Those who recall Dr Gene Amdahl’s failure to develop all the technologies he needed at his Trilogy Ltd to develop an IBMulator that would be sufficiently more powerful than the 3090 to create its own market will be well aware that even major development projects run by the very best people can go disastrously wrong, and John Rollwagen’s explanation of why Cray Research Inc cancelled Steve Chen’s 64-processor MP development project is revealing: the project, he says, would have to complete successfully some six different lines of research, each with only a 50% probability of success; nevetheless, IBM is sufficient a believer to be bankrolling Chen’s new Supercomputer Systems company in Eau Claire…
