[AT] Intel trying to release 7 nm in 2021

Page 4 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

dmens

Golden Member
Mar 18, 2005
1,842
21
81
#76
you have more info than we it seems !

care to share with us ?
No info, just what I consider to be the most likely status based on my translation of Intel corporate FUD which I am still well familiar with even after resigning from that cesspool. As far as I can tell the originally spec'ed 10nm process is functionally abandoned and replaced with some hugely scaled back 10nm, just so management can tick a checkbox and say that 10nm "ramp" was achieved to try to boost shareholder confidence. The fact that the ramp was achieved with low margin SKU's with the tiniest possible dies shall not be mentioned. Large core count Xeons? Forget it. They'd be burning money with every wafer.

Also any talk of Intel 7nm being "as good as" TSMC 5nm should be disregarded until actual size/power/perf numbers are released. Since TSMC 5nm will be physical reality long before Intel 7nm, that is a lot of waiting.
 

dmens

Golden Member
Mar 18, 2005
1,842
21
81
#77
Their processor architects haven't been on vacation this whole time, the new stuff they've been working on will be rolled out rapidly. Don't underestimate their "CPU design" just because they've been stuck on Skylake for years.
LOL.

That is all.
 

jpiniero

Diamond Member
Oct 1, 2010
6,327
226
126
#78
As far as I can tell the originally spec'ed 10nm process is functionally abandoned and replaced with some hugely scaled back 10nm, just so management can tick a checkbox and say that 10nm "ramp" was achieved to try to boost shareholder confidence.
There is no evidence so far they have done any kind of spec changes to the original 10 nm designs. If they had done that, maybe there would be real hope of achieving the kind of yield needed to move away from 14.
 

TheGiant

Senior member
Jun 12, 2017
365
47
76
#79
Mark, please don't cherry pick

a 95W limit CPU (scroll down to real system power, and btw the system power of 2600X system is the same as the power of the 9900K 95W limited system, the 2700X is higher) it is like to 10% faster (except pure avx loads) so from pure efficiency 9900K is in fact most efficient high end desktop CPU on the market
if techspot really did a benchmark where whole system of 9900K will consume the same as 2700X, the 9900K will be double digit more powerful

but what power is in 9900K is the ability to clock much higher (with more RAM support and higher RAM speeds), but that comes at the heat and power cost (that is why I didnt buy it)

9900K doesn't have a competition on the market right now

it would be stupid by Intel not to ask much more for it, from other point of view AMD should be happy Intel just didnt replace their previous flagship 8700K with 9900K at 370 EUR price
if they did, AMD will be pushed to sub 300, real 250 EUR prices and that a real profit cut

here comes the 7nm age and this thread is about it

what Intel can do is to let the 10nm pieces benchmark "accidentally leak" and when its ready 7nm too

as they did with this annoucement- marketing
 

Topweasel

Diamond Member
Oct 19, 2000
4,768
363
136
#80
Still, a stupid argument that lacks actual data. The main point is that 90% of the end users DO NOT CARE or even know what 10nm or 7nm means. You keep making these statements that Intel hasn't innovated for 8 years, for over 3 years TSMC is doing all these incredible things and Intel is just stagnate.

But you fail to realize that Intel still owns 90% of the over all market, despite all these "downfalls" you claim. Because, be brutally honest with yourself, no damn normal person has a clue how many nanometer their chip is, lol. Hell, I don't even know. We've for Ryzen 7 2700X in my son's gaming computer and I have an Intel 9700K. There's virtually no difference. I chose Intel because I do video production and it's a far superior chip for rendering and AE. We chose Ryzen for my son's gaming computer. We don't actually care about the die size or manufacturing process. If it performs, it performs.

