It's not a some thing. It's the entire lineup. We're talking about maybe doing a paper launch of Meteor Lake at the end of 2023. Maybe.
It's not a some thing. It's the entire lineup. We're talking about maybe doing a paper launch of Meteor Lake at the end of 2023. Maybe.
Atari2600
Golden Member
- Nov 22, 2016
- 1,409
- 1,655
- 136
I think instead of "Lakes" and "Rapids", Intel needs to switch to a "Creek" codename basis for a wee while.
I'm sure its not hard to figure out what the first one would be.
I'm sure its not hard to figure out what the first one would be.
IntelUser2000
Elite Member
- Oct 14, 2003
- 8,686
- 3,787
- 136
In addition to what @repoman27 has said, if you know what architecture the CPUs uses you can extrapolate pretty easily. Silvermont Atom to Airmont Atom is a small change in architecture and will barely make a dent. Yet on 14nm Airmont manages to be 2.7x dense. Get it now?
@repoman27
Great thoughts. Yes they are in trouble but exactly. You also answered the question on what happens in server? End Early 2021, Sapphire Rapids and it doesn't even look like it'll fully ramp and Icelake will live with it. End 2022 is Emerald Rapids and End 2023, early 2024 is Granite Rapids.
Remember the cycles are generally 12 month+ so 2024 will be an indicator how Intel really does with 7nm.
Last edited:
Yet you might compare a soft IP version of Silvermont with a hard IP version of Airmont. In any case you will not see close to 2.7 density increase for Atom from 22nm to 14nm everything else being equal.
ECC is kind of the whole point of Xeon E. There'd be no point of the product if Core had ECC.
Hulk
Diamond Member
- Oct 9, 1999
- 5,141
- 3,734
- 136
Seems like in the past node shrinks used to provide a trifecta of advantages, namely increase transistor budget for a given area, lower power consumption, and higher frequency. Now process shrinks seem to provide for more transistors and lower power.
As current technology hits the wall of physics I can see processes kind of stalling with development being tweaks here and there rather than entirely new nodes. Kind of like how the internal combustion engine has been refined for the last 100+ years but over the last 10 or 20 improvements have been relatively minor.
As current technology hits the wall of physics I can see processes kind of stalling with development being tweaks here and there rather than entirely new nodes. Kind of like how the internal combustion engine has been refined for the last 100+ years but over the last 10 or 20 improvements have been relatively minor.
Oh, how generously time heals all wounds and promises... let me know if you really want to go into this debate about the most 'ambitious' process design ever.
That's why I was asking if there was any evidence it existed for Alder Lake. Unfortunately there is. If it didn't exist it may imply Intel was moving ECC to consumer platforms. Obviously I was hoping for that.
If that is really the case, they better hope that 10esf is something amazing to behold...
If it can get to where N6 is, coupled with Intel's ability to push serious volume, then they will still be in the market. I doubt, beyond EPYC, thread ripper, and some high end ryzens, AMD will be shipping massive N5 volume in six quarters. We're three quarters minimum from any sort of N5 product announcement. Four to six from any sort of availability.
If it can get to where N6 is, coupled with Intel's ability to push serious volume, then they will still be in the market. I doubt, beyond EPYC, thread ripper, and some high end ryzens, AMD will be shipping massive N5 volume in six quarters. We're three quarters minimum from any sort of N5 product announcement. Four to six from any sort of availability.
Wait, where are people getting these late 23-24 numbers for MTL from? If it taped out now, in mid-21, that would be 2 1/2 years from tape out till shipping, which would be extraordinarily long. The industry standard is closer to half that.
IIRC, Intel uses tape in for internal manufacturing and tape out for external manufacturing in the same way. I'm less than convinced by attempts to draw meaningful conclusions from their use of that term. More to the point, perhaps:
Intel: Sorry, But Our 7nm Chips Will Be Delayed to 2022, 2023
In an earnings call, Intel CEO Bob Swan said the company identified a 'defect mode' in the 7nm process that resulted in 'yield degradation' when manufacturing the chips.
If not Meteor Lake, then what would that be? Makes far more sense than a stopgap product pushing things out by a whole year.
No first product on a node tapes out on a 1.0 PDK. But that distinction doesn't matter in this context.
coercitiv
Diamond Member
- Jan 24, 2014
- 7,359
- 17,443
- 136
Think of it the other way around: if MTL is late 2022 and ADL is late 2021, why are we still hearing about Raptor Lake?
Let's say MTL is more like H1 '23, since when's the last time the optimistic end of the range been a good reference? That gives roughly 1.5 years (6Q) between ADL and MTL. Smack a low-effort ADL refresh in the middle for holiday '22 sales, and I think it makes perfect sense.
Plus, Intel mentioned mobile specifically for the first 7nm product. That's in contrast to ADL, which seems like desktop first. If nothing else, it's easy to imagine a 2+ year gap between ADL-S and MTL-S.
And again, if that 7nm product isn't MTL, then what it is? Nothing else makes sense.
coercitiv
Diamond Member
- Jan 24, 2014
- 7,359
- 17,443
- 136
So you're going through all this to prove that MTL cannot be a H2 '23 product, but instead a H1 '23 one on mobile with likely H2 '23 launch for the entire stack?!
One would think "launching spread out over the course of 2023" is quite a different statement than "paper launch in Q4'23/Q1'24", no?
And I'm not sure why it's perceived as a stretch to quote Intel's own statement on the matter...
IntelUser2000
Elite Member
- Oct 14, 2003
- 8,686
- 3,787
- 136
We shouldn't be hearing about Emerald Rapids and Raptor Lake but we are.
