<Table> <Tr> <Td> 8 </Td> <Td> 16 </Td> <Td> 32 </Td> <Td> 64 </Td> </Tr> </Table> <Tr> <Td> 8 </Td> <Td> 16 </Td> <Td> 32 </Td> <Td> 64 </Td> </Tr> <Tr> <Th> Binary floating - point precision </Th> </Tr> <Tr> <Td> <Table> <Tr> <Td> 16 </Td> <Td> 32 </Td> <Td> 40 </Td> <Td> 64 </Td> <Td> 80 </Td> <Td> 128 </Td> <Td> 256 </Td> </Tr> </Table> <Table> <Tr> <Td> × 1⁄2 </Td> <Td> × 1 </Td> <Td> × 2 </Td> <Td> × 4 </Td> <Td> × 8 </Td> </Tr> </Table> </Td> </Tr>

A double word (on x86 systems) is 32 bits