Take a look at the Real-Time Monitoring of Reconfigurable Application given by PREESM-SPIDER-PAPIFY Integration

Visit PAPIFY Wiki to see the usage of PAPIFY in the context of  Energy estimation & automatic parallelization of dataflow applications

Key Features

  • Run-time KPI monitoring is a key enabler for adaptivity
  • Best practices rules to minimize the overhead impact


  • Heterogeneous system designers/users with a need to monitor at run-time the actual values of the incumbent KPI

Benefits for the User

  • Common high level abstraction to different technologies Actor-based seamless instrumentation
  • PREESM Application Specification: At design time, PAPIFY uses both algorithm and architecture specifications described in PREESM by the user


  • PAPI available events: PAPI library provides PAPIFY with the available events existing within the target platform
  • User Monitoring Configuration: In the scenario file, users define for each actor the monitoring mode to be performed in execution time
  • PAPIFY-VIEWER: performance results to be extracted using PAPIFY


  • Instrumented Application: the application code automatically generated using PREESM is instrumented accordingly to the monitoring mode defined by the user at design time
  • Performance Data: performance results extracted directly from the Performance Monitoring Counters existing within the SW cores and HW resources of the platform. These data are associated to each actor of the application and to the resource that is executing that specific actor
  • PAPIFY-VIEWER: Graphical representation of a chronological view per actor of the activity of a dataflow specification. In addition, PAPIFY-VIEWER generates per-actor histograms of PAPI events

Role in the CERBERO Toolchain

  • Software and hardware monitoring

Tool Highlight

[Table: PAPIFY]

[Table 2: PAPIFY]

[Table: PAPIFY]

Web Page




Return to Toolchain