Today in order to compare runtimes, I need to build against a single runtime (the lowest common denominator) with -f and then I can execute against multiple runtimes with --runtimes. That works in the majority use case, however sometimes you want to answer the question "how does this code compare on different runtimes?", and that includes compile-time impact due to new overloads being introduced, which necessitates building against the appropriate reference assemblies for each version.
Today in order to compare runtimes, I need to build against a single runtime (the lowest common denominator) with -f and then I can execute against multiple runtimes with --runtimes. That works in the majority use case, however sometimes you want to answer the question "how does this code compare on different runtimes?", and that includes compile-time impact due to new overloads being introduced, which necessitates building against the appropriate reference assemblies for each version.