Academic part-time employees Krista Andersen, Alex McKee, Timofey Ovcharenko and Camilo Trevino recently completed a major task of mapping, disconnecting and reconnecting more than 2,000 node interconnect cables for Cosmos—the Dell Xeon cluster of 1024 CPU multiprocessors—when it was moved from Building 126 to Building 600. Since downtime for the supercomputer affects many JPL engineering, technical and scientific activities, performing the detailed job quickly was critical. Reassembly under the supervision of a senior member of the Supercomputing and Visualization Group involved installing the fiber in cages under the floor and reconnecting all the Myrinet switch parts. Six days after the system was turned off and 25 cabinets, each weighing 2,500 lbs., were relocated, Cosmos was available again for all users.