When will cannon lake desktop hit town? Coffeelake Q4 2017? So CL Q3/4 2018?
Don't see why Intel couldn't just lower HEDT prices and add re-branded a Xeon for the top. I don't think ignoring Ryzen is a sound business decision. Any real response has to come with Skylake X. HT is better with Skylake so AMD's SMT advantage is going to vanish. I doubt there will be a eight core at $500 but it sure will be under $1000.
Right, so icelake Q3/4 2018? Is that a new uarch AND 10nm?There's a chance they will do a Cannonlake-X on X299 but that would be it. No mainstream model. Icelake would be next and presumably expect that a year after Coffee Lake arrives.
Right, so icelake Q3/4 2018? Is that a new uarch AND 10nm?
Things could get dicy for AMD.
I disagree. If the chips stop selling prices will go down. You can get a whole eight core rig for less than a 6900K soon.Unlikely, as that would hurt their bread and butter server chips. Segmentation is what has made Intel a boatload of money, in a market with no competition. But, it is also an Achilles Heel now. They don't have a lot of options, aside with fiddling with SKU's for the next several years.
Not good for AMD!! Better hope Zen++ comes out H1 2019 on 7nm.See: https://en.wikipedia.org/wiki/List_of_Intel_CPU_microarchitectures#Roadmap
Ice Lake is on 10 nm and is the "Architecture" step.
One-year cycles:
Kaby Lake: 2016 2H
Coffee Lake: 2017 2H
Ice Lake: 2018 2H
Right, so icelake Q3/4 2018? Is that a new uarch AND 10nm?
Things could get dicy for AMD.
I disagree. If the chips stop selling prices will go down. You can get a whole eight core rig for less than a 6900K soon.
They where talking about architecture its an assumption that it also means process. Just like CON cores each new chip/soc was a uarch refresh they moved form 32nm SOI to 28nm bulk in there as well.
http://www.pcworld.com/article/3155...hitecture-is-expected-to-last-four-years.html
The next Zen will i think 100% be 14nm LLP ( 18)
The Zen after that i think 40% 14 LLP , 40% 14 HP, 20% 7nm ( 19)
The 4th Zen i think will be 100% 7nm ( 2020)
I wish you are right, but I personally don't think that even Coffee Lake will bring any IPC improvements. Whatever improvement Intel guys developed until now after Skylake, it has never been designed for 14nm, but only for 10nm. Cannonlake and Ice Lake will definitely bring bigger or smaller improvements, but that's all about 10nm. I don't think Intel had to be forced to port 10nm improvements to 14nm uArch. It's much too risky in this timeframe. Maybe it was possible, but they've had to decide about that before Skylake release. I don't think this was the case. It's too late.As I explained in my post both Cannonlake and Icelake µarch must be ready, heck they displayed a working 10nm Cannonlake so that's done and exist 100%.
It may be just 5%, but it adds to Skylake advantage, then 14nm++ and some clock gains... they won't come with Skylake-X because that already done (with its own possible gains against desktop arch.) but there's no reason to not bring all IPC gains to Coffe Lake if it comes end 2017/2018. Tick tock may be dead and 10nm late but cpu designers worked anyway these years.
No need, AMD will just release higher clocked SKU's to stay competitive.Beside by the time 2ng gen Ryzen hits AM4+ will be ready and have much higher IPC to undercut Skylake-X.Not good for AMD!! Better hope Zen++ comes out H1 2019 on 7nm.
I'm waiting until mid April at the latest. AMD is probably getting my $$$ this time around.
I agree. Coffee Lake is probably a late addition to roadmap. Usually Intel had slack with finishing product unless process trouble, but in this case it might not be the case.I don't think Coffee Lake is done TBH. Cannonlake yes but of course that's DC mobile only and the fabs may not be ready yet.
People shouldn't be surprised.
Look at all the AMD fanboys who use Intel CPUs today because they have no other choice because of performance.
Intel fanboys can see a great CPU at a great price too, and that is what they haven't had for years from Intel.
If I am honest I think that intel fanboys who buy Ryzen are even more admirable, because they probably don't need it.
I think that sums it up so we are in agreement... Zen+ and Zen++ are both on ~14nm in 2018 and 2019. (If other foundries want to call it 14nm+ node to encompass some improvements over previous 14nm, that would probably bring it to parity with Intel 14nm)
My point is that for these two years Intel will be on 10nm.
If it becomes a #cores race, the process advantage is quite an advantage IMO.
Global Foundries starts mass production of their 7nm process node in 2018 and it will be more dense than Intel's 10nm. Enven if for some reason some people want to deny that fact.
The big intel SOC's on 10nm aren't going to be soon, skylake E/EP isn't even out yet. So it will tkae intel time to get a cores race going if they wanted to try and do that. But AMD still have a big binning/ yeilding advantage from the multi chip path they are going.
Not an actual hard number of course but we can at least understand the scale of their roll-out. http://www.digitimes.com/news/a20170222PD211.htmlLet me know when Intel or AMD releases sales numbers for any of their CPUs.
I agree not "soon," we are talking 2018-2019.
I think, the improved transistor performance is at least, if not more important than scaling, as there are limits in dissipated heat per sqmm.Global Foundries starts mass production of their 7nm process node in 2018 and it will be more dense than Intel's 10nm. Enven if for some reason some people want to deny that fact.