sped up marching cube triangle generation by 10x by using local list batches with IJobParallelForBatch to reduce NativeList<>.ParallelWriter locking adds (~0.35ms -> ~0.05ms)