Intels new 10 Core?! - How true is this leak do you think?

sao123

Lifer
May 27, 2002
12,648
201
106
Interestingly enough, according to the table in the article I read, all 4 variations i3, i5, i7, i9 are all back to hyperthreading on... The i7-10700k is an 8C/16T part again.
 
Last edited:

Markfw

Moderator Emeritus, Elite Member
May 16, 2002
25,483
14,434
136
https://www.forbes.com/sites/antony...-including-5-2ghz-10-core-beast/#1c9dd0571f2a

This has just been released by Forbes. How true do you all think this is? Is this a helpless attempt at staying relevant? What about the whole fact you would have to change entire socket as well??

What do you all think?
Possible... But by the time they come out, threadripper 3000 series will probably be out, and destroy anything they can come up with for the same price. A 64 core threadripper ? Even at $3000 it would be outrageous. And based on what we have seen, no memory problems like 2990wx and 2970wx. And the 32 core will trounce the 2990wx, and be less than $2000
 
  • Like
Reactions: Drazick

IEC

Elite Member
Super Moderator
Jun 10, 2004
14,323
4,904
136
Predictable response. I bought my R9 3900X knowing what was expected from Intel. More 14nm+++ squeezed to the max. Which isn't a bad product by any means, but it's overdue for a shrink. Yet another socket change is a real bummer for Z390 folks, though.

As for the competitive landscape... 16c/32t R9 3950X on AM4 in September. And Threadripper 3000-series waiting in the wings likely in Q4, this time with few to none of the downsides of previous Threadrippers due to the I/O die + chiplet configuration. And with every expectation of a Zen 7+EUV refresh in 2020, I fully expect AMD to take the unequivocal lead with Zen 3.
 

BigDaveX

Senior member
Jun 12, 2014
440
216
116
The socket being listed as "LGA1159" does make me wonder if this slide might be a fake. I just can't think of any good reason why Intel would change sockets again so quickly, unless they're making alterations to allow them to start cobbling together Pentium D/Core 2 Quad-style MCM chips.
 
  • Like
Reactions: XiandreX

fleshconsumed

Diamond Member
Feb 21, 2002
6,483
2,352
136
The TDP numbers make me question the chart. i3-10300 has higher clocks than i3-10100 but has 62W TDP as opposed to 65W TDP. Similarly i5-10400 has same all core boost as i3-10100 and yet they both have same TDP rating of 65W despite i5 having 50% more cores. Either the TDP numbers are bogus or this whole chart is bogus.

But assuming the chart is true it's nice to see Intel bring back HT for all models and actually price their CPUs competitively, which funnily enough makes me question the authenticity of this chart yet again. Even if it's true I'm sticking with AMD. Not having to buy a new motherboard when upgrading CPU is really nice. I have 5 desktops at home and having one unified AM4 socket that can handle anything from 2c4t to 16c32t across multiple generations when I shuffle parts around is a big plus for me.
 

CHADBOGA

Platinum Member
Mar 31, 2009
2,135
832
136
Will Socket LGA1159 motherboards come with PCI 4?

How hot will this 10 core run? Doesn't the 9900K have serious heat issues?

Unless Intel has silicon fixes to mitigate the various security flaws that are still hitting their range, then I can't see this chip being a winner for them.

Even if Intel gets a few percent IPC bump from security mitigation in hardware, AMD will have a 16 core CPU on offer and may choose to move their CPU's down a price level or two, so that the 16 core Ryzen 3000, is the same price as today's 12 core Ryzen 3000 and with obvious flow on effects for the 8 and 6 core Ryzens.
 

ClockHound

Golden Member
Nov 27, 2007
1,108
214
106
It's a fake. But for those in the fake news biz, a fake Intel 10 core is just as good as real FUD
 

SPBHM

Diamond Member
Sep 12, 2012
5,056
409
126
not sure how much good a 10 core 9900K would make, it would hardly move things forward in terms of ST performance, and MT would be bellow AMD anyway,
also power usage would be even worse...
 

jpiniero

Lifer
Oct 1, 2010
14,510
5,159
136
not sure how much good a 10 core 9900K would make, it would hardly move things forward in terms of ST performance, and MT would be bellow AMD anyway,
also power usage would be even worse...

Marketing wise it lets them give HT to the lower models. Perhaps they could squeeze a little bit more out of 14 nm as well to fit 10 but no more. In any case it looks like 10 is the max they will do on the ring.
 

moinmoin

Diamond Member
Jun 1, 2017
4,934
7,619
136
Just a note for the future, "Forbes Contributors" like the page linked in the OP is essentially an open blogging platform that says nothing about the articles' trustworthiness (more the opposite really).
 

StinkyPinky

Diamond Member
Jul 6, 2002
6,761
777
126
No english speaking person ever puts the dollar sign in that position, that's an asian thing. So that immediately outs it as a fake.
 

Yeroon

Member
Mar 19, 2017
123
57
71
No english speaking person ever puts the dollar sign in that position, that's an asian thing. So that immediately outs it as a fake.

I do it all the time - in casual messaging situations. Mostly because I've already typed the number and dont feel like going back to insert the dollar sign.

I did like the 14+++ though. That got a chuckle out of me.
 
Nov 3, 2004
10,491
22
81
Once Intel finally manages to shrink its process, I expect them to come out with a fury and vengeance. Might take a few years though.
 

epsilon84

Golden Member
Aug 29, 2010
1,142
927
136
If true, there is merit to the idea as it allows Intel to be more competitive with the Ryzen 3000 lineup. Obviously it won't beat the 3950X but everything below that should look more competitive. Like others I'm dubious about how hot the 10C part will run though. Is 14nm fully tapped out or is there one last round of optimisation left in it?

I actually wrote on these forums a while back that enabling HT on all SKUs was Intel's last 'low hanging fruit' until they sort out their 10nm mess. It makes sense because there is almost no reason at all to get a 9th gen core CPU unless you're chasing high fps gaming.
 

cortexa99

Senior member
Jul 2, 2018
318
505
136
Searching with google and find tweaktown, wccf, theinquier, notebookcheck, and some japanese websites also report it.

I think, even if this slide was pointed out as a fake, these SKUs looks logical to be a disappointing reality. Intel has no choice that other than adding cores in coffeelake 14nm++++++++ to face against Zen2 or even Zen3 before the crappy 10nm yielding is good enough. It's better-than-nothing action from Intel.

OTOH LGA1159 Wtf??? If it needs a new socket that incompatible to LGA1150 then Intel might lost so much reputation from users.

here's old thread about CML
https://forums.anandtech.com/thread...-comet-lake-die-to-preempt-zen-2-am4.2557511/
 
Last edited:

jpiniero

Lifer
Oct 1, 2010
14,510
5,159
136
I think, even if this slide was pointed out as a fake, these SKUs looks logical to be a disappointing reality. Intel has no choice that other than adding cores in coffeelake 14nm++++++++ to face against Zen2 or even Zen3 before the crappy 10nm yielding is good enough. It's better-than-nothing action from Intel.

The other option is to port an updated core back to 14 nm, but they would have to had planned for that, and of course it would be really big and hot. We'll have to see if Rocket Lake does indeed do this.

They certainly aren't going to play the core count game with AMD, and they can't anyway.
 

DrMrLordX

Lifer
Apr 27, 2000
21,582
10,785
136
10C Comet Lake would still be less lame than the 9900KS, which is (sadly) not fake. Merry Christmas from Intel?