[Digitimes] AMD updates product roadmap for 2014 and 2015

[MisterMac]

Yeah me too. This way there is at least a chance they may actually deliver something I might want to consider purcharsing.


Aren't Excavator modules / cores are supposed to be quite diffrent from Buldozer&Piledriver and be more like "full true 2 cores per module" instead of this "~1,5" per module we have now?

Better hope so - i'm worried were beyond 2016+ before intel goes mainstream Hexacore.

They're not gonna push it before HW-E and 8 core extreme atleast.


If AMD improved ST performance and throughput on quadcore it will put pressure right where we need it for intel not to forget us enthusiast.

Well... "Cheap ass" enthusiasts as Shintai would say

 

[guskline]

Im with ya but, im expecting at least one more FX AM3+ product.

High end , mid or low end?

Hard to see anthing faster than the 9590 out of a PileDriver core. Looks like most if not all of the Steamroller technology has gone over to the FM socket.

I would love to see it for my rig 2 below but the longer this drags on the less likely I see any new cpu being released for the AM3+ socket other than "souped up" PileDrivers (9370/9590).

 

[guskline]

Better hope so - i'm worried were beyond 2016+ before intel goes mainstream Hexacore.

They're not gonna push it before HW-E and 8 core extreme atleast.


If AMD improved ST performance and throughput on quadcore it will put pressure right where we need it for intel not to forget us enthusiast.

Well... "Cheap ass" enthusiasts as Shintai would say

Hard to call anyone now buying a FX 9590 or even a FX 9370 "cheap ass":awe:

 

[monstercameron]

amd should definitely keep going with their bd/pd plan, why compete with intel on st when the can more cores at the problem...I say MOAR COARS, they should keep going until the cores edge out intels ST ipc lead.

 

[Homeles]

Because there isn't a demand for it.

/argument

 

[Ajay]

amd should definitely keep going with their bd/pd plan, why compete with intel on st when the can more cores at the problem...I say MOAR COARS, they should keep going until the cores edge out intels ST ipc lead.

I assume you are joking, or haven't been paying attention to how well that worked for AMD w/BD&PD.

 

[Abwx]

Because there isn't a demand for it.

/argument

That s a void argument and total non sense when looking
at the numbers which says FXs account for 30% of their
CPUs/APUs total shippements , so there s a demand for them
whatever your hopes...

 

[Phynaz]

I assume you are joking

I don't think he is.

 

[frozentundra123456]

That s a void argument and total non sense when looking
at the numbers which says FXs account for 30% of their
CPUs/APUs total shippements , so there s a demand for them
whatever your hopes...

AMD has what 20% of the cpu market. So I wouldnt call 30% of 20% an overwhelming demand. Plus they are selling a big die dirt cheap. I actually would like to see AMD continue FX, because it might eventually pressure Intel to bring out a mainstream hex core, but economically I am not sure it is easy to justify the development and production costs, except they have to have something to manufacture at GF.

Edit: The costs of development are even harder to justify because they have very little presence in the high margin server market, and seem to be emphasizing APUs and ARM for servers as well.

 

[AtenRa]

High end , mid or low end?

Hard to see anthing faster than the 9590 out of a PileDriver core. Looks like most if not all of the Steamroller technology has gone over to the FM socket.

I would love to see it for my rig 2 below but the longer this drags on the less likely I see any new cpu being released for the AM3+ socket other than "souped up" PileDrivers (9370/9590).

Same space they occupy today

 

[Abwx]

AMD has what 20% of the cpu market. So I wouldnt call 30% of 20% an overwhelming demand. Plus they are selling a big die dirt cheap. I actually would like to see AMD continue FX, because it might eventually pressure Intel to bring out a mainstream hex core, but economically I am not sure it is easy to justify the development and production costs, except they have to have something to manufacture at GF.

Edit: The costs of development are even harder to justify because they have very little presence in the high margin server market, and seem to be emphasizing APUs and ARM for servers as well.

Cost is about 40$ per chip , still better in margins
that any APU.

30% of 20% is 6% of the market , that s considerable
given that it s for a single die and very few SKUs.

 

[monstercameron]

I assume you are joking, or haven't been paying attention to how well that worked for AMD w/BD&PD.

Here is how I see it, keep going or start over/redesign[which they would have done if they could]...at this rate which is more feasible? besides, we know it isn't because the chips are slower in ST perf. why they aren't selling all too well [in a blind test could you tell the difference between a i5-4670 and a 8350], it all comes down to oems and SIs...without there support amd doesnt stand a chance.

