while we are still years away. i cant help but be giddy at the thought of a 1 thz cpu and what graphene is going to do to electronics such as Oculus VR. 8 to 12k resolution, with graphics better than Avatar, and a world bigger than GTA V..
Think 1 thz cpu (20+ cores) solid state drives. (your cellphone will be far more powerful than the best personal computer we have now...) now hook that up to VR and your not locked into a seat 😛
the future is exciting and i cant wait to play in it!
IBM managed to get 350 ghz silicon germanium transistors in 2002 so all I am saying is that it doesn't immediately scream 1THz cpu processors just cuz they got some logic gates working at high speed. There are other things that are important like heat dissipation etc.
"lossofmercy" wrote: IBM managed to get 350 ghz silicon germanium transistors in 2002 so all I am saying is that it doesn't immediately scream 1THz cpu processors just cuz they got some logic gates working at high speed. There are other things that are important like heat dissipation etc.
You are kind of right. On the other hand. Who expected that braekthrough from Samsung? Last time i have read they managed to grow a few mm graphene onto copper. Now they a few thousand folded that amount. I think progress can also proceed in a non-linear, non-exponential way. Take for example cambrian genomics. They invented a machine with which you can now synthsize a human genome for 1000$ instead of 1 billion. (revolution!) Just let Samsung throw their billions onto R&D. But anway we wont see a bigger jump than Moores law would allow. Money thing..... But can be assured that computers will still get faster for a few decades. Samsung makes these announcement mainly because of stock market.
"lossofmercy" wrote: IBM managed to get 350 ghz silicon germanium transistors in 2002 so all I am saying is that it doesn't immediately scream 1THz cpu processors just cuz they got some logic gates working at high speed. There are other things that are important like heat dissipation etc.
Here's a promising approach towards solving energy dissipation and current leakage:
"cerebral" wrote: You are kind of right. On the other hand. Who expected that braekthrough from Samsung?
The breakthrough from Samsung is absolutely fantastic, no doubt. I am 100% excited about that development.
I am just cautious about the whole 400 GHz graphene xor gate thing. Having smaller and smaller chips is definitely a good thing, we just don't know how good these are compared to regular silicon chips in a fair comparison. I am guessing it's going to be faster, but not 100x much faster.
"cerebral" wrote: You are kind of right. On the other hand. Who expected that braekthrough from Samsung?
The breakthrough from Samsung is absolutely fantastic, no doubt. I am 100% excited about that development.
I am just cautious about the whole 400 GHz graphene xor gate thing. Having smaller and smaller chips is definitely a good thing, we just don't know how good these are compared to regular silicon chips in a fair comparison. I am guessing it's going to be faster, but not 100x much faster.
I wouldn't count on big jumps either. It will probably just allow the normal moore's law-like progression to continue in the next decade or two.
When cut into sub 10nm ribbons or quantum dots graphene does have a band gap useful for digital logic. Such nanoribbons can also make memristors for high density memory, so cache becomes RAM and RAM becomes storage. EUV lithography has stalled for some time now so perhaps instead contact lithography is the way forward, maybe a stamp made out platinum http://www.geek.com/science/scientists-accidentally-turn-graphene-into-diamond-1590210/ Pure diamond is a good electrical insulator. As the stamp will hopefully be reusable you can cut it with precise but lengthy electron beam lithography.
As well as being faster than silicon graphene also creates less heat and power draw so you could mount your graphene APU on the rift for tether free low latency 8k direct video link. As it creates less heat and diamond and graphene are good thermal conductors you can also stack them. If you start from a normal large chip size of 32x32mm if its 1 micron thick cut such chip in half and stack the halves and repeat 10 times and you will have a 1mm cube. Signals also travel pretty close to the speed of light in graphene wires.
