Soldato
Yeah it would be ridiculous for it to need a new motherboard, that would be like a change in motherboard from Skylake to Kabylake?.... oh.
-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
Intel Cannonlake, Cascade Lake, Ice Lake, Tiger Lake & Sapphire Rapid thread
- Thread starter TheF34RChannel
- Start date
Yeah it would be ridiculous for it to need a new motherboard, that would be like a change in motherboard from Skylake to Kabylake?.... oh.
odd now with ryzen sure, but intel wasnt expecting ryzen to be as good as it is so may have engineered coffee lake to need a updated socket/chipset for new "features". now this is all tinfoil hat tom foolery but wouldn't put it past them after the raid shenanigans on x299.
Soldato
- Joined
- 11 Mar 2013
- Posts
- 5,501
http://wccftech.com/intel-ice-lake-processors-taped-in/
They haven't even released x299 socket yet or Coffeelake or Cannonlake
They haven't even released x299 socket yet or Coffeelake or Cannonlake
Caporegime
- Joined
- 18 Oct 2002
- Posts
- 33,188
It's almost certainly backwards compatible in reality, that doesn't mean Intel will enable the chips to work on z270/older boards. Testing on available boards makes sense, but gimping them so the chips won't boot on a z270 is extremely easy for final silicon.
Exactly, Skylake to Kabylake doesn't need a new motherboard.
- Joined
- 12 Jul 2015
- Posts
- 1,694
http://wccftech.com/intel-ice-lake-processors-taped-in/
They haven't even released x299 socket yet or Coffeelake or Cannonlake
The latter will only be a low power mobile part so for 10nm we're looking at Ice Lake. It probably also means no Cannonlake-X but likely the 14nm+ Cascade Lake-X instead, in turn leading nicely into either 14nm++ Coffeelake-X or 10nm(+?) Ice Lake-X.
Last edited:
Soldato
- Joined
- 11 Mar 2013
- Posts
- 5,501
Cannonlake is mobile? Didn't know that
It wasn't originally meant to be mobile-only, but it looks like it's going the way of Broadwell. So desktop is getting Coffeelake instead, and then MAYBE Icelake next (there's a rumour they're bringing it forward, as it's the next real arch change, instead of refinements of Skylake)
I'm not sure if we'll see Cascade Lake on HEDT, or Coffeelake-X HEDT. One or the other will surely get dropped, partly because they'll be a feeble upgrade to Skylake-X anyway, but also because Intel probably really need to bring forward Ice Lake, otherwise they'll be putting what is a revised Skylake-X on 14nm against Zen2 on 7nm. If Zen2 is only 5% higher IPC, and can clock to 4.5 GHz because 7nm is much better than the 14nm botched Samsung process (as in GloFo botched what it's meant to be like), then a Skylake-X revision would be in pretty big trouble.
they'll launch a z370 chipset, but most z170/270 should just need a bios update (like kabylake for z170)
zen 2 is still on the 14nm process, or zen+ whatever they're calling it now, technically zen2 is 7nm, but that's not until 2019, 2018 is zen+ which is essentially rx480 to rx580, so maybe 5% ipc, and maybe more chips at 4ghz instead of 3.8ghz I would say at best, since the 'refined' 14nm from gloflo has been shown to not be much better performing.
https://www.kitguru.net/components/...cpus-on-14nm-before-jumping-to-zen-2-and-7nm/
I mean, what I said still stands, just maybe add 6 months to the timeframe.
If that report is true, they're only bringing out a refreshed Zen1 on a refined process. Nothing to do with Zen2.
And unless Intel bring forward Ice Lake and Ice Lake-X on 10nm, they'll still be putting a very refined Skylake on 14nm against Zen2 on 7nm.
ice lake is 7nm and due next year and they've said they're on schedule, which means Intel will be on 10nm before amd gets on 7nm (which are actually the same density as Intel uses a better node)
Don't know where you heard this, but that's complete rubbish.
Intel don't even have the capability to manufacture at 7nm yet (expect maybe in tiny tiny quantities in the lab). All they've said is they're building a 'pilot plant' this year to iron out 7nm mass production, and they have no plans to ship 7nm products for 3 years or more.
Also currently their 10nm process is worse than their 14nm process (for performance, obviously it's smaller), so they're still working on maturing that before they can even think about putting an HEDT platform on it.
