Re: Large Floating Point Numbers



On Apr 10, 1:12 am, "D. Baruth" <SeeBe...@xxxxxxxxxxxxxxxx> wrote:
user923005 <dcor...@xxxxxxxxx> wrote in messagenews:1176190696.055721.322150@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

: On Apr 9, 3:28 pm, "D. Baruth" <SeeBe...@xxxxxxxxxxxxxxxx> wrote:
: > user923005 <dcor...@xxxxxxxxx> wrote in messagenews:1176148438.662566.33550@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
: >
: > : On Apr 7, 12:11 pm, "D. Baruth" <SeeBe...@xxxxxxxxxxxxxxxx> wrote:
: > : [snip]
: > : > What are "x" and "y" in z=x*y, z=x/y, etc... ? I was asking about the
: > : > "operands" in the operations.
: > :
: > : They are multiple precision numbers, with precision set to 100, 1000,
: > : 10,000 for the three different runs.
: >
: > Can you specify?
:
: Yes, you specify the precision on the command line. Look at the
: source for the driver that I supplied and the usage is obvious.

For precision 100, 1000 or 10,000, please specify: x = ?, y = ?;
You may also point to a file containing the high precision numbers.

You supply the on the command line to the main() program as the first
argument passed into main, as is clear from the snippet shown below.
I guess you are not a C++ programmer.

int
main(int argc, char *argv[])
{
int n, prec, st, st2, N, i;

if (argc != 2 && argc != 3) {
fprintf(stderr, "Usage: timing digits \n"); exit(1);
}
n = atoi(argv[1]);

float_format_t precision = float_format(n);
cl_F x = sqrt (cl_float (3, precision)) - 1;
cl_F y = sqrt (cl_float (5, precision));
cl_F z;

cout << "Prec=" << precision << endl;


.