Open mayank0403 opened 5 years ago
I can confirm that there is an issue with the IEEE versions of the FP circuits.
This is due to some extra input and output bits that are included in the circuits, that are used for signaling exceptions in the FP calculation (e.g. overflows, divisions by zero, etc.). These are currently not handled correctly. Unless you really need these extra bits, please use the no_status FP gates for now.
In the floating point example, present in
examples/float/abyfloat.cpp
, changing the last parameter of PutFPGate from no_status to ieee gives either SIGFPE or SIGSEGV.The situation is like this:
PutNonLinearVectorGate()
.PutPrimitiveGate()
.With no_status, the example runs fine without any issues.