... Computations

Work reported herein has been supported in part by the Deutsche Forschungsgemeinschaft (DFG).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... small

say, smaller than the maximal representable machine integer, i.e. smaller than INT_MAX

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... machine

By default, we declare Exponent_t to be of type short which seems to be a good alternative between flexibility and efficiency.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...)

Using our previously introduced terminology, we have s_e = 2 for t_n,s (resp. 4 for t_n,c); and m = s_w/s_e = 2 (resp. 4 for t_n,c) on a 32-bit machine (like the HP C160, or Pentium Pro) and m = 4 for t_n,s (resp. 8 for t_n,c) on a 64-bit machine (like the DEC Alpha).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... examples

We tried to follow the paths of these examples to their origins as much as possible.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.