Performing Floating-Point Accumulation on a Modern FPGA in Single and Double Precision | IEEE Conference Publication | IEEE Xplore