My point still stands: They will have a deficient CPU design, and the only good thing they have is their GPU part that might help accelerate some workloads, but not all. Once you move to more conventional workloads, AMD deficiencies will be highlighted again.
And assuming that HSA will ever take off, given that HSA is an open standard, what prevents someone from outspending AMD in designing and get a better CPU and better GPU, and better HSA performance? That's right, nothing. So even if HSA succeeds, there won't be much market for them.
True about Jaguar in consoles- I was mainly talking about Temash, sorry. Its complete lack of CPU turbo (a step backwards from Bobcat!) crippled it in low-power performance. Can you imagine if Bay Trail had no turbo and was limited to 1.5GHz?
Hopefully some console SoCs will be made at GloFo, yes, but AMD still need someone to make a process suitable for high performance CPUs. I don't think either GloFo or TSMC has anything on their roadmap which would make a high performance desktop CPU feasible- there's a reason why Carrizo is going down to a 65W TDP... Being free from GloFo doesn't help them compete with Intel if they don't have a better alternative to jump to.
And a Playstation 4.1 sounds like a terrible idea! D: As Sega demonstrated in the 90s, fragmenting your console user base is a bad move...
Management prior made a statement of a minimum balance of 700m. Whatever bs they said about cpu marketshare they reached that important goal and then some.
Please no. I don't want to go back to the bad old days of Pentium II pricing.
My point still stands: They will have a deficient CPU design, and the only good thing they have is their GPU part that might help accelerate some workloads, but not all. Once you move to more conventional workloads, AMD deficiencies will be highlighted again.
And assuming that HSA will ever take off, given that HSA is an open standard, what prevents someone from outspending AMD in designing and get a better CPU and better GPU, and better HSA performance? That's right, nothing. So even if HSA succeeds, there won't be much market for them.
They made a statement of 1.1 billion for a comfortable operation, while 700 million being the very minimum to operate. They only reached that comfortable level only after starting to use the short term debt line.
They also stated that they are on track to reach the WSA commitments, and they weren't even ramping up Kaveri. And we didn't get a WSA commitment by now, as they did they would in Q313.
So I think they don't deserve the A+ grade here, maybe a C.
What would you do instead?
14nm process node, 10, 7 and 5nm following at a constant pace after 14nm with high yields, and other manufacturing advances? Haswell, Broadwell, Skylake and Cannonlake, Crystal Well, gen8 and gen9, Atom with Silvermont, Bay Trail, Merrifield, Moorefield, Goldmont, Broxton, SoFIA, Quark and other Internet of Things stuff, SSDs, chipsets, 450mm wafers, III-V semiconductors?
down $1k today on my long position
WARSAW is out folks.
http://www.maximumpc.com/amd_adds_more_12-core_and_16-core_processors_opteron_6300_series
That reported TDP seems too good to be true.
Lets hope its not just on paper.
It's just an underclocked piledriver.
In the past or now?
In the past I would have killed Bulldozer circa 2009 and going for a shrunk K10, on top of that would have extreme focus on Bobcat, not trying to bring it upwards but bringing it down. AMD did a very nice job in placing bobcat in the huge gap between Intel to product lines, but instead of bringing it down towards the mobile space, they just improved the chip enough to let it stay on the same market bracket. As the gap closed, the chip become less of a success than it was before.
Today? Probably would go get Nvidia's bacon. There's plenty of profits to be made in the workstation market, and that market won't go away any sooner. But in order to do that they would need to improve their software suport to enterprise levels, which it seems they can't.
Cheaper, smaller scope and less time to mature than HSA.
Intel basically already had.
For compute performance, Intel's design are neck and neck with AMD, and the edram they have on Iris Pro is likely to be a huge advantage. Not to mention that Intel's gpus have been "fused" to their cpus for far longer than AMD's.
Not to mention that Intel's gpus have been "fused" to their cpus for far longer than AMD's.
Intel are missing HSA, hUMA and stacked DRAM.
The first already available, the second soon available. Both soon to be utilized going forward.HSA is stillborn and got 0 apps.
hUMA...still nothing seen.
KL is not mainstream. When will we see it in Intel mainstream products?Stacked DRAM? In what product? Intel will deliver stacked memory before AMD with KL. When will AMD deliver? It will not be the next GPU nor Carizzo.
The first already available, the second soon available. Both soon to be utilized going forward.
The trend is clear; CPUs are not getting much faster from one generation to the next, GPUs are. You need to adapt to that and utilize GPUs for other tasks than pure graphics in the best way possible. Intel is late to the party.
KL is not mainstream. When will we see it in Intel mainstream products?
Stacked DRAM? In what product? Intel will deliver stacked memory before AMD with KL. When will AMD deliver? It will not be the next GPU nor Carizzo.
Nothing new for years already, AMD's CPU division is a disgrace. If not for APUs, they wouldn't even be around anymore. Intel obsoleted AMD's FX lines a long time ago.
The FX-6300 is a pretty good deal. It beats out Intel's offerings at that price point in most benchmarks (Intel does still win in single-threaded benchmarks and in power consumption). And you can overclock it, which you can't do with a modern i3.
FX6300 is a die salvaged chip operating in an ancient platform. It's also a very big chip from the kind of market bracket it sells. Not a good product for AMD, that's for sure.