

and where is it stated that two communicating Zeppelins must be symmetrical with each other too?

 

[Topweasel]

and where is it stated that two communicating Zeppelins must be symmetrical with each other too?

lolzfail already made that point. No one has done two different core count processors therefore lawyered or something like that. I guess because someone hasn't done it already is enough of a reason to say it can't be done.

But I still think a 3+2 3+2 and a 4+3 4+3 Is just as possible and wouldn't break that "rule" inter die CCX communication isn't anything like MCM or second socket communication.

 

[The Stilt]

and where is it stated that two communicating Zeppelins must be symmetrical with each other too?

PPR downcoring section.

 

[Shivansps]

8C/12C/16C, thats what i think AMD HEDT will be... 10 and 14C seems unlikely.

 

[lolfail9001]

lolzfail already made that point. No one has done two different core count processors therefore lawyered or something like that.

I am glad you failed to get my point. No, it is far simpler: it makes no sense to do in the first place.

 

[lobz]

PPR downcoring section.

Thanks

Sent from my VTR-L09 using Tapatalk

 

[Ajay]

This is probably a repost, but for those interesting in the Product Programming Reference, look here: http://support.amd.com/TechDocs/54945_PPR_Family_17h_Models_00h-0Fh.pdf

 

[Ajay]

8C/12C/16C, thats what i think AMD HEDT will be... 10 and 14C seems unlikely.

Yes, if it not divisible by four, it's not Snowy Owl
Lucky I had all those classes in advanced Calculus

 

[formulav8]

8C/12C/16C, thats what i think AMD HEDT will be

Yeps, I definitely expect an 8c/16t. That would be the sweet spot for many peoples.

 

[Topweasel]

I am glad you failed to get my point. No, it is far simpler: it makes no sense to do in the first place.

It does if it allows them a greater spread of price ranges for higher market penetration.

Think of it this way. While I doubt this is true. If it was AMD would have products 4c4t, 4c8t, 6c12t, 8c16t, 10c20t, 12c24t, 14c28t, and 16c32 all with their own rising price range on the same dies. This isn't counting the 4 MCM options. Why not offer a wide array to make sure you have a CPU for every need? Not only that but it would further them salvaging just about any die on the wafer and maximise ASP at the same time.

 

[IEC]

I'm expecting 12c and 16c. Yields would have to be atrociously bad to have a 8c (unless they sacrifice a few 100 chips just to say they have it, which I find unlikely).

 

[formulav8]

Yields would have to be atrociously bad to have a 8c

I'm inclined similar, but there is a strong entry level area for more pci-express lanes and features market to not have an entry level cpu where their mainstream ends/drops off.

 

[Topweasel]

I'm expecting 12c and 16c. Yields would have to be atrociously bad to have a 8c (unless they sacrifice a few 100 chips just to say they have it, which I find unlikely).

If you were AMD and you had the option to sell two sub $200 R5 1500 or 1400's or sell one $500+ and get someone into the HDET platform?

 

[Saylick]

If you were AMD and you had the option to sell two sub $200 R5 1500 or 1400's or sell one $500+ and get someone into the HDET platform?

I was just thinking this too... the nice thing about Ryzen dies is that AMD can either put them into AM4 or X399 SKUs, and like you say, it's more profitable to put them into HEDT where the margins are higher. Who knows? There might be a sizable market for a high-frequency 8C X399 part where a lot of PCIe lanes are required, like basically an R7 1800X but with more L3 cache and PCIe lanes for those who need it.

 

[cellarnoise]

Can a 4 die chip be used in this socket along with a 2 die chip. AMD could salvage crappy ccx complexes into odd ball 10 and 14 chips?

 

[dnavas]

I'm expecting 12c and 16c. Yields would have to be atrociously bad to have a 8c (unless they sacrifice a few 100 chips just to say they have it, which I find unlikely).

The 8C could be useful, though, given extra I/O and memory channels.

 

[Vaporizer]

I think the crappy ccx will be salvaged in r5 and r3 Ryzen processors

 

[Ajay]

Can a 4 die chip be used in this socket along with a 2 die chip. AMD could salvage crappy ccx complexes into odd ball 10 and 14 chips?

No.

 

[Atari2600]

The 8C could be useful, though, given extra I/O and memory channels.

Indeed.

As others have said - some won't need a 12 or 16C CPU, but could do with the extra I/O. It'd particularly make sense if they are sensitive to a few hundred dollars difference in price.

 

[Oddzz]

If these are true and hopefully they wont charge you $$$ like intel does, this just might be my next work station.

However The naming scheme could be abit better.. would have thought something like:

Ryzen 9 1980X, 1970X, 1960X, 1950X sound better than 1998.. almost as bad as Titan Xp.

I agree, your naming scheme sounds better but then you would have the problem with the three 12 core CPUs. You would end up with 1960X, 1960 and then something like 1960X+ when the third number is always equal to 1/2 the cores.

 

[Atari2600]

I agree, your naming scheme sounds better but then you would have the problem with the three 12 core CPUs. You would end up with 1960X, 1960 and then something like 1960X+ when the third number is always equal to 1/2 the cores.

1963
1962
1961

Done.

I don't give two sh*ts how a number "sounds". I want it to clearly tell me what I'm looking at.

1 = first gen Zen
9 = Ryzen9 series
6 = 12 cores
1/2/3 = clock tier of this CPU

 

[DrMrLordX]

you're looking at the wrong quarter though

Okay, which quarter past or present (no, Q2 2017 doesn't count) should I be looking at here? I may be one of the biggest AMD fanboys on this forum in terms of what I will buy or promote, but even I have to admit that they are still losing money. That's why they took at hit to their stock price after reporting their Q1 2017 earnings.

Apparently my decision to buy an 1800x wasn't enough to get them into the black. Guess I'll have to buy a Vega too.

 

[Atari2600]

I may be one of the biggest AMD fanboys on this forum in terms of what I will buy or promote, but even I have to admit that they are still losing money.

Your not wrong.

But - with the investments AMD should be putting out that this point to get Naples to market in fine fettle, they still *should* be losing money. Q4 reports of this year should be more reflective of where Zen has landed them.