Software
I am the main author of the Geneva library of distributed optimization procedures. The term "Geneva" stands for "Grid enabled evolutionary algorithms."
While, as the name suggests, the library currently implements predominently evolutionary algorithms, the intention is to transform this into a collection of different parametric optimization procedures, such as gradient descents or swarm algorithms. Geneva is targetted at multi-core machines, clusters, and ultimately also at wide-area computing infrastructures, such as Clouds and Grids.
Geneva is available under the Affero GPL v3 from the Launchpad Open Source portal. While the code is still under development, we believe it is mature enough for you to try it out in your own projects. Please do contact me, should you have questions, or if you wish to support this work.
Part of Geneva's development was aided by Karlsruhe Institute of Technology (KIT), as well as the Helmholtz Association of German Research Centres. I'd particularly like to thank these institutions for their kind help.
Geneva is also the main product of an upcoming spin-off from KIT - Gemfony scientific.