Fast random number generation using 128 bit multimedia extension registers on Pentium class machines

In this work it is shown how 128 bit SSE2 multimedia extension registers, present in Pentium IV class 32 bit processors, may be used to generate random numbers at several times greater speed then when regular general purpose registers are used. In particular, a 128 bit algorithm is presented for the Marsaglia MWC1616 generator from the DIEHARD battery of random number generator tests, and its performance is compared to that of the conventional approach.