| haxxiy said: Samsung claimed a SRAM size of 0.030 um², CCP of 44 nm and MMP of 36 nm on their proposed roadmap. Intel's 10 nm has a SRAM size of 0.031 um², CCP of 54 nm and MMP of 36 nm. It's unknown if they will even come close to these since we know their newer, proposed 5 nm demoed alongside IBM a few weeks back has a CCP of 48 nm (coupled possibly with a rather agressive gate scaling). Obviously no upcoming technology and the ones sponsoring it are going to make a presentation claiming it's going to be worse. EUV has a couple of theoretical advantages, but it's extremely complex and expensive to get it right - so, time and time again, for years now, another lithographic technology was preferred over implementing EUV. Both TSMC's 10 nm and GloFo's 7 nm are denser than Samsung's 10 nm. It's just a question of whether GloFo can get it to the market before Samsung's 8-7 nm manage to do on H2 2019 at earliest. GloFo had 32 nm and 28 nm ready before Samsung, anyway, not to mention Samsung is following a different path this time around, improving their 10 nm design, so I'm not sure why you are so offended by the possibility GloFo manages to have a denser node for a year or so. Besides, do try to do with less teenage internet slang and passive-agressiveness and more arguments the next time around, otherwise I'm going to assume you don't know much about what you are talking about, and not even bother to discuss. |
What other lithographic technology ? The only other lithography technology we used was 193nm ArF immersion lithography which we used for over a decade. Scanner technology has lagged for years until EUV came around ...
Kind of funny that you're pitting GloFo with Samsung at the time when the latter only entered high end chip manufacturing in a little over half a decade ago whereas the former started as a spinoff from AMD's foundry unit with chartered semiconductor. Samsung had 32nm ready with the Exynos 4 at the same time as GloFo did with Bulldozer despite only having one year of experience in chip manufacturing. What you said about who offered 28nm first wasn't true either as just after 3 years of experience, Samsung went on to humiliate GloFo by offering 28nm first with the Exynos 5 6 months before GloFo was ready with Steamroller ...
GloFo had to cancel 20nm which resulted in AMD cancelling project SkyBridge and then GloFo went on to cancel their own 14nm technology so that they could license 14nm from Samsung and they still failed hard with that one since they couldn't meet orders for the Apple A9 ...
In that timeframe just after licensing Samsung's technology, GloFo was faced with taking in another failing chip manufacturing unit from IBM by being paid $1.5B and before that GloFo had joined the IBM reseach alliance so what does all this have to say about GloFo as a supplier for their customers ?
I've never once seen GloFo having an advantage in terms of density or time to market and LOL for you implicating that I'm offended about GloFo beating Samsung to the punch when I'm just laying down facts and the realistic scenario considering GloFo is created from a bunch of rejects ...
It's hilarious that you assume that I don't know much because of temperament but that doesn't get in the way from addressing your points ...