It's just fanboys trying to act like "their" company is better. It's pretty stupid, if you ask me. But Intel hasn't lost a tremendous amount of market share and they will catch back up.
Did you read what I wrote. This isn't a Financial assessment of Intel. While I do think they will struggle a bit in the future financially as their server parts for the next year or two will be primarily dual huge dies that they will sell out of but will eat a lot of margin and slow down availability of all parts. As long as they are a Primarily 14nm supplier as they expand core count (like a 10c Commet Lake) they will take an already strained process and strain it more. That's why they are tossing in another 3-5 billion on expanding 14nm production. But and this is the part you missed my statement was completely based not on sales and profitability but other performance benchmarks like processing tech, manufacturing ability, ability to stick to roadmaps, general CPU performance leads and so on. They have stumbled mightedly and if you can't see that then you are the fanboy.


Intel has done plenty in the past 8 years, it's just been held back because of the 10nm issues. Their processor architects haven't been on vacation this whole time, the new stuff they've been working on will be rolled out rapidly. Don't underestimate their "CPU design" just because they've been stuck on Skylake for years.
Not buying it. There have been a few things like widening the pipline, redoing cache structure and adding things like AVX2 and AVX 512. But they haven't had a major arch change since Sandybridge.
 

dmens

Golden Member
Mar 18, 2005
1,842
21
81
#81
There is no evidence so far they have done any kind of spec changes to the original 10 nm designs. If they had done that, maybe there would be real hope of achieving the kind of yield needed to move away from 14.
And there is no public evidence that 10nm hasn't been blown up. Intel will never admit it but given business pressure and engineering reality that is the most rational course of action.
 

jpiniero

Diamond Member
Oct 1, 2010
6,327
226
126
#82
And there is no public evidence that 10nm hasn't been blown up. Intel will never admit it but given business pressure and engineering reality that is the most rational course of action.
I think they've been literally spinning their wheels trying to get yield up to a reasonable level and have essentially given up and will just produce what they can without killing the margins too much.
 

Markfw

CPU Moderator, VC&G Moderator, Elite Member
Super Moderator
May 16, 2002
17,593
1,060
136
#83
Mark, please don't cherry pick

a 95W limit CPU (scroll down to real system power, and btw the system power of 2600X system is the same as the power of the 9900K 95W limited system, the 2700X is higher) it is like to 10% faster (except pure avx loads) so from pure efficiency 9900K is in fact most efficient high end desktop CPU on the market
if techspot really did a benchmark where whole system of 9900K will consume the same as 2700X, the 9900K will be double digit more powerful

but what power is in 9900K is the ability to clock much higher (with more RAM support and higher RAM speeds), but that comes at the heat and power cost (that is why I didnt buy it)

9900K doesn't have a competition on the market right now

it would be stupid by Intel not to ask much more for it, from other point of view AMD should be happy Intel just didnt replace their previous flagship 8700K with 9900K at 370 EUR price
if they did, AMD will be pushed to sub 300, real 250 EUR prices and that a real profit cut

here comes the 7nm age and this thread is about it

what Intel can do is to let the 10nm pieces benchmark "accidentally leak" and when its ready 7nm too

as they did with this annoucement- marketing
Cherry pick ? I took the quote from the conclusion of YOUR link. Its virtually the same performance, but 70% higher cost for the 9900k.
 

dmens

Golden Member
Mar 18, 2005
1,842
21
81
#84
I think they've been literally spinning their wheels trying to get yield up to a reasonable level and have essentially given up and will just produce what they can without killing the margins too much.
If that is the case then the claim of volume ramp is likely a complete lie and the management should be sued for corporate fraud.
 

jpiniero

Diamond Member
Oct 1, 2010
6,327
226
126
#85
If that is the case then the claim of volume ramp is likely a complete lie and the management should be sued for corporate fraud.
I imagine they will run the manufacturing lines as if it was HVM, but of course only so many of the chips will actually be usable. Remember that the redirected most of the production lines intended for 10 back to 14, with some others retooled for EUV.
 

