Investigation of floating-point round-off errors within time-domain electromagnetic field computations | IEEE Conference Publication | IEEE Xplore