import jdk.incubator.vector.*;

static Pair<int[]> rowAndColSums_8x8_vectorAPI(int[] data) {
  VectorSpecies<Int> species = IntVector.SPECIES_256;
  int[] rowSums = new[8];
  int[] colSums = new[8];
  IntVector vColSums = IntVector.zero(species);
  for y to 8: {
    IntVector v = IntVector.fromArray(species, data, y*8);
    rowSums[y] = v.reduceLanes(VectorOperators.ADD);
    vColSums = vColSums.add(v);
  }
  vColSums.intoArray(colSums, 0);
  ret pair(rowSums, colSums);
}