Zucker2k

Senior member
Feb 15, 2006
757
65
136
#86
Looks to me like they trade blows. The 2700x wins the very first benchmark. It looks like its between the 95 watt and the unlimited benches on average.

And from your own link, in the conclusion:
It’s a $500 8-core desktop CPU competing with a $300 8-core desktop CPU. As we just saw with the 95-watt limit, it’s barely any faster than the Ryzen 7 2700X. In fact, in some tests it’s slower, and that’s an awful result for a CPU that costs ~70% more. "
All the bolded are not true. I don't even know what link you clicked on. The 95 watt 9900k won the first 9 benchmarks and barely loses by 0.7 points in the 10th benchmark (while consuming a whopping 33 watts less than the 2700x in that particular benchmark). Whut?! It then goes on to win the next two benchmarks, making a total of 11 wins and 1 loss.
That's a total domination from the performance and architectural point of view. Bear in mind, these benchmarks are mostly in heavy multithreaded apps where the AMD chip does better because of its bigger power budget. In single threaded scenarios, the 5Ghz boosting 9900k should open up the gap even further.
https://www.techspot.com/review/1744-core-i9-9900k-round-two/
 

Markfw

CPU Moderator, VC&G Moderator, Elite Member
Super Moderator
May 16, 2002
17,593
1,060
136
#87
All the bolded are not true. I don't even know what link you clicked on. The 95 watt 9900k won the first 9 benchmarks and barely loses by 0.7 points in the 10th benchmark (while consuming a whopping 33 watts less than the 2700x in that particular benchmark). Whut?! It then goes on to win the next two benchmarks, making a total of 11 wins and 1 loss.
That's a total domination from the performance and architectural point of view. Bear in mind, these benchmarks are mostly in heavy multithreaded apps where the AMD chip does better because of its bigger power budget. In single threaded scenarios, the 5Ghz boosting 9900k should open up the gap even further.
https://www.techspot.com/review/1744-core-i9-9900k-round-two/
Wrong. This is a perfect example. The first benchmark is cinebench 15. If you figure the 95 watt is what its supposed to be stock, the 2700x wins. But it looses to the unlimited setting. A perfect example of between the 2. I am not going through every benchmark and count who wins how many, but pushing this arguement constitutes a callout, since its linked, its not opinion.

Edit:Not counting gaming, in the top section, its between in 4 benchmarks. And again from the conclusion MY BIG POINT
"As we just saw with the 95-watt limit, it’s barely any faster than the Ryzen 7 2700X. In fact, in some tests it’s slower, and that’s an awful result for a CPU that costs ~70% more. "
 
Last edited:

Zucker2k

Senior member
Feb 15, 2006
757
65
136
#88
Wromg. This is a perfect example. The first benchmark is cinebench 15. If you figure the 95 watt is what its supposed to be stock, the 2700x wins. But it looses to the unlimited setting. A perfect example of between the 2. I am not going through every benchmark and count who wins how many, but pushing this arguement constitutes a callout, since its linked, its not opinion.
Actually, you're right about the first benchmark. I had a brain a fart there. My apologies.
However, saying they traded blows when the constrained 9900k won 8 out of 12 benchmarks is really a refusal to acknowledge facts.

Edited to account for my earlier mistakes.
 

Markfw

CPU Moderator, VC&G Moderator, Elite Member
Super Moderator
May 16, 2002
17,593
1,060
136
#89
Actually, you're right about the first benchmark. I had a brain a fart there. My apologies.
However, saying they traded blows when the constrained 9900k won 10 out of 12 benchmarks is really a refusal to acknowledge facts.
Since it beat it on at least 4 , how can that be ? Cinebench, 7-Zipm handbrake x264, Adobe premiere.
 

Zucker2k

Senior member
Feb 15, 2006
757
65
136
#90
Since it beat it on at least 4 , how can that be ? Cinebench, 7-Zipm handbrake x264, Adobe premiere.
Corrected. I swear, I'm having a bad day here. But the point still stands. The 9900k won 100% more of the benchmarks. That's 8 to 4. If the shoe was on the other foot, you won't be calling for a draw. The main point here being, the 2700x used more power and still lost 8 out of 12.
 

