SST v7.0.0 Released
SST v7.0.0 is now available and can be downloaded here
The SST v7.0.0 release follows the standard release process of having our point releases provide improved performance and stability. In this release, we have focused on many under the hood fixes to address bugs as well as improving predictive accuracy and simulation speed.
A selection of the improvement highlights are:
SST-Core:
- Overhaul of the SST element information description system to more easily enable externally development elements to interact with the SST framework
- Removal of Boost as a dependency from the SST core
- New SubComponent handling to make configuration/model construction easier
- Improved thread scaling
- Several bug fixes for output information related to component connectivity graphs
- Implementation of a “stop at” wall time feature for working within scheduled cluster environments
- HDF5 support for statistics output
- Support for describing co-ordinates of components in Python configuration (to assist in visualizations of simulations)
SST-Elements:
- Initial SST models for Non-Volatile Memory (Messier)
- Models for TLB handling (Samba)
- Improvements to CramSim (developed by IBM) to enable threaded simulations
- Support for Boost as an external dependency
- SimpleSimulation demonstration element added as an example component
- Support for memory modeling in large-scale network analysis
SST-Macro:
- Accuracy improvements in network models
- Improved documentation and updated example configuration files
- Integrated job launcher components for simulating PBS or SLURM-like batch systems
- Merlin integration. Trace replay and skeleton apps now compatible with Merlin router models.
- Beta release of OTF2 trace replay skeleton
- Beta release of Clang-based auto-skeletonization source-to-source tools
Current release information can be found here
We look forward to hearing your successes with the latest release!
SST Research and Product Teams
Sandia National Laboratories, USA
