The pla85900 computations were carried out on clusters of Intel Xeon compute servers and AMD Athlon compute servers, between February 2005 and April 2006. The clusters were also used for other purposes during the computing period, with the TSP computations taking over when the machines were otherwise free.
The total amount of computer usage for the computations was approximately 136 CPU years, scaled to a 2.4 GHz AMD Opteron 250 compute node. This is just under 4.3 billion CPU seconds.