Ravi Pandya
+1 425 417 4180

syndicate this site

Ravi Pandya   software | nanotechnology | economics


2007 11 10

2004 10 09 08 07 06

2003 04 02 01

2002 12 11 10 09 08

2001 11


Ravi Pandya
Cloud Computing Futures
ravip at microsoft.com

00-02 Covalent
97-00 EverythingOffice
96-97 Jango
93-96 NetManage
89-93 Xanadu
88-89 Hypercube
84,85 Xerox PARC
83-89 University of Toronto, Math
86-87 George Brown College, Dance
95-Foresight Institute
97-Institute for Molecular Manufacturing


The opinions expressed here are purely my own, and do not reflect the policy of my employer.

Thu 29 Aug 2002

IEEE Nanotechnology Conference 2002

About 200 attendees, ~90% academic, ~75% from the US.

Lots of papers on semiconducting carbon nanotubes, single electron transistors, quantum dots, quantum cellular automata, qubits, etc. There's a lot of work going on, but it's still at the level of trying to painstakingly design, build, and test a single rectifier or transistor. The big progress is simply that there are many different groups able to get this far.

David Vallett of IBM Microelectronics on the state of the art in failure analysis for CMOS - combinatorial simulation & test, scan chain readout, deprocessing the packaging, physically analyzing the cone of logic behind the fault, etc. Nanoelectronics will need all this eventually, but not for quite a while - the complexity and scope of the current technology base is pretty impressive.

Panel discussion on "Grand Challenges in Nanoelectronics"

  • Robert Trew, Virginia Tech: The 2001 International Technology Roadmap for Semiconductors shows CMOS reaching the 90nm technology node in 2004, and the 65nm technology node in 2007, and the industry is in fact doing better than that - it's closer to a two-year than a 3-year cycle.
  • Sandip Tiwari, Cornell: Any successor to CMOS needs to fit into the CMOS process path.
  • Guiseppe Iannacone, Univ of Pisa: Look at systems & manufacturing issues - QCA requires 1:10,000 geometric tolerance, very tight timing.
  • Bob Weir, SRC: I've seen a lot of technologies come and go over the last few decades - GaAs, Josephson junctions, etc. It's hard to predict what will succeed.
  • Consensus: CMOS will continue for 10-15 years, then something will need to be done, but nobody knows what.
  • My question: Exponential technology curves are great, but what are the economic drivers on the demand side? It can't be MS Word on THz desktops.
    • Robert Trew: Attitude of "we'll build it and they will come" has worked well so far, but a significant factor in the current slump is that customers aren't convinced of the value of new technology.
    • Bob Weir: If I knew, I'd be working on it.
    • Jim Murday, NNI: Portability is key - people are here from all over the world, but the conference is in English - what about portable realtime translation?
    • Sandip Tiwari: Mobile medical analysis - don't go into your doctor for a checkup, wear a device that analyzes and transmits information on a regular basis.
    • [In other words, noone knows. Personally, I think desktop CPUs are at a disruption point where cost, logistics & marketing are driving the market instead of peak performance. The main demand for CPU cycles is shifting to networking, media, and simulation, which have very different economic structures - compare Xilinx, Nvidia, and Globus to Intel.]

11:27 #

© 2002-2004 Ravi Pandya | All Rights Reserved