Higher radix floating-point representations for FPGA-based arithmetic | IEEE Conference Publication | IEEE Xplore