New Intel Roadmap leaks: 10 core Comet lake in 2020, Rocket Lake on 2021, both at 14nm
Click here to post a comment for New Intel Roadmap leaks: 10 core Comet lake in 2020, Rocket Lake on 2021, both at 14nm on our message forum
Fox2232
On 1st look it looks worse than it is. But intel is planning to deliver 10nm chips.
Here is their thinking:
- On desktop, people will not care about TDP, therefore we can write any power consumption in spec sheet we want. And we can get clock above 5GHz so we will stay competitive.
- On mobile AMD will invade our space with good power efficiency, therefore we'll bring 10nm to lower TDP space.
= = = =
It is pretty safe strategy. Very few of us care about actual power draw of our desktop CPUs. And even those who know way intel misrepresents TDP/SDP values will not boycott their chips over it. So, why would intel need 10nm?
And in mobile, there are going to be 10nm CPUs. They will be just in time to keep dominance. As it may took even year for people/manufacturers to start perceiving AMD's better performance per watt as viable/better solution.
(Side note: Intel always twists my nerves with those low SDP/TDP values, in many cases real world power draw is even 3 times as high.)
Taint3dBulge
My question is, will the comet lake cpu's work on the current z370/90 mobo's? Or will we have to upgrade to a z470? It would be really cool if they left it alone and we could upgrade since its practically the same chip, from what i understand anyways. Im running 8700k and while its a perfect cpu, iv been playin with the idea of goin 9900k, BUT if i could go with a 10 core i can wait for that........
tunejunky
Intel is just coasting (like the last 3 years) without any clear strategy. their contract works(14nm) gummed up the 8th gen cpu availability, now they lost that contract work so they're doubling down on fab capacity to earn a ROI (return on investment) from the capital expenditure.
the most stupidly irritating thing is that Intel's fattest margins are on their best products which are losing market share dramatically.
and as far as the mainstream doesn't care about TDP or (ultimate) FPS, the market does care about value.
try to explain to a non-gamer that Intel will give them 10 more FPS, but at a cost of (up to) several hundred dollars over AMD (with more cores) and the eyes will be rolling.
this isn't just an opinion, it's what's happening right now, especially in Europe.
user1
They probably could, though the biggest issue is the power draw, z370 boards with weak vrms could have issues, since they were intended for 6c cpus.
, some of those boards dont handle the 9900k all that well, adding another 2 cores isn't gonna be ideal. Overclocking would probably be a bad idea
Loophole35
RIP Intel. AMD will likely have 40% market by then and all the momentum. Staying at 14nm is not going to help regain what is going to be lost.
Calmmo
Well this just further solidifies my Zen2 jump about to take place in June
SamuelL421
Intel has made some quality hires and they also nabbed engineering guru / gun-for-hire Jim Keller last year. The mess will continue for another year or two but Intel reemerge unscathed.
I say all this as a hardcore AMD fan.
nick0323
Not fussed on power draw on desktop CPU's and I imagine most people aren't any ways. Let's see if they still bring 10nm for mobile (laptops) where a lower power draw matters.
Fox2232
Can AMD make that many chips? I do not really think that production capacity allocated for them is that high. Sure, AMD is going to improve their market share and make good money in process. But 40% is a lot of CPUs.
If report like that comes out, AMD's stock will double over night.
Loophole35
I would think it possible by 2021. Though I haven't looked to far into the numbers.
I hope Navi is a hit too.
waltc3
Vega obliterates any GPU Intel's ever made...;) Might want to rethink that...;)
waltc3
...And the moon is made of green cheese, too!
waltc3
Sounds like maybe you have a problem with government regs as opposed to AMD. I've never had a problem buying AMD.
Loophole35
Not a "big" advantage. They have an advantage but not a huge one. Zen does everything you could ask for all while using less power and running cooler. the first run of Zen was a little lack luster but the refresh has proven to be great. I can't wait to see true Zen2.
user1
i think they will manage, the zen2 chiplet is 7x10mm, very small , on the 300mm wafers tsmc uses, thats ~862 8core chiplets per wafer, and if we use the 70% yield amd claims thats ~603 dies that can be sold. that is probably more than double the chips per wafer than zen 1 (~213mm^2, like ~9x23mm) , now there is the io die, but i can't imagine glofo will be the bottle neck.
given that amd is supplying chiplets for sony as well, they will have probably have quite alot of allocation, that could easily be few million chiplets on launch day alone(ps5).
Now obviously demand could still end up being higher than expected, but if the current trends continue normally, it should be ok.
Aura89
Guess intel should just go back to 65nm or higher then, see where your performance will stand then.
Being low nm doesn't automatically make a product better in performance, but it does allow for more. If intel is stuck on 14nm for the next year and a half, it means we have no expectations from them then the same performance they have now. Sure, they can tac on more cores, make their CPU die larger, hotter, more expensive to produce, but thats it. The efficiency, max frequency, heat output (relative to performance) won't change, unless they get on a smaller node, or do extremely major changes and overhauls to their architecture.
You call it a "nerd trend" but that clearly means you just have no clue what you are talking about, as that would imply we could be on 65nm or larger and be at the performance and cost we are now.....thats not the case.
RooiKreef
Looks like AMD have a very good chance to overtake Intel. With that said though Intel is optimizing like crazy at this point so when they do drop node size then they will once again be on the top. But then again the people at Intel know how to make money so they might be looking at a whole nother angle than us. Only way is to see what they’re up to.
Reddoguk
Be fast or end up being last.
Zen2 could really shake up the market @ this rate.
Venix
seems the 10nm for intel do not want to work...although people stop telling about 7nm on amd the 7nm in tsmc are close to 10nm from intel the 14nm are actually 20 nm ....but the 20 nm non finfet node had almost no advantage for a lot more trouble so companies skipped it .... then glofo and tsmc made 20nm finfet ...but because it is so much better than their non finfet counterpart for marketing they named it 14 nm !!!!!their 7nm is a bit worse in density than intel's 10nm and better then the intel 14 nm , so tsmc and glofo "14 nm " are still inferior to intels 14nm ... with some respectable margin ...so far intel fighting on a superior lithography ! zen 2 .... is the time for the first time in history intel will not have the lithography advantage that's exactly why it is one of the most exciting releases if you ask me !
fredgml7
I believe Intel will "always" have competitive products, even back in P4 era, when Athlons were better P4 wasn't that bad at all, so competition will remain and Intel is not going anywhere, but if it loses market share to something like 50% it would be a great thing, better to all of us.
