It is more detailed than what we heard before. It's from last week

And the question is the obvious one,

Does AMD have any choice? FX is the CPU they have. FX-8150 is fast but uses a lot of power to compete with an i5-2500K; and it is finally priced the same at around $200

Well, it's going to take a complete re-design of Bulldozer for AMD to be competetive again with their CPUs. It's going to take them *years* to catch up, and wasting resouces on Piledriver will only lengthen the agony for them.

It's so bad right now that I don't think AMD will ever catch up to Intel again. Intel will definitely surpass AMD's integrated graphics, though, IMO. They'll probably do it with Haswell at which point there will be no reason to go with AMD unless their chips are significantly cheaper.

Well, it is in AMD's competitor's best interest *not* to crush them. i really didn't realize this until recently.
--AMD's competitors actually want AMD to survive.

Why is that? So the government stays off their back regarding intel being a monopoly?

Every reason to have a competitor. A monopoly is not healthy when it comes to governmental oversight and regulation.

Not only that, if you have a competitor, you can point to how much better your own products are in comparison.

Yeah it makes sense.

As already stated, its not like AMD has a lot of choice. What they are doing is backing out of the high end. PileDriver will not pile drive. Most of AMDs resources is going to integrating and slim downed CPUs. Even though this has been part of their focus, even a big part, its moved up to be their only objective. I suspect it all pivots on the bulldozer.

As much as fusion was touted by AMD with its importance to their overall strategy, the fact that bulldozer was released without it is a serious eye opener. It was a big deal. Just a few years after AMD announced their fusion future and its competitive edge intel had stepped up and was all over it. Before AMD had any of these type chips intel was putting a gpu and cpu on the same socket. In well executed steps, Intel has been successfully merging the two without any hiccups, every generation becoming better and more capable.

Many years after AMD announced their fusion, bulldozer creeps out. There was no chance of a GPU sharing the same socket much less the same die. Bulldozer was and is way to power hungry. Its a serious flaw in the design. There is no quick fix. The near future for AMD and bulldozer leaves little options. So this is where we are and to the heart of my post.

Where can AMD go, what routes are there?

1) abandon BD to continue the fusion on the older architecture such as Stars. Gives them mediocre CPU performance without a high TDP. They have tons of experience here and as they lower the consumption they can put larger and larger graphic capabilities on the die.

2) use the bulldozer modules merging it with their GPUs. This case would give them less cpu performance at a higher TDP so the fused GPU would be limited. They would be extremely limited with their modules as they are very power hungry. At start you would most likely have a dual core single module fused with a GPU. if you make it 4, your TDP is not ideal. Over time (and a ton of cash) they should be able to improve the modules performance and lower TDP allowing for better graphics performance.

They are already totally invested in fusion and there is no turning that back. So these are their only options. This is it, its real, and the sad truth.

So looking at the facts, we have these two paths. With option 2 they start way below where option 1 is. So starting off with 2 puts them way behind from the get go. Its hard to say which route they will take. The most reasonable choice is to abandon the BD module all together. That would be an embarrassing admission of total and complete failure. Tons of time, money, and engineering down the drain. This would be the smart and hard thing to do. But even if they do, what does this leave them?

Its blatantly obvious that in both cases their CPU is significantly behind intel. BD needs moar cores to remotely attempt to compete in the heavily threaded task, but as far as fusion goes, those cores cannot be there. Their older architectures are years behind intel as well. No matter what they do now, intel has a massively huge advantage. This cannot be remedied quick.

AMD must abandon the high performance CPU segment and focus on their only advantage. The GPU. They will try to offer weak CPUs with powerful GPUs. Most of their cash in development is on this. Its not on developing better CPUs or GPUs even. Its fusion. Fusion of weak x86 CPUs with current AMD gpus. Its all or nothing at this point. And if you dont read anything else in these statements you must see this.

The really sad thing is that AMD didn't trademark Fusion. All of their slogans went to hell when they had to drop the name. "The Future is _ _ _ _ _ _" is pretty embarrassing. And i am beginning to wonder if they have "VISION" even nailed down.

The problem with AMD focusing on their advantage, the GPU paired with a weak CPU is exactly what Nvidia is doing with Project Denver. Do they think they will be able to get an advantage over Nvidia when they failed against Intel?

Their future is very cloudy with this CEO at the helm. He is only stating the obvious and it doesn't look like they have a real plan.

Ocre's post is spot-on IMO. AMD is not in a good position at all right now. They are doing ok with their GPUs but I'm sure that *no one* is buying their CPUs right now.

Do you really think that AMD are doing great with their GPU sales right now?
- i mean other than HD 7850? Do you think the HD 7950 is selling well; or the reference HD 7970 compared to the GTX 670? They are very slow to respond with pricing and/or with overclocked models that match it. It is like they are still in denial.

Here is who AMD is going against in the smartphone field .. a real predator
viewtopic.php?f=97&t=24955