FERN: Fast Explicit Reaction Networks SVNCode

Fixed scanf buffer overruns due to smaller types than int

The Official Test Case checks out to acceptable precision for the abundant species. Small caveat documented at src/kernels.cu:550

Fixed Rate param computation on GPU, moved rate from Network to Globals

Fixed underallocation in Network; Kernel launches and times out with given IntegrationData params

Added print debugging for FplusMax and FminusMax

Appears to read alpha network files correctly, but does not continue to launch the kernel. Added back data directory

Restructured version compiles but is untested

Added launch example to main.cu

Completed ~90% of restructuring to fern OOP model

Test confirmed to ~4 decimal places for 150-isotope network with T9=7, massTol=1e-7, Precision=0.01, t_final=1e-9, rho=1e8