How long intel is going to stay on 14nm? Is there anything to improve anymore if we don't count yearly 1-2% better results?
How long intel is going to stay on 14nm...
Anthony Gray
Other urls found in this thread:
venturebeat.com
twitter.com
Julian Perry
Just buy an i9 9900k goy
Nicholas Hall
Node shrinks are anti-Semitic. 14nm++++++++ is God's chosen process.
Easton Nguyen
Intel 10nm is 2020 ramp at best estimate.
Daniel Allen
>tsmc will make 5nm in 2020
>intel 10nm in 2020
It's over intelbros
Easton Reyes
14nm+++++++++++ will be in production until 2024 or something like that until they stop providing support for the last 14nm Xeon line.
I imagine they might start fabbing their own chipsets on it though, so its probably staying around long after that.
Gabriel Foster
Why couldn't it be i9 9999k?
Jackson Foster
At this rate, Intel will likely file for bankruptcy before their 10nm goes into general mass production.
Gabriel Price
this
Dominic Ross
friendly reminder to see
Connor Taylor
>not an i9 11k
Levi Nelson
because that's going to be the 28c 500W thing
Wyatt Kelly
euv doesn't work and never will.
It's the end of the road.
From now on we'll only have affordable 14nm or expensive 10nm.
Matthew Smith
Incel will just buy in the tech for smaller transistors.
Brayden Ward
tsmc and samsung are the two last bleeding edge foundries. seems pretty sad actually.
i expect intel will outsource some low end cpus to tsmc in the near future.
Christopher Powell
All their current gen chipsets are already on 14nm. Apart from x370 which was rushed out early with coffeelake1 in a panic response to ryzen. x390 was destined to be what a 14nm x370 should have been but that idea was cut and it's also an older process (22nm afaik).
In unrelated news intel supposedly has a supply shortage for xeons, all that fab space taken up by mainstream chipsets has nothing to do with it though.
Nolan Bell
Problem is nobody managed to get the new euv machines to work.
TSMC is using the old immersion lithography machines, but that means cutting even more capacity away from 14nm because on those machines it takes longer to do 10nm than to do 14nm.
They could buy more immersion lithography machines, but then they'll probably have too many once they do get euv to work.
Jack Cook
I have a "friend" working at GloFo, and apparently the machines work fine. The problem they have, the biggest one anyway, which is leading to low yields is intensity of the light source. Single digit yields with EUV is where they're at right now. I think thats why Samsung has been hold off on their own 7nm process, waiting for a full EUV debug on their own lines.
GloFo just didn't want to spend the money that Samsung is willing to spend.
William Lee
Wyatt Jenkins
>Single digit yields with EUV
That's what I meant by "not working".
From what I understand they also require a lot of maintenance, making them very expensive to run.
Shit is just a fucking nightmare and they should have gone for a synchrotron for the light source instead of plasma.
Charles Nguyen
>plans
Intel also "planned" to be at 3nm by now.
Blake Moore
>it's going according to plan
Connor Harris
You seem to be confusing tsmc with intel