I’m going to give the technical management the benefit of the doubt here, and assume the $15M Cray wasn’t primarily for chip design, but rather architectural si

I’m going to give the technical management the benefit of the doubt here, and assume the $15M Cray wasn’t primarily for chip design, but rather architectural simulation. This was intended to be a parallel system across several layers, while wrapping in a number of novel architectural ideas. It would not have been ridiculous to assume having a massively parallel supercomputer to experiment with those ideas more quickly would be worth the expense.