Fix issues that are becoming visible on arm but might be a problem on all architectures. . Is it because of this: energy.c:539:53: note: within this loop for(i=0;i<ALPHASIZE;i++) for(j=0;j<ALPHASIZE;j++) for(k=0;k<=ALPHASIZE;k++) dr_dangle_dg_ar[i][j][k] = 0; . Perhaps that k<= in the k loop should be < like in the i and j loops so it doesn't go beyond the end of the array. . ALPHASIZE is 6, so the k loop would try go one too far. Every array allocated with ALPHASIZE certainly don't add one anywhere, so any look going from 0 to ALPHASIZE must use < not <= so there are two places in the file that are wrong.
