#2119: The binary neutron star gallery example gives different results with the release candiate.

Comment (by Roland Haas):

That may a quick(ish since it takes a day to get to merger) check.

Changing ID however is not quite the same thing as running with reduced IEEE compliance: I would still run the same code as before which differs from the code that ran on Zwicky by the fact that the Intel compiler defaults to imprecise math see eg the ever popular: https://www.nccs.nasa.gov/images/FloatingPoint_consistency.pdf and in particular the prec-sqrt section (the default is imprecise) https://software.intel.com/en-us/fortran-compiler-developer-guide-and-reference-prec-sqrt-qprec-sqrt.