Vampir 8.4 released

Nov
2014
At the SC14 in New Orleans we presented our most recent improvements in performance tracking and visualizing on the road to exascale.
Vampir
Vampir 8.4 introduces new performance charts, customizable performance metrics, and extended data filtering and preselection capabilities. Selected highlights include:
- New Summary Timeline – an intuitive and scalable overview of all concurrent activities
- Extended folding capabilities in the Master Timeline and Communication Matrix View to generate summary information
- New performance metrics (e.g. function invocation rates over time)
- Performance metrics now support log scales
- Enhanced event subset loading and event filtering for a better analysis focus
- Presentation mode for training activities
- Support for traces generated by Score-P 1.3
Score-P
Score-P is the primary code instrumentation and run-time measurement framework for Vampir 8, and also works natively with Scalasca, TAU, and Periscope. The latest version 1.3 provides numerous improvements and new features. Selected highlights include:
- Instrumentation and monitoring of parallel applications using Pthreads and SHMEM
- Fine-grained information about the system topology on Cray machines using the Process Manager Interface (PMI)
- Dependency tracking between CUDA host and device activities
- Platform support for the K computer and Fujitsu FX10 systems
- Support for CUDA 5 and CUDA 6
For more information and downloads visit www.score-p.org.