fix int overflow while compiling CLASS D,E

[yuxuan-z19]

Large problem sizes in CLASS D/E can cause integer overflow in memory indexing and array size computations, leading to compiler warnings and potential runtime errors.

This PR fixes #17 by updating setparams.cpp to use 64-bit integer literals (LL/ULL) for all relevant integer constants. This ensures that:

• Integer arithmetic in array size and memory allocation calculations does not overflow.
• Compiler warnings about integer overflow and sign change are resolved.
• Calculations are safe for large problem sizes on modern GPUs.