I have a multiple summation function with variables of the order of 10^-30 that quickly runs out of memory when I use SUM. INTEGRATE gives an acceptable approximation for variables of that Order Of Magnitude but I need to be more exact. Do I hire someone with access to a mainframe? If so, how? Is there a MATHEMATICA trick I’ve overlooked? (Logarithms aren’t an option for my expression.)
Thanks Frank. I’ve tried that and scaling does point to values I’m expecting my expressions to assume, but unfortunately it doesn’t give me the precision I need to make the work acceptable. Exact calculation will require a heavy dose of RAM and since I see no other alternative I’m now in the process of trying to rent some time on a super computer up in Orono.. I welcome any other suggestions.