http://www.foresight.org/nanodot/?p=3739 examining transformative technology Wed, 03 Apr 2013 18:23:47 +0000 hourly 1 http://wordpress.org/?v=3.0.4 http://www.foresight.org/nanodot/?p=3739#comment-866542 Eric Williams Tue, 09 Feb 2010 03:13:07 +0000 http://www.foresight.org/nanodot/?p=3739#comment-866542 Very exciting stuff, the carbon transition certainly looks more feasible in the short term now...thanks for the great link roundup. It seems like the cost of mass-producing graphene (with enough purity) is the next major hurdle, since the scaling down in size seems inevitable. One of those links quotes 1thz switching speeds as possible after a reduction in size to < 35nm. This may sound like a lot, but current MOSFETS switch upwards of 35ghz, and the limiting factor is interconnect speed (and heat), keeping cpu frequencies well under 5ghz. I wonder if we'll need <a href="http://web.mit.edu/newsoffice/2009/optical-computing.html" rel="nofollow">optical interconnects</a> to take advantage of this increased switching speed, and if this and other high-bandwidth interconnect research will be more of a priority for semi-conductor companies, given the clearer path to carbon and the increased transistor frequencies that come with it. Also wondering about efficiency. Maybe this decrease in PDP and apparent lower transistor leakage will help stave off the necessity of reversible computing to continue Moore's? J. Storrs Hall, do you have an opinion on this, and perhaps an estimate as to how much more efficient these graphene transistors might prove? Very exciting stuff, the carbon transition certainly looks more feasible in the short term now…thanks for the great link roundup. It seems like the cost of mass-producing graphene (with enough purity) is the next major hurdle, since the scaling down in size seems inevitable. One of those links quotes 1thz switching speeds as possible after a reduction in size to < 35nm. This may sound like a lot, but current MOSFETS switch upwards of 35ghz, and the limiting factor is interconnect speed (and heat), keeping cpu frequencies well under 5ghz.
I wonder if we’ll need optical interconnects to take advantage of this increased switching speed, and if this and other high-bandwidth interconnect research will be more of a priority for semi-conductor companies, given the clearer path to carbon and the increased transistor frequencies that come with it.
Also wondering about efficiency. Maybe this decrease in PDP and apparent lower transistor leakage will help stave off the necessity of reversible computing to continue Moore’s? J. Storrs Hall, do you have an opinion on this, and perhaps an estimate as to how much more efficient these graphene transistors might prove?

]]>