CONTR47@NOSC-TECR.ARPA (06/04/88)
I too look forward to Ada for Transputers. If you need raw compute power (as in realtime aircrew training simulators) they seem to be the most cost-effective. Imagine 150MFlops in your desktop PC for about $20K. The trick is, of course, to be able to port Ada code to them somehow. That's the type of research we are doing for the Navy. Here is what I heard from David Bye (214)490 9522, the Inmos salesman in Dallas, TX on 4/28/88: Alsys (England) is planning to have Ada for the transputer sometime in mid 1989. He placed my name on a list of people to be contacted when further information is available. He is a most pleasant fellow and I guess would place your name on the same list. Since he is from England and sounds like John Barnes I am tend to believe anything he tells me about what is happening in England (humor). Like they say "Its a good trick if you can do it". If Ada tasks can be mapped to transputer chips it should give Ada tasks great utility in realtime systems. Imagine even primitive static Ada task allocation where I can map 100 Ada tasks to the 100 transputers in my 150MFlop PC. With such a box I should be able to perform simultaneous realtime simulation for a formation of 10 F16's. This would be about a factor of 100 cost reduction in hardware. Of course the other trick is to do this without increasing software cost (or maybe even decreasing it). Hope this helps. regards, sam harbaugh ---------------------