 

[guskline]

Same space they occupy today

Thank you AtenRa. From reading all of your posts, you appear to be tied into AMD so I can appreciate the implication of a NDA.

Let me say as an AM3+mb owner your response gives me hope and meaning to the phrase "Hope Springs Eternal"()

 

[Homeles]

That s a void argument and total non sense when looking
at the numbers which says FXs account for 30% of their
CPUs/APUs total shippements , so there s a demand for them
whatever your hopes...

Last I checked, AMD only had 16.5% of the market, which is next to nothing. 30% of nothing is still nothing.

Since you like to stretch the truth (i.e., flat out lie about AMD's market share percentage), I'm sure you'll have no problem with my math.

 

[AtenRa]

Last I checked, AMD only had 16.5% of the market, which is next to nothing. 30% of nothing is still nothing.

Since you like to stretch the truth (i.e., flat out lie about AMD's market share percentage), I'm sure you'll have no problem with my math.

We are talking about DESKTOP market share, not total market share.
~2Mil AM3+ CPUs per quarter is not a small number.

 

[Abwx]

Last I checked, AMD only had 16.5% of the market, which is next to nothing. 30% of nothing is still nothing.

Since you like to stretch the truth (i.e., flat out lie about AMD's market share percentage), I'm sure you'll have no problem with my math.

The 20% was from Frozentundra who i was responding to ,
sqo much for you deffamation and lies accusation that straight
backfire to you along with YOUR stretching.

30% of 16.5% is a lot for a single reduced line of 3 SKUs ,
that makes 5% of the whole market , far from being nothing
unless one is maths defficient....

 

[Abwx]

We are talking about DESKTOP market share, not total market share.
~2Mil AM3+ CPUs per quarter is not a small number.

Actualy it s 5% of the whole CPU market ,
about 4.5millions/quarter.

 

[ShintaiDK]

Actualy it s 5% of the whole CPU market ,
about 4.5millions/quarter.

You count mobile+desktop. If 30% is AM3+ and 70% is FM2. How much is then Brazos, Kabini, mobile chips and so on?

I doubt AMD sells more than 6 million desktop CPUs per quarter. It was a tad over 7 million in 2012. But amount shipped is lower, plus more CPUs naturally turn to mobile.

No wonder AM3(+) is dead with those volumes.

 

[Centauri]

Last I checked, AMD only had 16.5% of the market, which is next to nothing.

LOL. It's more than Toyota's share of the US car market.

 

[Abwx]

You count mobile+desktop. If 30% is AM3+ and 70% is FM2. How much is then Brazos, Kabini, mobile chips and so on?

I doubt AMD sells more than 6 million desktop CPUs per quarter. It was a tad over 7 million in 2012. But amount shipped is lower, plus more CPUs naturally turn to mobile.

No wonder AM3(+) is dead with those volumes.

I count all production , FM2 and Kabini should be
more or less evenly distributed with their remaining 70%.

 

[nemesismk2]

So...APUs for all segments of the PC market?

I want an AM3+ Steamroller!!!!!!!!!

I want an AM3+ Steamroller too

 

[guskline]

I want an AM3+ Steamroller too

If all of us who want a FX Steamroller for our AM3+ mbs sing along with the link attached perhaps AMD will listen!:biggrin::awe:
http://www.youtube.com/watch?v=9VXr6U-BwMs

 

[krumme]

Why are they postponing the desktop Kabini?

Do you think of any other reason than the wsa? I cant...

This is just proof gf process is bad imho. Its good economics to produce huge rl dies here because nobody else will pay for it. A serious management issue this pathetic irrational show can continue.

When does mubadala think its enough. Who actually knows what is good business here anymore?

 

[Homeles]

LOL. It's more than Toyota's share of the US car market.

Toyota has far, far higher margins than AMD.

The 20% was from Frozentundra who i was responding to ,
sqo much for you deffamation and lies accusation that straight
backfire to you along with YOUR stretching.

30% of 16.5% is a lot for a single reduced line of 3 SKUs ,
that makes 5% of the whole market , far from being nothing
unless one is maths defficient....

Let's see: 5% and 95%; which is closer to 0?

And who are you to say that there's a demand for consumer-grade processors with more than 8 cores and garbage IPC? AMD doesn't even think so.

 

[Idontcare]

When does mubadala think its enough. Who actually knows what is good business here anymore?

The fact that Ruiz had them so easily and fully duped into thinking they were in a bidding war with some anonymous 3rd party says a lot of Mubadala's business prowess and savvy IMO.

That they continue to bring that business savvy into the decision making process post-acquisition is of no surprise, but does go some distance to explaining the realities of GF's 28nm timeline and customer situation.