Markfw

CPU Moderator, VC&G Moderator, Elite Member
Super Moderator
May 16, 2002
17,593
1,060
136
#91
Corrected. I swear, I'm having a bad day here. But the point still stands. The 9900k won 100% more of the benchmarks. That's 8 to 4. If the shoe was on the other foot, you won't be calling for a draw. The main point here being, the 2700x used more power and still lost 8 out of 12.
Did I ever say it won everything ? I said its close, and about 1/2 the price. And it only wins a lot when it does use a lot more power. And I will quote my main point a 3rd time, from the conclusion in the article:
""As we just saw with the 95-watt limit, it’s barely any faster than the Ryzen 7 2700X. In fact, in some tests it’s slower, and that’s an awful result for a CPU that costs ~70% more. " "

How many times do I have to make my point and re-quote this ? And this was not even my link, it was @TheELF 's.
 

JDG1980

Golden Member
Jul 18, 2013
1,581
67
136
#92
nm means nothing if it's not supported by the clocks, nobody is going to buy 7nm or 5nm if they are clocked at 4Ghz.
The vast majority of x86 chips sold run at maximum speeds lower than 4GHz, often significantly so. Given modern x86 IPC, and the ability of most current software to make at least some use of multithreading, these kind of high clock rates are really only needed on a handful of gaming/enthusiast PCs, and in a few specialized applications.

Yes intel would be pretty stupid if they released a 10nm chip at 4Ghz or even 4,5Ghz ,why on earth would they release a 10nm chip that is slower then their current one?
Because with better perf/watt and higher transistor density, it would be a step forward in laptops and servers - both markets significantly larger than enthusiast desktops, and both markets where chips hardly ever need to run individual cores at clock rates that high.

If Intel could do this with a 10nm product at acceptable costs and acceptable yields, they already would have released it.
 

Zucker2k

Senior member
Feb 15, 2006
757
65
136
#93
Did I ever say it won everything ? I said its close, and about 1/2 the price. And it only wins a lot when it does use a lot more power. And I will quote my main point a 3rd time, from the conclusion in the article:
""As we just saw with the 95-watt limit, it’s barely any faster than the Ryzen 7 2700X. In fact, in some tests it’s slower, and that’s an awful result for a CPU that costs ~70% more. " "

How many times do I have to make my point and re-quote this ? And this was not even my link, it was @TheELF 's.
I bolded the part of your post I was replying to. Here, I'll post it again

"Looks to me like they trade blows. The 2700x wins the very first benchmark. It looks like its between the 95 watt and the unlimited benches on average."

You're suggesting that:
1. They trade blows
2. That the 2700x may actually be faster than the 95 watt 9900k

I'm arguing that based on a result of 2:1 in favor of the 95w 9900k in the linked benchmarks, your suggestion seems to be ignoring the actual data in the links. The order in performance is:
1. 9900k Unleashed
2. 9900k Tamed (95w)
3. 2700x

When your opponent is punching you twice as much as you are punching them, you're not trading blows. You're getting pummeled into "bolivion," to quote the great Mike Tyson.
 
Jul 25, 2001
10,189
25
91
#94
9700k is actually the better overall chip and isn't as expensive as the 9900k. I purchased the 2700X and then found out, after the fact, that it doesn't perform nearly as well in Adobe CC. That's even taking into account that the 9700k doesn't have hyper threading! But the Intel wins in every single gaming benchmark.

https://www.tomshardware.com/news/ryzen-7-vs-core-i7-9700k,38046.html

"The battle between Intel’s Core i7-9700K and AMD’s Ryzen 7 2700X gets much more interesting when you factor productivity into the equation. AMD’s Ryzen 7 2700X is better at certain tasks, such as rendering with Cinebench and Blender and compressing files with 7-Zip. However, some tasks, such as encoding with Lame or Handbrake, run faster on Intel’s Core i7-9700K. "

"Intel’s latest CPUs also favor Adobe’s Creative Cloud suite by a considerable margin."

