AnchorPower is a very fast, simple, and general random number generation method based on Inverse Transform and Taylor Series. Its main contributions are speed, simplicity, and generality. Speed tests have shown that it outperforms many excellent RNG implementations, including GNU Scientific Library and UNURAN, by a factor of 2 or more. The quality of the method has also been confirmed by standard randomness tests.
| Tags | Scientific/Engineering |
|---|---|
| Implementation | C++ |
Recent releases


Release Notes: This release adds a major update of the C library, supports hundreds of distributions, provides computation of distributional functions (CDF, quantile, density, derivatives, etc.), and adds much better programming documentation.


Release Notes: A C programming interface has been created to allow any C/C++ application program to take advantage of the AnchorPower method for generating non-uniform random numbers. The Intel Compiler requirement has been removed without any impact on performance. This should allow many more users to utilize this tool.


Release Notes: In this latest release, AnchorPower demonstrates truly impressive performance. When pitched against the very fast and very optimized random number generators from the Intel Vector Statistical Library (probably the fastest implementation currently available on Intel processors), AnchorPower shows a speed advantage for every distribution tested. Furthermore, for the important Gamma and Beta distributions, the speed advantage is a factor of 4 to 6. This release should be of particular interest to people who are interested in Monte Carlo simulation.


No changes have been submitted for this release.