RNGSSELIB: Program library for random number generation. More generators, parallel streams of random numbers and Fortran compatibility

RNGSSELIB: Program library for random number generation. More generators, parallel streams of random numbers and Fortran compatibility. Abstract In this update, we present the new version of the random number generator (RNG) library RNGSSELIB, which, in particular, contains fast SSE real-izations of a number of modern and most reliable generators [1]. The new features are: i) Fortran compatibility and examples of using the library in Fortran; ii) new modern and reliable generators; iii) the abilities to jump ahead inside RNG sequence and to initialize up to 10 19 independent random number streams with block splitting method. Manuscript Title: RNGSSELIB: Program library for random number generation. More generators, parallel streams of random numbers and Fortran compatibility. No. of lines in distributed program, including test data, etc.: 9299 No. of bytes in distributed program, including test data, etc.: 1768030 Distribution format: tar.gz Does the new version supersede the previous version?: Yes Nature of problem: Any calculation requiring uniform pseudorandom number generator , in particular, Monte Carlo calculations. Any calculation requiring parallel streams of uniform pseudorandom numbers. The library contains both usual realizations and realizations based on SSE command set. Usage of SSE commands allows to substantially improve performance of all generators. Also, the updated library contains the abilities to jump ahead inside RNG sequence and to initialize independent random number streams with block splitting method for each of the RNGs. Reasons for the new version: 1. In order to implement Monte Carlo calculations, the implementation of independent streams of random numbers is necessary. Such implementation of initializing random number streams with block splitting method is added to the new version for each of the RNGs. Jumping ahead inside RNG sequence, which is necessary for the block splitting method, was also added for each of the RNGs. 2. Users asked us to add Fortran compatibility to the library. Fortran compatibility and the examples of using the library in Fortran for each of the RNGs are included in this version. 3. During last few years, the method of random number generation based on using the ensemble of transformations of two-dimensional torus, was essentially improved [5, 6]. Important properties, such 2 as high-dimensional equidistribution, were established for the RNGs of this type. The proper choice of parameters was determined, which resulted in the validity of the high-dimensional equidistribution property, and, correspondingly, the new high-quality RNGs GM29, GM55.4, GQ58.1, GQ58.3, GQ58.4 were proposed. These RNGs are …