FYI Cray developed an interesting allocator for scaling to a massive amount of threads https://dl.acm.org/doi/10.1145/1122971.1122999