s258 benchmark function could be cheated

[pinskia]

s258 contains this inside the inner loop

	  if (a[i] > 0.)
	    {
	      s = d[i] * d[i];
	    }

BUT a[i] > 0. is always true due to the initialization of a array which does:

if (!strcmp(name, "s258")) {
        set_1d_array(a, LEN_1D, any,frac);

set_1d_array does this for frac (SET1D_RECIP_IDX):

        for (int i = 0; i < length; i++) {
            arr[i] = 1. / (real_t) (i+1);
        }

If the compiler decides not to generate a conditional move, the HW might be able to predict the branch correctly through the loop and that is cheating what the benchmark is trying to benchmark here.