AbstractPAPI is a widely used cross-platform interface to hardware performance counters. PAPI currently supports native events, which are those provided by a given platform, and preset events, which are pre-defined events thought to be common across platforms. Presets are currently mapped and defined at the time that PAPI is compiled and installed. The idea of user-defined events is to allow users to define their own metrics and to have those metrics mapped to events on a platform without the need to re-install PAPI. User-defined events can be defined in terms of native, preset, and previously defined user-defined events. The user can combine events and constants in an arbitrary expression to define a new metric and give a name to the new m...
As useful as performance counters are, the meaning of reported aggregate event counts is sometimes q...
This paper describes the notion of event-driven metrics as an alternative to generating metrics from...
The dynamic behavior of parallel programs can be disclosed by event-driven monitoring. A hybrid moni...
AbstractPAPI is a widely used cross-platform interface to hardware performance counters. PAPI curren...
In this work, a standard and unified method for monitoring hardware accelerators in Reconfigurable C...
PULSE builds on the latest Performance API (PAPI) project and extends it with software-defined even...
Performance analysis is an essential step for better software optimization, which is critical for em...
The purpose of the PAPI project is to specify a standard application programming interface (API) for...
This paper proposes a performance tools interface for OpenMP, similar in spirit to the MPI profiling...
For industrial systems performance, it is desired to keep the IT infrastructure competitive through ...
High performance computing is playing an increasingly important role in the scientific community. As...
Abstract—For more than a decade, the PAPI performance-monitoring library has provided a clear, porta...
International audienceEstimating safe upper bounds on task execution times is required in the design...
The performance of parallel and distributed systems depends significantly on how the programs '...
Performance Monitoring Counters (PMCs) have been traditionally used in the mainstream computing doma...
As useful as performance counters are, the meaning of reported aggregate event counts is sometimes q...
This paper describes the notion of event-driven metrics as an alternative to generating metrics from...
The dynamic behavior of parallel programs can be disclosed by event-driven monitoring. A hybrid moni...
AbstractPAPI is a widely used cross-platform interface to hardware performance counters. PAPI curren...
In this work, a standard and unified method for monitoring hardware accelerators in Reconfigurable C...
PULSE builds on the latest Performance API (PAPI) project and extends it with software-defined even...
Performance analysis is an essential step for better software optimization, which is critical for em...
The purpose of the PAPI project is to specify a standard application programming interface (API) for...
This paper proposes a performance tools interface for OpenMP, similar in spirit to the MPI profiling...
For industrial systems performance, it is desired to keep the IT infrastructure competitive through ...
High performance computing is playing an increasingly important role in the scientific community. As...
Abstract—For more than a decade, the PAPI performance-monitoring library has provided a clear, porta...
International audienceEstimating safe upper bounds on task execution times is required in the design...
The performance of parallel and distributed systems depends significantly on how the programs '...
Performance Monitoring Counters (PMCs) have been traditionally used in the mainstream computing doma...
As useful as performance counters are, the meaning of reported aggregate event counts is sometimes q...
This paper describes the notion of event-driven metrics as an alternative to generating metrics from...
The dynamic behavior of parallel programs can be disclosed by event-driven monitoring. A hybrid moni...