Block Sieving Algorithms Block Sieving Algorithms

Quite similiar to the Sieve of Erastosthenes, the best-known general algorithms for fac-toring large numbers today are memory-bounded processes. We develop three variations of the sieving phase and discuss them in detail. The fastest modiication is tailored to RISC processors and therefore especially suited for modern workstations and massively parallel supercomputers. For a 116 decimal digit composite number we achieved a speedup greater than two on an IBM RS/6000 250 workstation.