gyrfalcON does not compile on tiger lake CPU's

[teuben]

This is a case where the compiler does not know about the CPU yet. gyrfalcON uses -march=native to get the best possible speedup, but this fails on current intel Gen11 CPUs with the tiger lake based CPU.
The solution is to put
export NO_ARCH_NATIVE=1
in your ~/.nemo_local.sh file on that particular machine.

Not until falcON is using NEMO's makedefs mechanism, can we expect this kind of problem to go away.

[teuben]

the configure --disable-native opption will also do it, but the ~/.nemo_local.sh is useful for a global setting on a machine.

[teuben]

usr/dehnen/falcON/src/public/lib/bodyfunc.cc has some hardcoded options (e.g. clang will not like the -mpreferred-stack-boundary=4 option)