How does this code calculate pi with high precision?

Question

Here is the code:

#include <stdio.h>

long f[2801];

int main()
{
    long i = 0, c = 2800, d = 0, e = 0, g = 0;
    for (i = 0; i < c; ++i)
        f[i] = 2000;
    for (;;) {
        d = 0; 
        g = c * 2;
        if (!g)
            break;
        i = c;
        for(;;) {
            d += f[i] * 10000;
            --g;
            f[i] = d % g;
            d /= g;
            --g;
            --i;
            if (!i) break;
            d *= i;
        }
        printf("%.4ld",e+d/10000);
        e = d % 10000;
        c -= 14;
    }
    return 0;
}

My question is: How does this code calculate pi with high decimal precision and what is the math formula it uses?

Answer 1

This is a formatted copy of the PI program written by Dik T. Winter of the CWI institute of Holland. Originally written in an obfuscated form, in two or three lines, there are several variations by Dik and other that output different numbers of places of PI (e.g. 800, 15,000, etc.) based on evaluation of a mathematical series.

It is a class of programs known as 'spigot algorithms', designed to output a specific number of digits. You can find out more via a Google search on Dik Winter and 'spigot algorithms'. Some example hits:

Computing Pi in C a detailed analysis of the algorithm with unanswered questions.

Pi the Number, not the Movie