The speed, bandwidth and cost characteristics of today's PC graphics cardsmake them an attractive target as general purpose computational platforms. Highperformance can be achieved also for lattice simulations but the actualimplementation can be cumbersome. This paper outlines the architecture andprogramming model of modern graphics cards for the lattice practitioner withthe goal of exploiting these chips for Monte Carlo simulations. Sample code isalso given.
展开▼