If 7nm didn't have the issue I bet Raptor Lake wouldn't exist. Either that or you are saying they'll have to cut the lifespan of Raptor Lake significantly. Or Alder Lake. I can see Ponte Vecchio coming earlier but the volume on that is going to be tiny. I think that also works with their claim "1 year delay but with tricks we got it down".
I really can't see how much earlier it can be considering we just had Tigerlake-H and Tigerlake-U Refresh launched. So Alderlake is going to be close to Holiday 2021. Maybe they'll abandon Rocketlake early but that's still Holiday 2021.
If 7nm didn't have the issue I bet Raptor Lake wouldn't exist. Either that or you are saying they'll have to cut the lifespan of Raptor Lake significantly. Or Alder Lake. I can see Ponte Vecchio coming earlier but the volume on that is going to be tiny. I think that also works with their claim "1 year delay but with tricks we got it down".
I really can't see how much earlier it can be considering we just had Tigerlake-H and Tigerlake-U Refresh launched. So Alderlake is going to be close to Holiday 2021. Maybe they'll abandon Rocketlake early but that's still Holiday 2021.
DrMrLordX
Lifer
- Apr 27, 2000
- 22,915
- 12,988
- 136
I know, that's what I was trying to point out. 7nm will be mostly a non-factor. Intel will be highly-dependent on an old node. If you think Mizuho is blowing smoke then so be it. 20 or 36 kwpm, it matters not.
There's a difference between using it for non-critical parts and for parts where perf/watt and perf/area will be critical. Even node slackers like AMD can get away with using N7 or N6 for I/O dice since their performance-critical silicon will be on N5 by 2022. Apple will be an early adopter of N3. No idea where Qualcomm will be (should be interesting to see if they're staying with Samsung). Intel 10nm is not competitive with any of those nodes.
Yup, that's the crux of it.
Wouldn't be the first time. Kaby Lake vs Coffee Lake. Heck, looks like they may do just that with Rocket Lake this year, though we'll have to see when products actually hit the shelves.
coercitiv
Diamond Member
- Jan 24, 2014
- 7,359
- 17,443
- 136
Because they lie when it comes to nodes and manufacturing. They lie so much that they rename nodes and make products disappear. I guess what I'm trying to tell you is your estimates are optimistic while @jpiniero 's take with a '23 paper launch is pessimistic, but both fall within the "margin of error" when it comes to Intel statements. That's how bad it is in terms of trust.
I think we can both agree that Intel needs and wants MTL on the shelves in 2023, so they will try their best to make it within this time window.
Intel is launching 7nm products in 2022, just probably not a high end x86 part. We know this based on the announcement the other day.
I don’t understand the claims that Intel is somehow behind. Tiger Lake is competitive with Ryzen Mobile, and Alder Lake is going to give us laptops with > 8 cores.
You seem to gloss over the fact that the thinnest, lightest, and most efficient laptops you can buy have Tiger Lake U chips in them, not Ryzen.
I don’t understand the claims that Intel is somehow behind. Tiger Lake is competitive with Ryzen Mobile, and Alder Lake is going to give us laptops with > 8 cores.
You seem to gloss over the fact that the thinnest, lightest, and most efficient laptops you can buy have Tiger Lake U chips in them, not Ryzen.
those are also going to be the hottest, and have the worst iGPUs
Anything in 22 or most of 23 is going to be a test device at best. Like Cannonlake was supposed to be. I got the impression with Cannonlake that they thought (hoped) that yields would improve enough in time that they would be able to sell it with the IGP enabled even if it had to be mostly busted.
I always take Intel's announcements with a very large grain of salt. What they announce and what actually happens is very rarely the same thing. When there are actually products on shelves, and in reviewers hands I'll pay attention and go through the reviews like the computer nerd I am. Until then though, I mostly ignore announcements.
W-1390P and W-11955 are both Q2 2021. A 1 year typical Xeon cadence puts Alderlake at Q2 2022.
Q1'17
Product Specifications
quick reference guide including specifications, features, pricing, compatibility, design documentation, ordering codes, spec codes and more.
ark.intel.com
Intel® Xeon® Processor E3-1280 v6 (8M Cache, 3.90 GHz) - Product Specifications | Intel
Intel® Xeon® Processor E3-1280 v6 (8M Cache, 3.90 GHz) quick reference with specifications, features, and technologies.
ark.intel.com
These are all 1 year apart
Product Specifications
quick reference guide including specifications, features, pricing, compatibility, design documentation, ordering codes, spec codes and more.
ark.intel.com
Intel® Xeon® W-1290P Processor (20M Cache, 3.70 GHz) - Product Specifications | Intel
Intel® Xeon® W-1290P Processor (20M Cache, 3.70 GHz) quick reference with specifications, features, and technologies.
ark.intel.com
Intel® Core™ i9-10900K Processor (20M Cache, up to 5.30 GHz) - Product Specifications | Intel
Intel® Core™ i9-10900K Processor (20M Cache, up to 5.30 GHz) quick reference with specifications, features, and technologies.
ark.intel.com
Intel® Xeon® W-1390P Processor (16M Cache, up to 5.30 GHz) - Product Specifications | Intel
Intel® Xeon® W-1390P Processor (16M Cache, up to 5.30 GHz) quick reference with specifications, features, and technologies.
ark.intel.com
Intel® Xeon® W-11955M Processor (24M Cache, 2.60 GHz) - Product Specifications | Intel
Intel® Xeon® W-11955M Processor (24M Cache, 2.60 GHz) quick reference with specifications, features, and technologies.
ark.intel.com
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes + WCL Discussion Threads
- Started by Tigerick
- Replies: 22K
-
News NVIDIA and Intel to Develop AI Infrastructure and Personal Computing Products- Started by poke01
- Replies: 384
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
Facebook
Twitter