SST v10.0.0 Released
SST v10.0.0 is now available and can be downloaded here
The SST 10.0.0 release contains several changes to the simulation core and its component/sub-component APIs. In addition, a number of performance and stability issues have been addressed allowing for reduced simulation times and better scalability.
A selection of the improvement highlights are:
SST-Core:
- Added support for Python3
- Generalized statistics model to support custom output formats and visualizations
- Additional Python options for configuring statistics on SubComponents
- Remove backward compatibility paths for features deprecated in SST 9.0 and before
- Fixes known issues with OneShot events; these events are no longer available for Element use
- Support for Pin3 IPC tunnels
- Added ‘–enable-preview-build’ configure switch which disables deprecated features
- Support for adding additional send latency on links
- Added convenience functions for handling fatal errors, including new assert-style conditional fatal function
- Removed leftover references to Boost
SST-Elements:
- Expanded QoS support in Merlin
- Added Pin3 support to Ariel and Prospero. Due to incompatibilities with Pin3, compressed tracing (Prospero), compressed memory allocation tracking (Ariel/MemSieve), and the Ariel/GPGPU-Sim integration require Pin2. All other features are supported with Pin3.
- Improved Merlin Python module, including Python3 support
- Added support for multiple virtual networks in Firefly
- Ariel frontend (i.e., dynamic trace input) is now isolated into sub-component interface for improved customizability
- Removed PyProto element
- Removed Vanadis element
- Scheduler element is deprecated
- Substantial updates to memHierarchy caches and directories to improve model fidelity, enable greater customization, and enhance debugging infrastructure
- Fixes for known bugs in memHierarchy’s NonInclusive MESI coherence protocol
SST-Macro:
- Miscellaneous bug fixes and performance improvements
- Added new contention models for quality-of-service
- Added new network statistics and visualizations
- Improved auto-skeletonizer stability and ease-of-use
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
