• Guest, The rules for the P & N subforum have been updated to prohibit "ad hominem" or personal attacks against other posters. See the full details in the post "Politics and News Rules & Guidelines."

Discussion Intel current and future Lakes & Rapids thread

Page 182 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

IntelUser2000

Elite Member
Oct 14, 2003
7,250
1,839
136
Where exactly did they write the gains are IPC exclusive? Fine-tuning an architecture can be done to improve fmax as well, not just IPC.
Exactly. Its details such as these we'll likely never get to hear about but nevertheless significant.
 

SAAA

Senior member
May 14, 2014
496
92
91

I'm sure IT World Canada is incorrect and those gains are from IPC and clocks, not through architecture alone. >10% IPC gain would be very unexpected.
Icelake is a much better base to build upon than Skylake so I'm not surprised if they found another 10% from µ-arch alone: there's between 25 to 50% raw increase in many areas inside Ice cores and that can be worked upon, Tiger lake has increased caches size and different subdivision, who knows what else.

Intel themselves posted that chart with 18% -average- IPC increase: there were tests ranging from a few to +40% normalized scores. What didn't benefit from Icelake increases (0-5%) might do from Tiger Lake implementations so the average could jump up a lot.

Also the fact that on the same process the cores are larger (they basically take twice as much space as Zen cores) I'm sure heat density won't be an issue at all with 10nm compared to TSMC 7nm, there's something else keeping the clocks low.
I'm certain they planned for 10 nm node to achieve 5+ GHz as 14 nm so the design is fine but the execution totally lacking, when they manage to find what's wrong with the process it will shine, for performance at least, maybe not so much power efficiency.
 

IntelUser2000

Elite Member
Oct 14, 2003
7,250
1,839
136
Also the fact that on the same process the cores are larger (they basically take twice as much space as Zen cores)
Zen 2 is 3.6mm2. Icelake is 5.6mm2. Intel's L3 layout is more integrated with the core in Icelake and beyond but the 6.9mm2 number from Wikichip is with the L3 cache. L2 is pretty much part of the core for both.

I'm sure heat density won't be an issue at all with 10nm compared to TSMC 7nm, there's something else keeping the clocks low.
Pretty sure process plays a part, but beyond 4GHz, its much less so process than its about thermals. That's why Cometlake-S is rumored to be at 300W. They want a 10 core to reach 5GHz. If they just backed off on the clocks by 10%, the power usage will drop drastically.

Also, Intel had the same core/process since 2015. Like with anything, if you work on it long enough, you can polish it to a state not previously possible.

Let's say in the future we're willing to accept 1000W CPUs, then clocks can rise close to 6GHz. Is it worth it? Likely, definitely not.
 
Last edited:

liahos1

Senior member
Aug 28, 2013
573
45
91
Thought: tigerlake is 146mm^2. 9900k is 177mm^2

given the increase in density at 10nm vs 14nm, why doesn’t intel release a 10nm equivalent of the 9900k.

don’t say clocks - tigerlake clocks have improved. A 95w tdp at 4.5ghz would outperform 9900ks in single thread

don’t say yields. Laptop market is 2x the size the Desktop market.

So why? Maybe intel just doesnt care about desktops anymore given they control enterprise desktop and diy is a niche within a niche.

come at me bros



This is trolling / inflammatory posting.

So no, we won't come at you bro.


esquared
Anandtech Forum Director
 
Last edited by a moderator:
  • Like
Reactions: Richie Rich

Markfw

CPU Moderator, VC&G Moderator, Elite Member
Super Moderator
May 16, 2002
21,149
9,267
136
Thought: tigerlake is 146mm^2. 9900k is 177mm^2

given the increase in density at 10nm vs 7nm, why doesn’t intel release a 10nm equivalent of the 9900k.

don’t say clocks - tigerlake clocks have improved. A 95w tdp at 4.5ghz would outperform 9900ks in single thread

don’t say yields. Laptop market is 2x the size the Desktop market.

So why? Maybe intel just doesnt care about desktops anymore given they control enterprise desktop and diy is a niche within a niche.

come at me bros
They control enterprise desktop ? In your dreams.
 
  • Like
Reactions: Drazick

Markfw

CPU Moderator, VC&G Moderator, Elite Member
Super Moderator
May 16, 2002
21,149
9,267
136
Yeah they do. Lol are you joking with me. I just met with senior execs at dell and hpe. Why don’t you keep your troll comments to yourself?
You mean they currently have more desktops that are Intel ? probably... Do they control it ? No. Do the execs think they control it ? Most likely. Upper management more of the time has no clue what is really going on.

