sdb@whuxlm.UUCP (Brener Stanley) (02/22/85)
Are there any computers currently working on tri-state logic (as opposed to binary logic)? Are there any research projects trying to develop such machines? Is there a fundamental paper dealing with the topic from a mathematical perspective?
spector@acf4.UUCP (David HM Spector) (02/23/85)
I seems to remember one of my college profs (Dr. Robt. BK Dewar..) telling us about a 10 state machine that was developed in Great Brit. a while ago, I think it was a tube machine....but I don't know...
lwe3207@acf4.UUCP (Lars Warren Ericson) (02/23/85)
[] There is such a fundamental paper. It is called "The mathematical theory of communication" by Claude Shannon. The conclusion is that any channel which can transmit, say, 4-state signals at K megahertz with an acceptably minimum level of noise, can transmit 2-state signals at 2K megahertz. Translation: building a computer with N-state logic, N > 2, doesn't buy you anything, because you can use the same technology to build a computer with 2-state logic which will solve problems just as fast. -- Lars Ericson -- UUCP: cmcl2!csd1!ericson -- ARPA: ericson@nyu
abc@brl-tgr.ARPA (Brint Cooper ) (02/25/85)
In article <290002@acf4.UUCP> lwe3207@acf4.UUCP (Lars Warren Ericson) writes: >[] > > Translation: building a computer with >N-state logic, N > 2, doesn't buy you anything, because you can use >the same technology to build a computer with 2-state logic which will >solve problems just as fast. > I don't believe that Shannon addressed the problem of squeezing more functions on a given area of silicon, however.
doug@terak.UUCP (Doug Pardee) (02/26/85)
> Are there any computers currently working on tri-state logic (as > opposed to binary logic)? A decade or so ago, I read that some Russian computer systems used trinary. A few years back I worked with a fellow who claimed to have (in a previous job) worked on a prototype system with base -2. -- Doug Pardee -- Terak Corp. -- !{hao,ihnp4,decvax}!noao!terak!doug
johnl@ima.UUCP (02/27/85)
I have heard claims that the 8087 uses multi-state logic internally. The
same sources have claimed that it didn't work as well as Intel had hoped,
which is why you don't see faster 8087 and 80287 chips.
Informed commentary by mail invited.
John Levine, Javelin Software, Cambridge MA 617-494-1400
{ decvax!cca | yale | bbncca | allegra | cbosgd | ihnp4 }!ima!johnl
Levine@YALE.ARPAdavid@daisy.UUCP (David Schachter) (02/27/85)
Some of the megabit dynamic RAMs being presented at the ISSCC (The IEEE Solid State Circuits Conference) use multi-level logic to reduce the number of cells. (This is according to the ISSCC preview in a recent Electronics Week magazine. The magazine is published by McGraw Hill and is available in many libraries.)