Well, so much for Silicon Valley . . .

Post #436,477 by Andrew Grygus 9/29/20 4:30:57 PM 9/29/20 4:30:57 PM Reply	Well, so much for Silicon Valley . . . . . if these guys succeed. Scientists are working on making computer chips from pure carbon, graphene, coaxing it to act as semiconductors, wires and insulators. They say all-carbon computers would be about 1000 times as fast as silicon, and use less power. I suspect Google will want to play with these as soon as they are available. Machine Design
Post #436,480 by drook 9/29/20 5:29:05 PM 9/29/20 5:29:05 PM Reply	Or will they be smaller? Sounds like they'll be 1,000 times as fast for the same design, but wouldn't they likely just make them smaller instead? -- Drew
Post #436,481 by InThane 9/29/20 6:01:04 PM 9/29/20 6:01:04 PM Reply	That's probably how they get the speedup. At those sizes, the speed of light is a very important consideration.
Post #436,488 by drook 9/30/20 12:02:34 AM 9/30/20 12:02:34 AM Reply	But what if they don't speed up at all? https://www.anandtech.com/show/11550/the-intel-skylakex-review-core-i9-7900x-i7-7820x-and-i7-7800x-tested/6 That shows the latest Intel Skylake 28-core design has a 698 mm^2 die size. If you reduce the complexity of that design by a factor of 1,000 you have a chip smaller than 1 mm^2. I know performance doesn't scale linearly with die size, but as a first approximation you could fit the performance of a high-end server in a microSD card. -- Drew
Post #436,482 by Another Scott 9/29/20 6:12:05 PM 9/29/20 6:12:05 PM Reply	Original press release https://news.berkeley.edu/2020/09/24/metal-wires-of-carbon-complete-toolbox-for-carbon-based-computers/ Intel, AMD, TSMC, etc., don't have anything to worry about yet. It's a step in the process of trying to figure out what graphene and related materials are practically good for. (Graphene is fast, but has no "energy gap" so you can't turn off transistors made from it, so the power dissipation from a real circuit of hundreds of millions or billions of them would be huge. You can do tricks to make graphene have a bandgap (like silicon and traditional semiconductors), but it's usually not robust because graphene is practically inert (which is why it exists in the first place).) The dichagenogides may be more practical, but there's still a lot of work to be done. https://pubs.rsc.org/en/content/articlehtml/2013/tc/c3tc00710c?casa_token=WLPTGGk32KUAAAAA:pYyxz2zaRRRSLdlpKfTWA9FlWatYJbvUOirwPEi3kOwivIvQHg0Yxn1cgGejwwE5H1uIKem2i7H_yRI Cheers, Scott.

Welcome to IWETHEY!