The latest version of the RIO API for ROSS checkpoints has just been released! The RIO project can be found here.
This version has many core features, including event capturing, allowing multiple partitions per MPI rank, and optimistic mode support. Additionally, all of the RIO documentation has been ported to the ROSS blog.
RIO is an API for creating restart checkpoints of a ROSS simulation. These checkpoints are created after the completion of a ROSS simulation and cannot be created during simulation execution. Thus, RIO is a system for checkpoint-restart operations.
RIO version 2 offers support for optimistic schedulers (along with serial and conservative schedulers).
This release includes a complete set of documentation. The documentation for RIO can be found on the ROSS website (Look for the RIO section on the archive page). The documentation includes:
A full implementation of the RIO API can be found within the PHOLD-IO model. Integrating RIO into an existing ROSS model should be straightforward and simple to understand.