We had a "false start" with fixed-point math: I added the relevant header but did not properly call its API the way I thought I had. Addressing this issue, it seems that fixed-point state vector simulation might be difficult and not necessarily better than floating-point. However, QBDD is an ideal candidate for fixed-point math, since branch pairs always exist in sets of two whose overall sum norm is 1, implying an absolute scale of 0.5 (or sqrt(0.5)) for all QBDD math.
We had a "false start" with fixed-point math: I added the relevant header but did not properly call its API the way I thought I had. Addressing this issue, it seems that fixed-point state vector simulation might be difficult and not necessarily better than floating-point. However, QBDD is an ideal candidate for fixed-point math, since branch pairs always exist in sets of two whose overall sum norm is 1, implying an absolute scale of 0.5 (or sqrt(0.5)) for all QBDD math.