Msieve / News: Recent posts

A web forum for Msieve help: https://www.mersenneforum.org/forumdisplay.php?f=83

This update contains many important bug fixes, plus primality proving and use of updated CUDA libraries

This release adds several important bug fixes plus a major overhaul of the linear algebra, which is now much faster and scales much better on modern CPUs.

This release adds an enormous speedup for NFS polynomial selection running on graphics cards, along with many fixes, updated documentation and a much cleaner build system

This version includes many improvements to GNFS polynomial selection, especially the graphics card branch.

An incremental release with several important bug fixes, completely revamped documentation and many improvements in GNFS polynomial selection

This version includes a major overhaul of NFS polynomial selection, along with some bug fixes and an MPI speedup in the linear algebra.

This version include some bug fixes in the NFS linear algebra, as well as a little tuning for the new Fermi GPU from Nvidia. If you have a Fermi card and are using the precompiled GPU binary, replace the PTX files with those in the 'fermi' subdirectory.

This release includes new support for parallel linear algebra using MPI, as well as improved Nvidia GPU code and new degree-6 polynomial selection.

This release features many performance improvements and fixes to NFS polynomial selection; it also includes a merge of the CPU and GPU algorithms

This release contains code that uses Nvidia graphics cards to search for GNFS polynomials, as well as a few small fixes

This version contains many cleanups and bug fixes. An experimental branch is now available with code that lets Nvidia graphics cards perform some of the steps in NFS polynomial selection (and hopefully more than that someday!)

Msieve is the result of a multi-year development effort to create a sophisticated library for integer factorization. The v1.42 release is the first to be hosted on SourceForge.