recFNToRecFN

in hardfloat_sys

extern entity recFNToRecFN<
    #uint inExpWidth,
    #uint inSigWidth,
    #uint outExpWidth,
    #uint outSigWidth
>(
    control: uint<1>,
    in: uint<{inExpWidth + inSigWidth + 1}>,
    roundingMode: uint<3>,
    out: inv &uint<{outExpWidth + outSigWidth + 1}>,
    exceptionFlags: inv &uint<5>
);

Description taken from the Berkeley Hardfloat documentation:

Converts a recoded floating-point value to a different recoded format (such as from single-precision to double-precision, or vice versa). This module requires no special explanation.