9700K actually out performs the 2700X in Illustrator, Premiere Pro and it beats it up in After Effects. AE is a whole different story and really what started me down the path of the Intel CPU in the first place. I ended up giving my 2700x to my son and built a gaming rig with him. So I've used both and we have both in our household.
 

TheGiant

Senior member
Jun 12, 2017
365
47
76
#95
I bolded the part of your post I was replying to. Here, I'll post it again

"Looks to me like they trade blows. The 2700x wins the very first benchmark. It looks like its between the 95 watt and the unlimited benches on average."

You're suggesting that:
1. They trade blows
2. That the 2700x may actually be faster than the 95 watt 9900k

I'm arguing that based on a result of 2:1 in favor of the 95w 9900k in the linked benchmarks, your suggestion seems to be ignoring the actual data in the links. The order in performance is:
1. 9900k Unleashed
2. 9900k Tamed (95w)
3. 2700x

When your opponent is punching you twice as much as you are punching them, you're not trading blows. You're getting pummeled into "bolivion," to quote the great Mike Tyson.
it is not 2:1 really, lets look at the benchces
for me a difference of 5% +- is a tie, you won't even see it in real world

1. cinebech tie
2. blender tie
3. corona 9900K small win (<10%)
4. vray tie
5. zip compress- 9900K wins
6. zip decompress- tie
7. excel 9900k wins
8.h265 9900k wins
9.h264 tie
10. premiere tie

gaming is 9900K win or tie

but with real power budget, while the ryzen 2700X (105W and other intel chips, like 8700K) doesnt hold to its power budget (ofc TDP isnt officially power bla bla, but test is made with power limit)

give that 9900K system a chance to consume 210W not 175W and things will be very different

need to have 7nm to get back with power to that 7700K while having 8 cores at the performance of CFL 5GHz

everywhere you see green world....use less resources, my whole house lights use the same power that my table light back in highschool years, yet we increase the power of desktop systems again

cpu- rise (not even counting that 500W 28C xeon)
GPU- total rise

I have a undervolted 6600K oced to 4.4GHz - while stress testing in intel extreme utility the package TDP fluctuates between 41 to 47W, system bought january 2016
to have more than 10% performance (to notice anything, barely) with first 4 threads I need to get 8600k+ and oc it to 5GHz, which means going to like 120W+ depending on the chip

WTF really, 3,5 years and you can get a sidegrade or upgrade at the cost 3x the power....
 

Zucker2k

Senior member
Feb 15, 2006
757
65
136
#96
it is not 2:1 really, lets look at the benchces
for me a difference of 5% +- is a tie, you won't even see it in real world


1. cinebech tie
2. blender tie
3. corona 9900K small win (<10%)
4. vray tie
5. zip compress- 9900K wins
6. zip decompress- tie
7. excel 9900k wins
8.h265 9900k wins
9.h264 tie
10. premiere tie

gaming is 9900K win or tie

but with real power budget, while the ryzen 2700X (105W and other intel chips, like 8700K) doesnt hold to its power budget (ofc TDP isnt officially power bla bla, but test is made with power limit)

give that 9900K system a chance to consume 210W not 175W and things will be very different

need to have 7nm to get back with power to that 7700K while having 8 cores at the performance of CFL 5GHz

everywhere you see green world....use less resources, my whole house lights use the same power that my table light back in highschool years, yet we increase the power of desktop systems again

cpu- rise (not even counting that 500W 28C xeon)
GPU- total rise

I have a undervolted 6600K oced to 4.4GHz - while stress testing in intel extreme utility the package TDP fluctuates between 41 to 47W, system bought january 2016
to have more than 10% performance (to notice anything, barely) with first 4 threads I need to get 8600k+ and oc it to 5GHz, which means going to like 120W+ depending on the chip

