Last year I worked for a lab at National Laboratory of Scientific Computing (LNCC) in Petrópolis, RJ, Brasil. Was a pleasant experience. My colleagues are good people and I like LNCC research environment.
My job there actually was manage and help to choose a cluster. A secondary role of my work (which became first because we had some problems buying the cluster) was help in the parallel implementation of cryptographic algorithms in C++.
I am not a fluent guy in C++ so many stuffs here are discoveries. We heavily used NTL library which helped me a lot with the theorical part. From our implementation, there are two functions that helped me a lot and maybe useful for someone else.
vec_GF2E slice_vec(vec_GF2E x, int n) {
// Function that returns a subvector [0:n]
vec_GF2E result;
result.SetLength(n);
for (int i = 0; i < n; i++) {
result[i] = x[i];
}
return result;
}
mat_GF2E slice_mat(mat_GF2E O, int i_ini, int i_end, int j_ini, int j_end) {
// Returns a submatrix M[i_ini:i_end][j_ini:j_end]
mat_GF2E R;
R.SetDims(i_end - i_ini, j_end - j_ini);
for (int i = i_ini; i < i_end; i++) {
vec_GF2E temp_vet;
temp_vet.SetLength(j_end - j_ini);
int j_vet = 0;
for (int j = j_ini; j < j_end; j++) {
temp_vet[j_vet] = O[i][j];
j_vet += 1;
}
R[i - i_ini] = temp_vet;
}
return R;
}
One can see that we try to emulate some matrix and vector slicing in C++.
There is more NTL implementation-specific stuff but probably I will make a mess trying to explain.
That’s All.