AMD Transtioning To 12nm LP Process For Vega, Ryzen In 2018

J

juanrga

2[H]4U
Joined
Feb 22, 2017
Messages
2,804
14LPP+ has been replaced by renamed to 12LP

http://www.tomshardware.co.uk/amd-ryzen-vega-12nm-lp-2018,news-56797.html

http://wccftech.com/amd-announces-2nd-gen-ryzen-vega-launching-12nm-2018/

EDIT:

Toms mentions 14nm, but Glofo is really using TSMC 16nm as baseline

Global-Foundaries-12LP.jpg


Note that a true 16nm to 12nm node shrink would provide 78% better density, but this 12LP node only provides "up to 15%" better density because this "12" is a marketing node label.

"Up to 15% improvement" over TSMC 16nm means this 12LP has basically the same density than the current 14LPP node.

TSMC 16FF: 0.07µm² HD SRAM

Glofo 12LP: ~0.061μm² HD SRAM (using the 15% official claim)

Glofo/Samsung 14LPP: 0.064μm² HD SRAM

"Risk production in 1H 2018". Aka volume production in 2H 2018.

"More than 10% improvement in performance" over TSMC 16nm. 16FF has higher clocks than 14LPP (ask Nvidia). So I expect 5--10% higher clocks over 14LPP.

If above numbers weren't enough to demonstrate that this 12LP is just a relabeling of the former 14nm+ node, the official slide says "12LP build on the GF 14LPP". almost confirming officially this 12LP is just a rename for the old 14nm+ node

aHR0cDovL21lZGlhLmJlc3RvZm1pY3JvLmNvbS9HL0cvNjc3MTA0L29yaWdpbmFsLzE1LkpQRw==



So the AMD roadmap continues being

Zen on 14nm on 2017.
Zen on 14nm+ on second half of 2018 (with 14nm+ renamed now to 12nm).
Zen2 on 7nm somewhat in 2019.


At the Global Foundries Technology Conference, AMD’s CTO Mark Papermaster announced that the company will be transitioning “graphics and client products” from the Global Foundries14nm LP FinFET process it uses today to the new 12nm LP process in 2018.

We followed up with Papermaster in person and confirmed directly that the company will transition both Vega GPUs and the Ryzen line of processors to the 12nm LP process.
Click to expand...
 
Last edited:
  • Like
Reactions: noko
like this
I wish some tech site would gander a closer look at all the PR-nm being tossed around.
16NM FF is a false label..techincal it should have been named 20 NM FF.
The same false claims are being used abot 14nm, 12nm, 10 nm and 7 nm...

Only nm claims I currently trust are those from Intel...the rest....pure PR FUD...and it will keep being like this as long as they can away with it without anyone rasing the finger and sayin "Erhhm...no!"
 
Yep, its just 14LP+ renamed to 12LP, nothing else. They dont even want to compare with their own 14LP, just 16FF from TSMC.
 
Honestly the way fabs are naming their processes is getting rediculous..

"This is the new 1 atom ULFFPQQAT process that's actually just a cleaner version of 20nm..."
 
TSMC 12FFN is a 4th generation 20nm FF process. Glofo´s 12LP is a second generation 16nm FF give or take. Just to show how silly it has become.

12LP is actually a third generation.
 
Last edited:
Nathan_P said:
basically its all marketing bollocks
Click to expand...

The problem is....that retarded PR-FUD draws headlines, but not critique...it's like claiming a PIII CPU could do +1GHz and no one cares about that it's not +1GHz...but it's +1PR-FUD-GHz...

I am slighty dissapointed that one gets called out...the other gets reported as "facts"....
 
A Richland-like refresh of RyZen was planned many many time ago. Some of us estimate we will get 100--300MHz higher clocks on Pinnacle Ridge compared to Summit Ridge.
 
You must log in or register to reply here.
Back
Top