Two more examples: * Nvidia intentionally capping floating point precision on 1/8th on their GTX 480 cards, since they cost a fraction of their Tesla cards. Bo

Two more examples: * Nvidia intentionally capping floating point precision on 1/8th on their GTX 480 cards, since they cost a fraction of their Tesla cards. Both based on latest Fermi architecture. * In automotive industry: two otherwise identical cars, but with different versions of engine microcontroller driver, can have many thousands of dollars difference in price.