WTF really, 3,5 years and you can get a sidegrade or upgrade at the cost 3x the power....
I agree with you, and it's not just for you, statistical ties and dead heats do exist. It's just that the partisanship on Anandtech leaves little room for such considerations. You acknowledge a loss, no matter how tiny the margin or get called out for your 'bias.' Do you remember when AMD demoed Zen2 against the 9900k in CB15. Go look at the Ryzen Speculation thread. In any case, a 4:0 win is even better, hehe.
 

Markfw

CPU Moderator, VC&G Moderator, Elite Member
Super Moderator
May 16, 2002
17,593
1,060
136
#97
I agree with you, and it's not just for you, statistical ties and dead heats do exist. It's just that the partisanship on Anandtech leaves little room for such considerations. You acknowledge a loss, no matter how tiny the margin or get called out for your 'bias.' Do you remember when AMD demoed Zen2 against the 9900k in CB15. Go look at the Ryzen Speculation thread. In any case, a 4:0 win is even better, hehe.
So the Giant agrees with me, then you agree with him, so why all the crap that you gave me ?
 

Zucker2k

Senior member
Feb 15, 2006
757
65
136
#98
So the Giant agrees with me, then you agree with him, so why all the crap that you gave me ?
So the Giant agrees with me, then you agree with him, so why all the crap that you gave me ?
How does he agree with you? You set the tone and I went along with it. Everywhere you claimed a win, he gave it a tie. In the end, the 9900k came out ahead with the same 4-win difference over the 2700x even if we apply the "a win is a win" filter. This is a clear win for the same chip you suggest is below the 2700x in performance when restricted to a 95w tdp. Moreover, this same 95w chip dominates the 2700x in gaming and almost every single threaded scenario I can think of. This is the best the 2700x is going to look and yet it loses in 50% of the benchmarks so what's your argument again?
 

TheELF

Platinum Member
Dec 22, 2012
2,817
111
126
#99
Did I ever say it won everything ? I said its close, and about 1/2 the price. And it only wins a lot when it does use a lot more power. And I will quote my main point a 3rd time, from the conclusion in the article:
""As we just saw with the 95-watt limit, it’s barely any faster than the Ryzen 7 2700X. In fact, in some tests it’s slower, and that’s an awful result for a CPU that costs ~70% more. " "

How many times do I have to make my point and re-quote this ? And this was not even my link, it was @TheELF 's.
Achoo...someone's thinking about me...
No it wasn't my link but the giant's I believe.

Also why are you focusing solely on the one side that is suiting you?
Yes ryzen is a low power design and competes very nicely at that low power.
On the other hand how much would it cost to get a 2700x up to 5Ghz?How much power would it consume?
Oh it's impossible?Well then 70% more for a CPU that does these things at stock seems like a steal to me.
You are paying 70% more to get something that no other CPU in the world can give you,if you don't need it you don't have to buy it and that's the opinion of whoever wrote that article, but if you do need it,really need it, for work or whatnot that 70% is nothing, it's paying for itself if you use something that runs much faster at 5Ghz.
 

TheELF

Platinum Member
Dec 22, 2012
2,817
111
126
Because with better perf/watt and higher transistor density, it would be a step forward in laptops and servers - both markets significantly larger than enthusiast desktops, and both markets where chips hardly ever need to run individual cores at clock rates that high.
If Intel could do this with a 10nm product at acceptable costs and acceptable yields, they already would have released it.
Yes and wonder of wonders this post is about intel announcing that they will bring 10nm to servers first...
Desktop is the moneymaker for intel,they make more from PC centric than from everything else combined.
https://www.intc.com/investor-relat...t-Quarter-2019-Financial-Results/default.aspx
 


ASK THE COMMUNITY