libfann2 cannot train new neural network on OpenSuse Leap 15.3 returns floating point error (core dump)

[moscowboy1974]

Have gone through the many hints and tips and normalised my data prior to training the neural network. Despite using the provided example fanndemo.cpp but with my own dataset I get a floating point error(coredump) run time error. I can repeat this error on a Raspberry Pi4 running the latest 32 bit Raspbian and latest updates so the issue is not AMD x86_64 related. Any help would be greatly appreciated