@Zoby, this time it is my program you should try to analyse. It isn't that hard but I hope you have some fun with it. As small hint: it is related to a mathematician who participated in an infamous project.

            #include <stdio.h>
     #define p(i,t,x)for(i=t;i<x;i++)
    main(){int m=0400,i,j,r,u[4*(m*m+m
)+1],p=2*(m+m*m)          ,t=0,l[]={2,0,1,1
,1},q[4*(m*                     m+m)+1],d[]=
{1,-m*2-                          1,-1,m*2+1
,1},x;u                           [p]=t++;p(j,
 0,4                              *(m*m+m)+1)q[j
                                   ]=j<2?0:1;p(j,
                                    2,4*(m*m+m)+1)
                                    {x=j;while((x
                                   +=j)<4*(m*m+
                                 m)+1)q[x]=0;}
                                p(i,1,m+1)p(r,
                               0,5){l[r]+=r<
                             1?0:2;p(j,1,l[
                      r])u[p+=d[r]]=t++;
                   }printf("P%d\n%u"\
                " %u\n%d\n",2|
                1,(m*2
                +1),(m
                *2+1),
                0xff);
                p(j,0,



             4*(m*m+m)+1)
            printf("0 %u"\
            " 0 ",q[u[j]]*
                128);}

Use gcc test.c -o test && ./test > test.ppm to generate an image.

Update: Of course, it generates the Ulam spiral. It starts by generating an array with prims using the sieve of Eratosthenes (0 == not a prime, 0 != a prime) and a translation array to map positions inside a 513x513 sized image of a spiral to its linear offset of the unrolled spiral. The last step is to print the informations to stdout.