the info i was getting was 10nm
 
which would explain the large dies, power draw and the heat issues
 
if they can beat AMD with 10nm++ rather than 7nm(thats non + til next year)
thats a good deal and put hopper on 7nm+ when it matures
 
 
corrections that 10nm refined to 8nm LPP without EUV
 
lol if they beat AMD with 8nm LPP then they are ahead of the ballgame
 
 
confirmed Orin uses 8nm LPP
 

Xavier Zepherious
if they beat AMD with 8nm LPP then they are ahead of the ballgame

"If"? It's almost certain that they will 'beat' AMD when it comes to performance. I don't think that there is a single person who expects otherwise.

AMD fights the performance-per-dollar battle, and the winner of that battle is far more subjective and opinion-based. So will AMD win when it comes to price-to-performance? It's hard to say. Even long after the products are released, that argument won't be settled. Everyone will have their own opinion and use-case. It won't matter to AMD; AMD will just keep on happily selling cards to the mainstream.

Wrong.

When you have 5,000+ symmetrical multiprocessors (CUDA cores) on one big massive chip... Between 8nm and 7nm - even 10nm - makes zero real world difference.

If it did - then AMD would be ruling the GPU market... and, of course... not even close.

On full-fat cpu’s or very small footprint phones, it makes a difference.

MasterMiner
When you have 5,000+ symmetrical multiprocessors (CUDA cores) on one big massive chip... Between 8nm and 7nm - even 10nm - makes zero real world difference.

If it did - then AMD would be ruling the GPU market... and, of course... not even close.

On full-fat cpu’s or very small footprint phones, it makes a difference.

Smaller GPU and CPU dies do make a difference in energy efficiency. This is why Intel's current CPUs are not as energy efficient as AMD's CPUs in my personal opinion. 
 
Smaller die sizes also allow for increased cores and speed. 

Miguell
they are reaching silicon limits.
under 10nm its obvious by now its becoming a real pain! we are stuck!
 
 
i wonder when they will make the transition to graphene!

Right now that is not a consideration. TSMC for example already has planning for the next few years all the way done to 2nm.

Miguell
welll they realllllly gonna struggle then....  really struggle....   if they already have all these issues getting to 7nm much less 2nm .....

Who? TSMC? TSMC hasn't been struggling with 7nm. TSMC is already making 5 nm with a 80% yield rate.
And TSMC has already started constructing its 3nm fab late 2019 at the cost of 19.5 billion US dollars. TSMC plans to be producing 3nm chips in 2023.

Intel has been struggling big time with less than 14nm. Not TSMC.

The problem with graphene is that no one (as far as I know) has been able to figure out how to make graphene have a band gap (act as a semiconductor). As far as I know, all the graphene technology currently in development uses a traditional silicon semiconductor, and just uses the graphene for all of the connecting interconnects. Graphene has excellent electrical conductivity and thermal conductivity, so it can still be a very useful thing, but its not going to solve the limitations of silicon on its own. Until they find some other semiconductor or figure out how to to make graphene act as a semiconductor, graphene isn't going to be the miracle material. The band gap problem means that they have no way to turn it off. If it's always on, it's not a switch, is it? And they still need to develop the graphene manufacturing process in order to make it into something which can be used in mass production at a reasonable cost.

Edit: Did some more research and discovered that one research group claims to have made graphene behave like a semiconductor by making it grow with uniform spaced pores. I don't know if further progress has been made. It will be interesting. Either way, we aren't there yet. And according to most slides available, graphene commercialization projects are currently using a silicon base layer.

I actually think it’s a problem at the sub-atomic level, and not materials, that is limiting sub nanometer growth.

Well beyond my brain-power - but quantum entanglement appears to be a problem sub-nanometer.

weasel333
Wrong.

You win the internet for the most 2020 thing I've read all day. Posts a 1 word opinion and doesn't bother to explain why. Thanks for my quota of entropy for the day. 

Xavier Zepherious
NVIDIA (NASDAQ: NVDA) Surpasses Intel to Become the Third-Largest Semiconductor in the World by Market Cap

:O

Wow, that's incredible. 

atfrico

Sajin

Xavier Zepherious
NVIDIA (NASDAQ: NVDA) Surpasses Intel to Become the Third-Largest Semiconductor in the World by Market Cap

:O

Monopoly anyone? If they went after Microsoft doing the same you dont think Nvidia will be next? Fines to come you will see😼

Do you know what market cap is? do you understand that AMD are in all home consoles where 99% of games are developed?

NVIDIA Ends Production of GeForce RTX 20 ‘Turing’
 
https://wccftech.com/nvidia-geforce-rtx-20-turing-production-end-geforce-rtx-30-ampere-gaming-graphics-card-launch-close/

Xavier Zepherious
NVIDIA Ends Production of GeForce RTX 20 ‘Turing’
 
https://wccftech.com/nvidia-geforce-rtx-20-turing-production-end-geforce-rtx-30-ampere-gaming-graphics-card-launch-close/

:O

leaked info 
rtx 30 series
aug 17-31 - announcement
 somewhere in there
 
released date sept 17
 
 
rtx 20 was aug 20 and sept 20
 
so exactly 2yrs later
 
 
 
and expect card pricing to be high because there will be high demand
 
seems Nvidia advised AIB partners that the mining crazing is winding up again
 
 

 
Can't help myself  even if it might be fake  everyone needs there fix of techporn
 
rtx 3080