I already said which pack formats are needed in order to properly encode floating point numbers

I asked for the format of the packet, and you have no yet provided that. There can only be one of those.

Maybe you mean you control the decoding, so we're free to create a packet format that's composed of those number formats. If so, packing the numbers as double-precision numbers sounds most appropriate.

my $packet = pack('d>*', @nums);
[download]