Do developers ever need to deal with this problem by writing multiple implementations of the same code, and the implementation that runs on a given machine is chosen at compile time based on the specs of that machine? Or is it just assumed that the "same" software package shouldn't be run on machines of vastly different sizes?
Do developers ever need to deal with this problem by writing multiple implementations of the same code, and the implementation that runs on a given machine is chosen at compile time based on the specs of that machine? Or is it just assumed that the "same" software package shouldn't be run on machines of vastly different sizes?