Parallel GC, for some reason implements specialized local methods to iterate over array slices (PSPromotionManager::process_array_chunk_work):

template <class T>
void PSPromotionManager::process_array_chunk_work(oop obj, int start, int end) {
  assert(start <= end, "invariant");
  T* const base = (T*)objArrayOop(obj)->base();
  T* p = base + start;
  T* const chunk_end = base + end;
  while (p < chunk_end) {
    claim_or_forward_depth(p);
    ++p;
  }
}

Investigate and potentially remove this code duplication.