Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
TCaE
Intel would be wise to play this game
For what purpose?
morrowinder
TSMCs 10nm is likely inferior to Intel's 14nm since they are just using bullshit marketing names at this point
Unless we can see data on Ion, Ioff, sram cell size, gate pitch, metal pitch etc, then you're right. It's all marketing bullshit at this point and it looks like Intel is the only one that's not playing the game.
TCaE
The "naming" that Intel uses for example is meant to loosely approximate the scaling in the linear dimension from one node to the next.
That's what I said.
However, there have been times in the past when Intel (and others) have named their nodes in a manner to suggest the 2x area scalling when in fact that was NOT the case.
I think you missed my point. With planar transistors, the process was named by the channel length, not the scaling.
TCaE
Thank you for the explanation, this starts to make some sense. So to restate, you are saying that seeing that even though we no longer have planar transistors, we're accustom to describing process generations in terms of transistor channel dimensions. So if we can scale down the transistor density of a process generation by the equivalent ratio(0.7x), we should be able to call it 10nm because that's what it would be if we were still on planar transistors, even though there is no actual dimension that measures 10nm with FinFets, or is there? FinFets are tapered so where should such a measurement be made?
Additionally, what information do we have on the scaling factor from the foundries when going from 14nm to 10nm? Is it 0.7x?
He proclaimed “Moore’s Law is alive and well”
Really? Does this mean that there is a geometry on this process that is actually 10nm as the name implies? Is the number of transistor's doubling per given Xmm2 going from 14nm to 10nm? What justifies the name 10nm?
remember when AMD was dancing in the aisles?
That one will live forever in infamy
The solution is to simply rename their 14nm process 10nm.
morrowinder
We can be sure that what Marketing showed us was the very best they had to show and what we didn't see, which was everything else, was what didn't look nearly as good. That's what Marketing gets paid to do and that's what they did. The question is can AMD make any money? I'm inclined to think that maybe they can. Needless to say, we need to see unbiased 3rd party evaluations. The comments section of the leaked benchmarks contain a lot of skepticism as to the authenticity of those benchmarks.
GE paid out $2.1 billion in dividends (again, counting common shares only) just this morning
Yesterday morning? You could have fooled me...
Andy
How do you feel about this new naming convention, 14nm, 10nm,7nm etc when no feature is actually 14nm, 10nm,7nm etc? Seems to be to be just the battle of the marketing slides.
Andy
Doesn't make a lot of sense to me. Intel empowering it's rival and infusing it with cash, for patents? Why not get more patents from NVidia instead? A dropin design would make more sense.
Francis
Will Zen be a high end CPU unit or it's more of a medium segment?
We don't know. AMD's long standing track record is to over promise and under deliver. Will that pattern continue here? Only time will tell.
Andy
only one somewhat cryptic claim.....that's why I said many questions need to be answered...
The only thing I'm aware of is a claim of 40%+ IPC improvement over their previous generation. However it is unclear how this translates into real world performance and AMD has made no attempt to clarify. They had an opportunity to do so when they did a side by side comparison with a "similarly configured" Haswell (IIRC) showing a slightly faster Zen performance but they wouldn't give any details whatsoever about what the configuration was so no one could verify. The software was open source so we don't even know how it was compiled. The funny thing was AMD could have done the obvious thing and run the demo against their own previous generation CPU to demonstrate their claim of a 40% IPC improvement and remove all doubt but instead they chose to run it against an unvarifiable and completely unknown setup leaving more questions than before the demo.
Needless to say the fans went wild...
TCaE
OK but I don't know what "competitive" means, does anybody? They've made no specific claims. Nothing that can be verified or anything they can be held to.
If Summit Ridge can hit the performance AMD claims
AMD hasn't made any performance claims whatsoever.
Any diet lacking nacho flavored corn chips borders on malformation.
this news has no positive effect on share price?
What news?
Bozo
I am still getting over Otis
That's because you are amongst the crybaby set.
borusa
There is still a bit of shock digesting the election results. All will return to normal is due time.
TCaE
Intel's silicon photonics began volume shipments earlier this year
Didn't Apple actually ship a product based on it before that?
Silicon Photonics
3-5 years away. Seems like I've been hearing that for nearly 20 years...
TCaE
I don't see this a valid comparison to AMD's schedule slips. Intel isn't facing bankruptcy and doesn't need to fool anyone into holding false hopes for it's 10nm release date. The term "10nm" has no meaning any more, it's just a marketing term with no specific definition. Is there anything about anyone's "10nm" process that's actually 10nm? Zen refers to a specific product that AMD is very dependent upon to maintain it's viability.
TCaE
FWIW, Intel is not so innocent when it comes to product schedules slipping
Sure, Intel does slip from time to time but Intel does not publish schedules intended only for the gullible. A schedule that claims a new uArchitecture from tapeout to product availability in 10 months is intended only for the gullible and it worked.
Chipguy
AMD's initial schedule had product availability 10 months after tapeout. That was an entirely new uArchitecture on a new process at an outside foundry. The Faithful bought it hook line and sinker. It has now slipped from "late 2016" to "1H 2017". The more things change the more they stay the same.
Chipguy
While anything is possible, I see absolutely no difference at this point between this Zen buildup and the buildups for Bulldozer and Barcelona before that. The psychology is the same amongst the rabid fans and of course it's a completely new generation because the previous generations were crushed and gave up, just like the ones before them. Maybe this time it's different but at this point it's still the same old same old and Lucy is still promising to hold the football...
Dmcq
I don't understand what your point is
I guess it doesn't really matter any more.
One has to be careful of what one means by 'world class CPU'
Indeed. I suggest you be more careful next time.
Dmcq
You think an AMD CPU is doing all the work there?
AMD is the only semi with a compelling story. No one else has world class CPU + GPU
AMD does not have a world class CPU.
TCaE
No. Intel is clearly behind in terms of modem features, quality, and peak upload/download speeds. Apple appears to be giving Intel a "pass" in order to allow Intel to win some business.
That was not my statement. I said it was likely Apple gave Intel the specifications and Intel met them.
I wonder which it is, Chip Thinning or Wafer Thinning?
Yes.
Ducati
Isn't a large portion of that debt already spoken for a long term rates? Higher rates would apply to new rates and what if the Fed buys the new debt? I'm no expert here. Far from it. Just asking questions.
jbbrown
Who said I didn't think so? I simply asked for an explanation.
Andy
Don't you think it's more likely that this chip was designed to Apple's specifications, rather than Intel's?
When the “nasty woman” doubles or triples spending, Metals (especially Silver) will be KING
Can you explain to us why rising interest rates will cause PMs to increase in value?
Andy
...what a pile of psycho babble from you
Perhaps but it perfectly describes the past, perfectly explains the present and perfectly predicts the future.
What we have here is a failure to communicate...
morrowinder
Just let Lucy hold that football for you Charlie Brown, really it will work out this time
The real key is that this is a matter of human psychology. As each generation of cult members is finally disillusioned, there is always a new generation of suckers willing to step forward and take their place because the need to believe in an impossible dream is a human need and AMD fits the bill perfectly. Losers hate winners so Intel became the embodiment of evil. AMD could make up any lie about Intel and the believers would believe it because they needed to. This is the human condition. It always has been and always will be.
Bozo
Did you notice that AMD will pay the other foundry for wafers while also paying GF at the same time for using those same wafers from the other foundry? What is it about GF wafers that AMD is willing to pay GF twice to NOT use GF wafers?
EU fine ain't over yet:
That will break FPG's heart.