Deciding floating-point logic with systematic abstraction | IEEE Conference Publication | IEEE Xplore