From APE to APE-100: from 1 to 100 GFLOPS in lattice gauge theory simulations

Abstract We briefly describe the APE processor, a parallel computer currently used in lattice gauge simulations. We also present in greater details the architecture and the implementation of APE-100, a fine grained SIMD processors for lattice gauge theory simulations, similar in structure to APE and designed to provide floating-point performance in the 100 Gflops range.