You and them are free to have your delusions.
 

liahos1

Senior member
Aug 28, 2013
573
45
91
Yes they absolutely control enterprise desktop. The proof is in the numbers. The numbers are based on corporate inertia. But none of this has to do with the question I posed.
 

jpiniero

Diamond Member
Oct 1, 2010
8,911
1,645
126
don’t say yields. Laptop market is 2x the size the Desktop market.
Yields are that bad. We're talking maybe getting 10-20% out of a wafer, and that's including the cut down models. But that's just an estimate... it could be worse than that.
 

mikk

Diamond Member
May 15, 2012
3,044
850
136
Yields are that bad. We're talking maybe getting 10-20% out of a wafer, and that's including the cut down models. But that's just an estimate... it could be worse than that.
You have no source as usual I guess, this is the (your) problem. What happened to your LGA 1159 by the way? Debunked?
 

coercitiv

Diamond Member
Jan 24, 2014
4,184
5,120
136
So... Intel isn't making 8-10 core 10nm TGL chips in the near future because they don't wanna, not because they can't.

It feels like we're in some kind of ludicrous claim contest here. On today's menu we find out Intel is full of unstable geniuses which have both the know-how and the means to mass-produce TGL chips for the entire computing market, but lack the basic desire to do so.
 

jpiniero

Diamond Member
Oct 1, 2010
8,911
1,645
126
You have no source as usual I guess, this is the (your) problem.
Let me put it this way, if yields were even just terrible there would be no shortage. Now either the OEMs are going to have to suck it up or use AMD. I think you might see both.

Remember you can still get a decent amount of chips if you are willing to burn the wafers, even at 10-20% yield. But it's not that much when you need enough for 260M PCs a year.
 

liahos1

Senior member
Aug 28, 2013
573
45
91
So... Intel isn't making 8-10 core 10nm TGL chips in the near future because they don't wanna, not because they can't.

It feels like we're in some kind of ludicrous claim contest here. On today's menu we find out Intel is full of unstable geniuses which have both the know-how and the means to mass-produce TGL chips for the entire computing market, but lack the basic desire to do so.
Maybe thats the case? It's the smallest market within the pc space and the one that is declining the fastest. Intel is capacity constrained.

A tick of the 9900k would be a pretty small die. Smaller than TGL-U for sure. I'm just trying to go through the thought processes.
 

lobz

Golden Member
Feb 10, 2017
1,577
2,023
136
Exactly. Its details such as these we'll likely never get to hear about but nevertheless significant.
I gotta tell you, everyone who doesn't want Intel to have even bigger problems than they have right now, they / we better hope that TGL is at least a double digit increase over ICL on a laptop to laptop basis, seeing how the mere existence of 10th gen 14nm chips is making ICL nothing more than a cool technical demo demonstrating Foveros right now. It's nice to have them, but nobody would ever really miss them.

You have no source as usual I guess, this is the (your) problem. What happened to your LGA 1159 by the way? Debunked?
There is no need for calling out someone on an assumption which they already acknowledged as wrong. The socket naming confusion originated from a specific piece of software, that is usually very reliable, but this time it turned out to be either a placeholder or a typo. Bringing this up to him being so smug is pointless, it means nothing for the point you're trying to make, because of how different the 2 topics are.
The low yield assumptions are coming from inside sources (not mine, of course, but Charlie Demerjian's) followed by somewhat basic math (the deepening and persistence of 14nm shortage, which seems to be completely unaffected by 10nm, despite it being officially praised as 'high volume production'), and most importantly, all this backed up by real life: it's Q1 2020 and Ice Lake server parts are nowhere to be found. It's Q1 2020 and half the ICL mobile lineup is yet to be seen in anything more than a slideshow. I don't even see the point in talking about Tiger Lake all the time, when the rest of the ICL mobile lineup will enter the market only now in Q1 (confirmed by Intel). I totally believe that intel is throwing a maaaaaassssive number of 10nm wafers into production, even if it means effectively losing money on every 10nm chip they ship, because the execs seem to be worried about nothing but short time stock market perception. Sure, let's call it HVM but it clearly doesn't mean the high volume of actual usable chips returned. Otherwise there would be a significantly lower number of low core count 14nm laptop designs (meaning up to 4 cores), which is the VAST majority of the whole mobile market and ICL could have flooded that market. That alone would have alleviated the 14nm shortage, because we're not talking about a shortage of 40-50% higher demand than what intel can deliver. It's almost an order of magnitude lower than that.

