Accelerating ODE solving with FPGA

I am trying to accelerate ode solving with FPGA. As the first 
experiment, I used the Lokta Volterra model to solve with the Euler 
method. I did write the VHDL code for 1-time step. Now I want to write 
the code to run for several time steps and should be run in a parallel 
manner on Stratix IV FPGA. I found this attached paper but couldn't 
understand the method. Can anyone give me some help?
Thanks