• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

Intel 10nm Product Lineup for 2020 Revealed: Alder Lake and Ice Lake Xeons

AMD FTW now is in the lead...sorry Intel :D

AMD covered all gaming,from consoles to pc,for everybody,from entry level to high lvl.

Also with the path AMD has created u can keep up with new cpu whitout change mobo,etc.

I will keep the i9 9900k but no more invest on intel platform.
 
For what exactly do you need AVX in a mobile chip?
Lakefield uses the same idea as you said above, "both the Big and Small core (Sunny and Tremont) have same Instruction set, but with different throughput"
If AVX not that important then why Intel was bragging about AVX-512 on Ice Lake(Sunny Cove) and Ice Lake currently availble on mobile.
And DRM software like Denuvo use FMA3 (which Lakefield lacks), with out this new instruction your cpu cant run new games. This days old but still capable Phenom II and Core 2 Quad cant run many game because of lack of new instruction like AVX and FMA3.
 
No Rocket Lake S this year then by the looks?

Rocket Lake is 14nm.

Everything is written in the article. It states "SOME" of the products are 2020 on 10nm+, and others on 10nm++ in 2021.
Alder Lake is 2021 on 10nm++.
 
Wtf is the problem with +++? All major manufacturers improve nodes over time.

The fact that they need to market it. I don't remember a 22nm +++ but we still got quite a few gens on it and they still were improvements.

These process refinements are exactly like you say, everyone does it. So why mention it? Answer: because its all Intel has, and all the pluses are now, is just a few hundred mhz less base for a few hundred mhz more turbo. So its an empty shell and that is obviously what my sarcasm was about.

10nm intel may be denser, but is just a lot worse. Look at the ryzen 4000 mobile 15units, they all smash the ice lake mobile units, all of them, graphics as well. The tiger lake cpus (10+) are sill losing to the amd apus according to the leaks so far.

It kinda still echoes that Intel's design goals were way too difficult to achieve. And then they got stuck so deep there was no way back.
 
Last edited:
The fact that they need to market it. I don't remember a 22nm +++ but we still got quite a few gens on it and they still were improvements.

These process refinements are exactly like you say, everyone does it. So why mention it? Answer: because its all Intel has, and all the pluses are now, is just a few hundred mhz less base for a few hundred mhz more turbo. So its an empty shell and that is obviously what my sarcasm was about.


I don't know why they need to market its manufacturing process. No other industry does it.
Better focus on thermals, power consumption, price and performance :laugh:
 
If AVX not that important then why Intel was bragging about AVX-512 on Ice Lake(Sunny Cove) and Ice Lake currently availble on mobile.
And DRM software like Denuvo use FMA3 (which Lakefield lacks), with out this new instruction your cpu cant run new games. This days old but still capable Phenom II and Core 2 Quad cant run many game because of lack of new instruction like AVX and FMA3.
Again, I am repeating because you don't understand. Lakefield is a mobile chip. That is very low power, akin to Snapdragon 865/Apple A13. Regarding newer games, Lakefield doesn't have enough graphical horsepower to run these games anyway, so no need for this. You think you've spotted something than a few hundred Intel/Microsoft/partners from industry employees didn't, but trust me, the market is studied very well and 99% of the users won't try to run Control with RTX enabled on Lakefield devices.
 
I'm pretty sure these slides are Old, that's the ONLY way they make sense with respect to what we now know is true about comet lake and what could be true about rocket lake
 
The fact that they need to market it. I don't remember a 22nm +++ but we still got quite a few gens on it and they still were improvements.
These process refinements are exactly like you say, everyone does it. So why mention it? Answer: because its all Intel has, and all the pluses are now, is just a few hundred mhz less base for a few hundred mhz more turbo. So its an empty shell and that is obviously what my sarcasm was about.
I kind of agree with what you say, but given their stagnation with process along these years, they want to convey that they still updating the process, even if it not such a big change. You probably agree that the original 14nm wasn't no where near capable of running at 5.3Ghz like the 14nm++ (actually these was no +++, only press invents these pluses).
 
I kind of agree with what you say, but given their stagnation with process along these years, they want to convey that they still updating the process, even if it not such a big change. You probably agree that the original 14nm wasn't no where near capable of running at 5.3Ghz like the 14nm++ (actually these was no +++, only press invents these pluses).

Personally I believe Coffee Lake was the last real improvement. The only real plus after that was moving to solder out of pure necessity due to heat, ergo, they just pushed it over the edge.
 
I kind of agree with what you say, but given their stagnation with process along these years, they want to convey that they still updating the process, even if it not such a big change. You probably agree that the original 14nm wasn't no where near capable of running at 5.3Ghz like the 14nm++ (actually these was no +++, only press invents these pluses).


This is a product of the natural ramp up of the node and improved yields. You've got more dies with better characteristics.
Every node does it, and yet you hadn't got 32nm++ or 65nm++.
 
But Intel's 10nm is like TSMC's 7nm, no, wait, its probably better... and when they start adding pluses, you know AMD is doomed.

How about Intel releases a product that isn't 14nm+++++++++++++++++++++++ before saying AMD is doomed, eh?
 
How about Intel releases a product that isn't 14nm+++++++++++++++++++++++ before saying AMD is doomed, eh?
it could be worse, they could release 14nm------------------------------
 
How about Intel releases a product that isn't 14nm+++++++++++++++++++++++ before saying AMD is doomed, eh?

I see the sarcasm was lost on you :)
 
Again, I am repeating because you don't understand. Lakefield is a mobile chip. That is very low power, akin to Snapdragon 865/Apple A13. Regarding newer games, Lakefield doesn't have enough graphical horsepower to run these games anyway, so no need for this. You think you've spotted something than a few hundred Intel/Microsoft/partners from industry employees didn't, but trust me, the market is studied very well and 99% of the users won't try to run Control with RTX enabled on Lakefield devices.
Unlike the Intel cpu those ARM cpu dont lack NEON instruction.
 
I'm interested about it actually until I remember it's Intel making 16 cores with their mythical 10nm so it's gonna be mighty expensive
 
For what exactly do you need AVX in a mobile chip?
Lakefield uses the same idea as you said above, "both the Big and Small core (Sunny and Tremont) have same Instruction set, but with different throughput"
If you're thinking just ~5-7 W CPUs, then I assume no AVX support is fine for right now.
But there are numerous productive applications and a few games I believe which requires AVX, and usage of AVX is likely to increase since it offers massive performance gains and good efficiency. While I wouldn't recommend doing any real productive work on a laptop anyway, laptops still can get substantial performance gains from AVX, e.g. with laptops running Intel's Clear Linux (compiled with various AVX optimizations).

Adding basic AVX support shouldn't be hard, even in a smaller core. The full desktop architectures uses multiple units spread across three execution ports, a smaller setup can scale this down.
 
Let's see how they will complete. Not that I'm hoping for much but benchmarks are always better than theories.
 
2020 - Comet Lake 10 cores and Rocket Lake 8 cores on LGA-1200
2021 - Alder Lake BIG.little 8+8+1 cores on LGA-1700

Intel Alder Lake-S Desktop CPUs Will Only Feature Support on the New LGA 1700 Socket Platform, LGA 1200 To Last Two 2020 CPU Generations:
1588580335926.png

1588580299526.png


1588580360890.png




1588582729643.png

 
Last edited:
@ARF No need for that wall of text/pictures. A design so different always needs a new chipset/socket.
 
Back
Top