Within this, it is worth noting the GloFo 7nm process (developed by IBM) is not really 7nm. It's much more like Intel's 10nm. Same for TSMC and Samsung (who's processes are each different from each other, and GloFo).
Once GloFo, TSMC, and Samsung are manufacturing 7nm+ (7nm with EUV), that will be superior to Intel's 10nm.
sorry meant Intel's 10nm is due next year, but that's a full year ahead of zen2 on 7nm, giving Intel an even higher ipc lead than what zen+ Will even have.
if I'm not mistaken Intel's 7nm is around 2020/21 because their new plant should be up and running by 2019 iirc (which will be closer to the 5nm amd has planned for 5+ years time)
won't Intel's 10nm+ be out in 2019 the same time amd has 7nm? so the differences will stay largely the same
Except there won't be a desktop 10 nm Intel part until Zen 2 (7 nm) arrives, at least according to current roadmaps. I wouldn't bet against Intel bringing 10 nm on desktop forward. Also, expecting higher IPC from a node reduction is naive considering what has happened over the last decade. On the contrary, the biggest IPC improvements have come from iterations on the same node as the previous one. That's why Skylake was such a disappointment; people were expecting another Sandy Bridge or Haswell.
Haswell to broadwell had a bigger ipc jump than broadwell to skylake, it was just that there wernt any 5700ks etc to teat vs the 4790k directly, where as the 6700k is what most people jumped to.
not sure if you're getting confused here but ice lake is being released on desktops next year, and Intel has recently said they're on track for that, amd is releasing a refresh next year on 14nm, and zen 2 will be 2019 on 7nm.
it'll be a good whole after ice lake releases that zen2 comes along
My Skylake motherboard supports kabylake with a bios update along with quite a few others.
Exactly, hence it was a disappointment.
No, Intel's official timelines still show Icelake is due in 2019, same year as Zen 2. There are a couple of outlets suggesting it might come out in 2018 (and one merely saying they "need to") but nothing concrete. However, Coffee Lake was brought forward so it's possible Icelake will be brought forward too. Remember that Kaby Lake is only 5 months old; Intel's public roadmap doesn't mention anything after Kaby Lake on the mainstream socket for 2017! If Icelake does come out in 2018 you have AMD to thank for that.
Last edited:
Np, easy to make number errors will all these nm sizes and dates.
Their 10nm is due late this year/early next year, but only for laptop parts (and I'm not even sure the full-power 4 core parts, might only be 17W ultrabook and below).
Intel need to get HEDT onto 10nm to compete properly with Zen2. So they have a max of 2 years to mature the process enough, since GloFo say mass production of 7nm starts 2H 2018, and I believe they're already in risk production now (as well as TSMC for their 7nm).
Also mid-late 2019 will be 7nm+ for GloFo (7nm with EUV, so a decent improvement). So Intel BETTER have 10nm+ done by then, or again they're in trouble.
2020/2021 does appear to be right for Intel's 7nm, as they seem to be set on EUV from the get-go and also they're slightly behind as they dilly-dallied for a bit deciding whether to do EUV or lots of multi-patterning.
Also it's worth bearing in mind, although Intel's 10nm (once mature) will be almost identical to GloFo's 7nm - Intel going from their 14nm to 10nm will be a smaller jump than AMD going from GloFo's 14nm to 7nm.
So whatever jump you expect to see Intel make going 14nm-10nm, AMD will be making a bigger jump. And add to that, it appears GloFo's 7nm+ will be superior to Intel's 10nm+, because Intel still won't be using EUV at that point.
What I'm trying to put across in all this, is potentially as early as late 2018/early 2019 Intel need to compete with Zen2 with 24 cores on 7nm, which is a process which is reported to be much MUCH better than the current 14nm GloFo is using, for both power and clockspeed potential. And I just can't see how they're going to do that, outside of bringing forward Ice Lake HEDT to 2019 on their 10nm+ if it's ready.
Last edited:
yeah 10nm cannonlake is due this year for notebooks etc, desktop is going straight to ice lake much like it did with broadwell, which should be early next year judging from previous time lines.
with Intel going up to 6/12 on consumer chips though it will be much closer toryzen than we think, already their skylake x 6 cores will do 5ghz, so it stands to reason their consumer chips will too
a 5ghz 10nm should be equally as good in multitasking as the 1800x at 4ghz, including any ipc gains from the smaller node etc.