If 10nm yields were even a bit better than what @jpiniero says, the 14nm shortage would be long gone now, coinciding ICL's market debut.

These are still all small, LP mobile chips, imagine what a financial disaster every single monstrous Ice Lake server chip is they have to make. And they have to.
 
Last edited:
  • Like
Reactions: spursindonesia

Gideon

Golden Member
Nov 27, 2007
1,196
2,200
136
Maybe thats the case? It's the smallest market within the pc space and the one that is declining the fastest. Intel is capacity constrained.

A tick of the 9900k would be a pretty small die. Smaller than TGL-U for sure. I'm just trying to go through the thought processes.
1. The desktop market might be small but it's acutally growing.
2. > 4core chips are also heavily used in laptops (15-45W)

Even just ta 6-core Tiger-lake would address a market way larger than the 4-core ever could. It could even work for desktop gaming due-to the IPC increases (as they can already clock it to 4.4 on mobile, ~4.5 would be totally enough to compete with 9900K).

The ony reason I can explain the lack of at least a 6-core Tiger Lake on the roadmap is ... yields. Especially if you consider that they plan to make the 8-core Rocket Lake (with supposedly the same core @ 14nm). Why would they do that if they had 0 manufacturing issues?
 
  • Like
Reactions: spursindonesia

liahos1

Senior member
Aug 28, 2013
573
45
91
1. The desktop market might be small but it's acutally growing.
2. > 4core chips are also heavily used in laptops (15-45W)

Even just ta 6-core Tiger-lake would address a market way larger than the 4-core ever could. It could even work for desktop gaming due-to the IPC increases (as they can already clock it to 4.4 on mobile, ~4.5 would be totally enough to compete with 9900K).

The ony reason I can explain the lack of at least a 6-core Tiger Lake on the roadmap is ... yields. Especially if you consider that they plan to make the 8-core Rocket Lake (with supposedly the same core @ 14nm). Why would they do that if they had 0 manufacturing issues?
thats not true. pc market grew in 2019 based on windows 7 eol support. If you look at last 5 years of data the desktop market shrunk considerably faster than the notebook market.
 

Hitman928

Diamond Member
Apr 15, 2012
3,409
3,561
136
do you live in a world of numbers?

Epyc revenues grew 80mln in q3 from q2.
Intel grew DCG 1.4bln.

That is your definition of eating alive?
We've been through this before.

Intel's Q2 was unusually low in terms of both revenue and volume due to companies being concerned about the trade war and so they stockpiled parts in 4Q of 2018 which left 1H 2019 lite on sales. This stockpile cleared out by 3Q 2019, so a large increase from Q2 to Q3 was more of them returning to normal than growing. Actually, the Y/Y DCG volume shrunk by 6%.



Whereas Epyc volume grew over 50% from Q2 to Q3 and grew Y/Y as well (though no specifics were given). Yes, Intel is outselling AMD in every market (except DiY where AMD appears to be dominating), but it's obvious that AMD's share continues to grow overall and Epyc continues to increase in both revenue and volume while Intel is struggling to stop AMD's progress. Throwing their current revenue around is a straw man, no one is disputing this. It's how much AMD is growing while Intel's market share is shrinking.

If you go back and look at the timeline with Opteron, AMD is basically following the same curve, though maybe a quarter or two behind. If they continue to follow the same curve then by this time in 2021 AMD will have ~20% of the server market and then, if they don't slip up like last time, another 10% within a year.
 

lobz

Golden Member
Feb 10, 2017
1,577
2,023
136
1. The desktop market might be small but it's acutally growing.
2. > 4core chips are also heavily used in laptops (15-45W)

Even just ta 6-core Tiger-lake would address a market way larger than the 4-core ever could. It could even work for desktop gaming due-to the IPC increases (as they can already clock it to 4.4 on mobile, ~4.5 would be totally enough to compete with 9900K).

The ony reason I can explain the lack of at least a 6-core Tiger Lake on the roadmap is ... yields. Especially if you consider that they plan to make the 8-core Rocket Lake (with supposedly the same core @ 14nm). Why would they do that if they had 0 manufacturing issues?
It really is about believing more in your own sanity than in anything Intel officially states about things they know you can never check for yourself.
They are backporting an otherwise excellent new uarch to 14nm. In 2021...
This says to me that yields are not just bad, but that it's not even viable for a high volume product launch at all.
7nm to the rescue! :)
 
  • Like
Reactions: krumme

ASK THE COMMUNITY