bhv@igloo.scum.com (Bronis Vidugiris) (08/21/90)
We just had a talk on the 'connection machine' and Spice, which I found rather disapointing. Though there were some interesting algorithms discussed for process simulation, the status on circuit simulation (to replace Spice) was that only 'relaxation methods' had proven suitable, and these only converged for non-analog circuits. TI has apparently reported good luck with the 'CSG' algorithm, but the folks giving the talk said they didn't have any experience with it themselves, but that they were in the process of investigating the TI algorithm. The disapointing part was that it took about 3 hours to say this. Logic simulation was also in the 'there must be a better way' phase. I'm not sure about the status of design-rule checking. It's a great machine for global weather simulation and oil prospecting, and fluid dynamics simulation and nuclear weapons simulation, though :-)