I have code running on two separate 64-bit Linux machines. One machine is running IDL v8.2 while the other is running v8.2.2 (these machines are owned by different entities, thus the different versions and licenses). When I run my code on the v8.2.2 machine, however, I get "floating illegal operand" errors that I do not get on the v8.2 machine. Tracking this down reveals that the errors are created when using the "EXP" function with real numbers (as opposed to NaN values). For example, "EXP(1.3994603D0)" creates a math error on the newer version and not the older one while "EXP(1.3994603)" does not create a math error on either. This can happen with many other seemingly innocuous numbers. I am sort of at a loss as to whether or not this is related to the different versions or different systems. I was wondering if this was something that had been noticed before and perhaps fixed in a newer version of IDL?