Using floating-point arithmetic on FPGAs to accelerate scientific N-Body simulations | IEEE Conference Publication | IEEE Xplore